Inorganic Materials

Chemistry
Mark T. Weller

QD
151.2
.W444
1994
(\

Inorganic
Materials Chemistry
MarkT. Weller
University of Southampton

OXFORD
UNIVERSITY PRESS
 QtMbl.Z .W444 in4

OXFORD
UNIVERSITY PRESS

Great Clarendon Street, Oxford 0x2 6dp

Oxford University Press is a department of the University of Oxford.
It furthers the University’s objective of excellence in research, scholarship,
and education by publishing worldwide in
Oxford New York
Auckland Bangkok Buenos Aires Cape Town Chennai
Dar es Salaam Delhi Hong Kong Istanbul Karachi Kolkata
Kuala Lumpur Madrid Melbourne Mexico City Mumbai Nairobi
Sao Paulo Shanghai Taipei Tokyo Toronto
Oxford is a registered trade mark of Oxford University Press
in the UK and in certain other countries
Published in the United States
by Oxford University Press Inc., New York
© M. T. Weller, 1994
The moral rights of the author have been asserted
Database right Oxford University Press (maker)
First published 1994
Reprinted 1996, 1999 (twice), 2001, 2002
All rights reserved. No part of this publication may be reproduced,
stored in a retrieval system, or transmitted, in any form or by any means,
without the prior permission in writing of Oxford University Press,
or as expressly permitted by law, or under terms agreed with the appropriate
reprographics rights organization. Enquiries concerning reproduction
outside the scope of the above should be sent to the Rights Department,
Oxford University Press, at the address above
You must not circulate this book in any other binding or cover
and you must impose this same condition on any acquirer
A catalogue record for this book is available from the British Library
Library of Congress Cataloging in Publication Data
Data available
ISBN 0 19 855799 X (Hbk)
ISBN 0 19 855798 1 (Pbk)
7 9 10 8
Printed in Great Britain by
The Bath Press, Bath
Series Editor’s Foreword
Solid state chemistry has undergone a renaissance in this country in the last ten
years. Synthetic methods, characterization techniques and underlying theories
have now been developed that make a wider range of problems accessible to the
chemist. Many of these deal with technologically important and intellectually
demanding problems that make this an essential component of an
undergraduate course.
Oxford Chemistry Primers are designed to give a concise introduction to all
chemistry students by providing material that would usually be covered in an
8-10 lecture course. As well as giving up-to-date information, this series
provides explanations and rationales that form the framework of an
understanding of inorganic chemistry. Mark Weller presents the reader with an
authoritative description of modem solid state chemistry, and the role of X-ray
diffraction in it. Few if any standard inorganic textbooks can offer the kind of
treatment that is achieveable by an expert in the field.
John Evans
Department of Chemistry, University of Southampton

Preface
The chemistry of inorganic materials is rapidly developing as an important
component of university degree courses, reflecting recent advances at the
forefront of research. Examples of such materials include the high-temperature
superconductors, molecular sieves, fullerenes and layer intercalates. These
topics also impinge on other undergraduate courses such as physics and
materials science. However, the subject of inorganic solids is rarely covered in
any depth in the major textbooks of inorganic chemistry. In addition the
techniques used by the solid state chemist to prepare and characterize inorganic
materials are seldom introduced alongside, or integrated with, the descriptive
materials chemistry. This primer is intended to fill this omission by providing
an affordable but comprehensive introduction to the chemistry of solid
inorganic materials. The introductory chapters provide the foundation,
reviewing basic crystallography and featuring a step-by-step guide to the
analysis of powder diffraction data from simple systems. The synthesis of a
wide range of inorganic material types is covered. These central techniques and
concepts are developed in relation to fundamental solid state chemistry with
discussion of simple oxide systems and a guide to the physical properties of
solids. Finally the contemporary aspects of the subject are reviewed,
integrating many of the ideas introduced in the preceding chapters.
I thank the members of my research group for their forbearance during the
construction of this primer and Dr David Currie for useful suggestions and
improvements. Finally I am indebted to three generations of my family for their
constant support and abundant affection.
Southampton M.T.W.
March 1994
Contents
1 Basic crystallography 1

2 Application and interpretation of powder X-ray diffraction data 15

3 The synthesis of inorganic materials 26

4 Transition metal oxides 37

5 Electronic, magnetic and optical properties of inorganic 49

materials

6 Non-stoichiometry 60

7 Zeolites, intercalation in layer materials and solid electrolytes 71

8 Some recent developments in inorganic materials chemistry 82

Answers to problems 89

Bibliography 90

Index 91
1 Basic crystallography

1.1 Introduction

The structures of inorganic materials, the majority of which are crystalline in

the solid state, are a key feature in the control of their chemistry and physical
properties. The techniques used to identify and investigate such materials
involve characterisation of the bulk structure. In order to discuss inorganic
materials in some detail in the later chapters of this book it is important to
cover the background to crystallography and structure determination.

1.2 Crystal systems and unit cells

All crystalline materials adopt, in the solid state, a regular distribution of

atoms or ions in space. The simplest portion of the structure which is repeated
by translation, and shows its full symmetry, is defined as the unit cell. In a
two-dimensional array of ions, such as that shown in Fig. 1.1, the unit cell
consists of a parallelogram. Any parallelogram may be chosen as a unit cell
provided that translation along either of the cell directions repeats exactly the
chosen unit.

■ o ■ o

Fig. 1.1 Two-dimensional array

demonstrating the choice of unit
cell.
A

■ O ■ O
o ■ o ■
However, the unit cell is normally selected to be the simplest of these
Note that it is possible, and equally
repeating units. Hence, in Fig. 1.1, the unit cell could be chosen as A for valid, to choose a unit cell with the
which displacements parallel to either edge of the square by the dimension of corners not coincident with the
the unit cell produce a new position which is indistinguishable, in terms of atoms, for example E.
2 Basic crystallography

cell content and environment, from the original. The parallelogram B is also
a suitable choice for the unit cell as it has the same area as A and shows the
full translation symmetry. Square D would also be an acceptable choice of
unit cell in terms of demonstrating the translational symmetry of the array, but
is larger than A and B. However, parallelogram C is not a unit cell as
translation parallel to one side by the length of the parallelogram places a
corner originally at a ■ site on a O site; that is, the 'cell' does not show the
translational symmetry of the ion array.
The basic unit cell in three dimensions is a parallelepiped with side lengths
and angles defined as in Fig. 1.2(a). The angles and lengths used to define the
size of the unit cell are known as the unit cell parameters; by convention the
angle between a and b is y, between b and c is a and between a and c is (3.
These parallelepipeds must stack together so as to form the structure and be
completely space filling. That is, translation parallel to any of the sides by the
length of that side will generate a new position for the unit cell; unit cells thus
generated will fit perfectly together so that no part of the space is excluded.
This is shown for two parallelepiped types in Fig. 1.2(b).
The unit cell shown as the parallelepiped in Fig. 1.2(a) has no symmetry
in that the cell parameters and angles may take any values. An increasing
level of symmetry produces relationships between the various cell parameters
and leads to the seven crystal classes. These are summarised in Table 1.1 and
sketched in Fig. 1.3.
A few conventions are worth noting when assigning particular lattice
parameters for a particular crystal class. In monoclinic systems the non-right
angle is normally taken as (3; in the hexagonal and tetragonal systems the
non-equivalent lattice parameter is normally taken as c. The trigonal systems
Fig. 1.2(a) General three- can all be redefined using hexagonal lattices but this group is still generally
dimensional unit cell definition,
regarded as a separate crystal system.
(b) Packing of unit cells in three
dimensions.

1.3 Fractional atomic coordinates and projections

The position of an atom within a unit cell is normally described using

fractional coordinates. With respect to the unit cell origin, an atom within

Fig. 1.3 Units cells of the seven crystal systems. From left to right, cubic, tetragonal, orthorhombic, monoclinic, triclinic
rhombohedral and hexagonal.
 Inorganic materials chemistry 3

Table 1.1 The crystal systems

Unit cell dimensions Crystal class Examplet

a= b= c a = p = y = 90° Cubic NaCI, MgAI204, C^Ka

a = bit c a = p = y = 90° Tetragonal K2NIF4, Ti02,

BaTiOg (298 K)

a* c
“CE>

CD
Oo
Orthorhombic YBa2Cu307
P
II
II
II
a* bit c a = y = 90° |3 * 90° Monoclinic kh2po4

ait bit c a * p * y * 90° Triclinic

a - bit c a = P = 90° y= 120° Hexagonal LiNbOg

a= b= c a = p = y it 90° Trigonal / Rhombohedral BaTiOg below -80°C

+ The majority of the example compounds are discussed later in the text.

the unit cell displaced by x x a parallel to a, y x b parallel to b and z x c

parallel to c is denoted by the fractional coordinates ix,y,z). This method of
describing the position of atoms within the unit cell is equally applicable to
all crystal systems and cell sizes.
Three-dimensional representations of complex structures are frequently
difficult to draw and to interpret and may be ambiguous. Often a clearer and
faster method of representing three-dimensional structures is to draw the
structure in projection viewing the unit cell down one direction, typically one
of the unit cell axes. Normally the axis chosen is at 90° to the others and,
hence, for triclinic systems a projection of the structure is, generally, of more
limited use. The positions of the atoms relative to the projection plane are
denoted by the fractional coordinate above the base plane. For example, the
structure of zinc blende, ZnS, shown three-dimensionally in Fig. 1.4(a) may
be viewed in projection down the z direction as in Fig. 1.4(b).

Fig. 1.4 Zinc blende structure viewed conventionally (a) and drawn as a projection (b).
4 Basic crystallography

1.4 Lattices

A lattice is defined as an array of equivalent points in one, two or, more

normally for inorganic materials, three dimensions. The lattice provides no
information on the actual positions of atoms or molecules in space but shows
the translational symmetry of the material by locating equivalent positions.
The environment of an atom placed on any one of these lattice points would

♦ ♦♦♦♦♦♦♦ be identical to that placed on any other lattice point. The simplest illustration
of this is a one dimensional lattice consisting of an infinite series of equally
spaced points along a line. The distribution of other lattice points about any
Fig. 1.5 A line of equally spaced randomly selected lattice point is identical, Fig. 1.5.
points as a one dimensional lattice. In three-dimensional crystal structures four lattice types are found. The
simplest lattice type is known as primitive, given the symbol P, and a unit
cell with a primitive lattice contains a single lattice point. That is, the only
purely translational symmetry is that of the unit cell. Figs. 1.6a and b show
examples of primitive lattices; lattice points are normally located at the
corners of the unit cell parallelepiped
For the remaining three lattice types, as well as the translational symmetry
of the unit cell, there is additional transitional symmetry within the unit cell.
A second lattice type is body centred, which is given the symbol I, and an
example of a body centred unit cell is shown in Fig. 1.7. This shows lattice
points at the cell corners and, for body centring, the additional lattice point is
at the cell centre with fractional coordinates (VV/V/2). This means that if an
atom is placed on a general position within a body centred unit cell with
fractional coordinates (x,y,z) the lattice will generate a second, identical, atom
at the position with fractional coordinates (X+V2, y+Vi, Z+V2).
A three-dimensional lattice which has lattice points at the centre of all the
unit cell faces, as well as at the corners, is known as face centred and given
the symbol F, Fig. 1.8. The additional translational symmetry in this case
consists of the three elements +06,V4,0), +(0,1/2,1/2) and +(1/2,0,1/6), hence for an
atom on a general site (x,y,z) three additional identical atoms will be
generated within the unit cell with the coordinates (x+V2,y+l/2,z), (x,y+V2, Z+V2)
and (x+l/2,y,z+V2).
Finally a lattice which has points in just one of the faces is also known as
face centred but given the symbol A, B or C, Fig. 1.9. A C type lattice refers
Fig. 1.6 Primitive lattices (a) cubic
to the case where the additional translational symmetry places lattice points
and (b) triclinic. at the centres of the faces delineated by the a and b directions as well as at
the origin. In fractional coordinate terms, in addition to the general site (x,y,z)
for a C centred lattice, a second site is generated at (x+Vz, y+lA, z). The A and
The environment of the additional
B face-centred lattices are obtained in an identical manner but the additional
atoms generated by the centring
will be identical in all aspects to lattice points occur in the be and ac planes respectively. However, the A and
that of the original atom. B descriptions are not normally used, as redefinition of the a, b and c
Redefining the origin of the lattice directions will produce the C centred description.
at the face centred or body centred The lattice types describe the pure translational symmetry of the structure
positions produces exactly the
and show equivalent positions within unit cells related by translational
same arrangement of points.
symmetry. It should be noted that the lattice type does not provide definitive
information on the number of atoms or molecules in the unit cell. In the case
 Inorganic materials chemistry 5

1.7 1.8 1.9

Figs. 1.7, 1.8 and 1.9 Body centred, I, face centred, F, and face centred, C, lattices
respectively.

An example would be the unit cell

of a material crystallising with a primitive cubic lattice there may more than of YBa2Cu307, Fig. 8.5 which is
one atom or molecule of a particular kind within the unit cell, but their primitive but contains two copper
positions in terms of their environment will be quite different. atoms in different environments,
square planar and pyramidal.

1.5 Bravais lattices

The four lattice types (P,I,F,C) can be combined with the seven crystal
classes, in Table 1.1, to give rise to all the possible variations, which are
known as the Bravais lattices. These fourteen unique possibilities are
summarised in Table 1.2 and it can be seen that not all combinations of lattice
type with crystal system give rise to a different Bravais lattice. This is
generally because an alternative description in terms of one of the fourteen

Fig 1.10 Equivalence of face

centred and body centred
tetragonal lattices.

(a) (b)
6 Basic crystallography

Table 1.2 The fourteen Bravais lattices.

Crystal system Bravais lattices

Cubic P.I.F

Tetragonal P,l

Orthorhombic P.C.I.F

Monoclinic P,C

Triclinic P

Hexagonal P

T rigonal/Rhombohedral P(R)*

* The primitive description of the rhombohedral lattice is normally given the symbol R.

unique lattices is possible. For example in the case of a 'face centred

tetragonal’ unit cell, Fig. 1.10a, a new unit cell may be chosen which is body
centred tetragonal, Fig. 1.10b.

1.6 Lattice planes and Miller indices

The lattice points which form an array in two or three dimensions showing
the translation symmetry of the structure may be connected by lattice lines
(two dimensions) or lattice planes (three dimensions). Each line or plane is
a representative member of a parallel set of equally spaced lines or planes
and each lattice point must lie on one of the lines or planes. Two possible
lattice lines for a two-dimensional square lattice are shown in Fig. 1.11.
These lattice lines and planes are labelled using Miller indices. For a three-
dimensional unit cell, three indices are required and designated conventionally
h, k and /; the Miller indices for a particular family of planes are usually
written (h,k,l) where h, k and / are integers, positive, negative or zero. In the
case of a three-dimensional structure the derivation of Miller indices may be
illustrated by considering the two adjacent lattice planes from the same set
which cut the a, b and c axis at the origin and at the least distance along the
cell directions. The Miller indices of this family of planes are given by the
reciprocals of the fractional intercepts along each of the cell directions. For
example, using the lattice plane shown in Fig. 1.12, the intercepts are at Vi
x a, 1 x b and Vb x c. The Miller indices of this plane, one of a set of planes,
are thus (2,1,3). The intercept along the unit cell a direction gives h, along b
gives k and along c gives l. For planes which are parallel to one of the unit
cell directions the intercept is at infinity and, therefore, the Miller index for
this axis is l/°° = 0. Fig. 1.13 shows an example of a (0,0,1) plane for a
triclinic unit cell.
The separation of the planes is known as the d-spacing and is normally
denoted dhk[. From Fig. 1.12 it follows that this is also the perpendicular
distance from the origin to the nearest plane. The relationship between d-
spacing and the lattice parameter can be determined geometrically but is
dependent upon the crystal system. In two dimensions simple trigonometry
 Inorganic materials chemistry 7

Fig. 1.11 Lattice lines in a two-

dimensional lattice and the
definition of dhkl.
103 planes 102 planes

Table 1.3 Expressions for d-spacings in the different crystal systems.

Crystal System Expression for dhk/ in terms of lattice parameters and Miller indices

Cubic
1 h2 + k2 + l2
d2 a2

Tetragonal
1 h2 + k2 ^ l2
2 2
d2 a c

Orthorhombic
1 h2 k2 i2
d2 a2 + b2 c2

Hexagonal
1
d2
4f
3l
h2 + hk
a2
+ k2 )
J
l2
c2

Monoclinic
1 _ 1 {hi k2 sin2p ( l2 2Wcosp)
d2 sin2p { a2 b2 c2 ac J

Triclinic Complex expression

8 Basic crystallography

Fig. 1.12 Definition of Milter indices

in three dimensions. A pair of
planes with Miller indices (213).

may be used, with the definitions in Fig. 1.11, to derive the relationship
between the Miller indices and d-spacing. Hence, from Fig. 1.11,

dhoi dhot (1.1)

sin<p = COScp =
(alh) (ell)

giving

(1.2)

Extension of this expression to a three dimensional case with right angles

between the lattice directions, an orthorhombic system, produces

h2 (1.3)
d2
Fig. 1.13 (001) plane for a triclinic
unit cell. For the other crystal classes the relationships summarised in Table 1.3 apply.

1.7 X-Ray diffraction

X-Rays interact with electrons in matter and a beam of X-rays impinging on

an inorganic material will be scattered in various directions by the atomic
electrons. If the scattering centres are separated by distances comparable to
While picometres are the SI unit
the wavelength of the X-rays then interference between the X-rays scattered
that should be used in from particular electron centres can occur. For an ordered array of scattering
crystallography, angstroms are still centres this can give rise to interference maxima and minima.
 Inorganic materials chemistry 9

Distances between atoms or ions in solids are typically a few hundred very widely employed; 1 A = 1CT10
picometres ( equivalent to a few Angstroms ), which is comparable to the X- m. This is because interatomic
ray wavelengths produced by bombardment of metals by electrons. A study distances are conveniently
expressed in angstroms, being
of the X-rays scattered by crystalline solids provides a wealth of structural
typically between 1 and 4 A. Both
information. units are quoted where possible
throughout the text.

Generation of X-rays
*A beam of electrons striking a metal plate will, provided the electrons are
sufficiently energetic, eject an electron from one of the metal atom core
orbitals. Filling of this hole by electron decay from a higher energy orbital
occurs with the emission of radiation. For moderately heavy metal atoms, for
example first and second row transition metals, the transitions such as 3p—»ls
correspond to an emitted radiation wavelength in the range 0.5 to 3.0 A ( 50- s
<f>
c
300 pm). In copper, for example, core electron vacancies formed by ®

bombardment with electrons can be filled by decay from various higher

energy electrons and thus the spectrum of X-rays obtained contains a number
of intense maxima corresponding to the energies of these various transitions.
In addition, X-rays can also be generated by the slowing down (or Wavelength

bremsstrahlung) of the electrons as they enter the metal target. A typical X-

ray spectrum obtained from bombarding a metal target is shown in Fig. 1.14 Fig. 1.14 The X-ray spectrum
with a background formed from the bremsstrahlung radiation and intense obtained from a copper target.
sharp maxima corresponding to the quantised electron transitions. The intense
lines are labelled with respect to the electron shells which are involved in
producing the X-ray. The filling of a vacant Is orbital, principal quantum
number one, is given a symbol K; this may be achieved, in a copper atom, by
electron decay from the 2p or 3p levels and the X-ray produced by the
transitions 2p -> Is and 3p Is are termed Ka and KB respectively. Both these
lines are in fact close doublets due to the spin multiplicity in the p shells;
hence the Ka line ( A -1.54 A ) in the copper spectrum consists of two lines
at 1.5406 A and 1.5444 A termed Kal and Ka2. A transition which fills a 2s
or 2p shell, principal quantum number 2, is given the symbol L and so on
through the alphabet for higher principal quantum number shells.
In order to carry out a diffraction experiment a single X-ray wavelength is
desired. This may be achieved using a crystal monochromator in which the
X-ray beam emerging from the X-ray tube impinges on a single crystal at a
chosen, fixed orientation. For a particular angle 0 and employing Bragg’s Law
(see the following section) only one wavelength can be diffracted from the
crystal. By adjusting 0 it is possible to select one wavelength from the X-ray
spectrum, normally the most intense KaI. A second way of removing
unwanted wavelengths from the X-ray beam is to use a filter. This normally
consists of a foil of a metal with an atomic number one or two below that of
the metal used as the target to generate the X-ray spectrum. In the metal of
the foil, transitions between energy levels will require slightly less energy than
the corresponding transitions in the metal target due to a reduced nuclear
charge and this leads to a high absorption coefficient for these X-rays. Hence,
for an X-ray beam generated at a copper target with a nickel filter, X-rays
with wavelengths less than =1.5 A will be strongly absorbed. This will remove
10 Basic crystallography

from the copper X-ray spectrum the majority of the bremsstrahlung radiation
and, more importantly, lines such as the KB. The X-ray beam emerging from
the filter is, thus, almost monochromatic, consisting chiefly of Ka radiation.
Both components of the Ka line are still present but they are so close in
wavelength that diffraction experiments are possible producing close doublets
in the diffraction pattern.

1.8 Scattering of X-rays by crystalline solids

Scattering of X-rays from crystalline solids can be demonstrated by

consideration of the diffraction from points on a set of lattice planes such as
those shown in Fig. 1.15.

The scattering of the impinging X-ray beam from the points A and D in
neighbouring planes will produce in phase diffracted X-ray beams
(constructive interference) if the additional distance travelled by the X-ray
photon scattered from D is an integral number of wavelengths. This path
21
◄-► difference BD + DC will depend on the lattice spacing or dhkl, where hkl are
the Miller indices for the planes under consideration, and will also be related
to the angle of incidence of the X-ray beam, 0. For an integral wavelength
pathlength difference the following relationship between 0 and dhkl can be
obtained

path difference = BD + DC = 2.dhkl.sinQ = nl (1.4)

where n is an integer and 1 is the X-ray wavelength. The expression

2dsinQ=n\ is known as the Bragg equation. In general n is always taken as
1 unity; the reason for this is illustrated in Fig. 1.16. Rewriting the Bragg
equation as A, = 2. (dhk/n). sin0 and consideration of Fig. 1.16 shows that the
Fig. 1.16 Equivalence of n = 2 for
the (100) reflection and n = 1 for higher order diffraction maxima, n = 2, is equivalent in representation to
the (200) reflection. diffraction from a set of planes with half the separation. Hence, the second
 Inorganic materials chemistry 11

order diffraction from, for example, the 100 plane is identical to the first order Note that slightly off the Bragg
diffraction from the 200 plane. In general, it is always possible to select angle the diffracted X-ray beams
lattice planes such that n is unity. from successive planes will only be
slightly out of phase. However, in a
In a crystalline material an infinite number of lattice planes with different
crystal with a large number of
Miller indices exists and each set of planes will have a particular separation, planes, scattering from each
dhkl. All the various d-spacings possible can be obtained by consideration of successive plane will be further
Table 1.3 for a particular crystal system. Each of the d-spacings generated for and further out of phase. Over the
a set of lattice parameters and choice of integers for h,k and 1 can, in whole crystallite this leads to
destructive interference.
’principle, through the Bragg equation give rise to a diffraction maximum it
a particular diffraction angle, 0. The positions of these lines, in terms of the
angle 0, can be obtained by combining the Bragg equation with the
expression for the d-spacing for a particular crystal system and eliminating
dhkl. For example, in the cubic system

sin20 = { h2 + k2 + l2 } (L5)
4a2

The scattering of an X-ray beam by a crystal will thus give rise to a large
number of diffraction maxima, the positions of which are related to the lattice
parameters, the Miller indices and the X-ray wavelength. Hence, by studying
the diffraction of X-ray beams by crystalline solids structural information can
be obtained.
The diffraction maxima in X-ray diffraction patterns are normally called
reflections. Thus X-ray scattering from the 213 plane is termed the 213
reflection in diffraction pattern. This terminology whilst incorrect in terms of
the process involved in origin of the maxima is so widely used that it is the
accepted notation.

1.9 The powder technique

In order for an X-ray beam, diffracted from a particular lattice plane in a

crystalline sample, to be detected, the orientation of the X-ray source, crystal
and detector must be correct. A powder or polycrystalline sample contains an
enormous number of very small crystallites, typically 10~7 - 10-4 m in
dimension and these particles will adopt, randomly, the whole range of X-rays
possible orientations. An X-ray beam striking a polycrystalline sample will,
Sample
therefore, be diffracted in all possible directions as governed by the Bragg
equation. The effect of this is that each lattice spacing in the crystal will give
rise to a cone of diffraction as shown in Fig. 1.17. In fact, each cone consists Fig. 1.17 Diffraction cones
of a set of closely spaced dots each one of which represents a diffraction from obtained from polycrystalline
samples
a single crystallite within the powder sample. With a very large number of
crystallites these dots join together to form the cone.

Experimental methods
In order to obtain powder X-ray diffraction data in a form useful for analysis,
the positions of the various diffraction cones need to be determined. This can
be achieved by using either photographic film or a detector sensitive to X-ray
12 Basic crystallography

radiation. In both cases the basic idea is to determine the diffraction angle, 20,
of the various diffraction cones.

The Debye-Scherrer camera

The Debye-Scherrer camera is the simplest technique used for acquiring

powder diffraction data and is shown, schematically, in Fig. 1.18a. A strip of
film is placed inside a cylindrical camera which has at its centre the sample.
There is a trend for the diffraction
The X-ray beam enters through one side of the cylinder and is collimated on
maxima in the powder X-ray
experiment to get weaker as 20 to the powder sample which is normally mounted by coating a glass fibre or
increases. Strong reflections are, filling a narrow glass capillary. The sample is often rotated around the axis
typically, seen up to 20=50° but at of the fibre in order to present as many orientations of the crystallites as
high angles, e.g. above 90°, the possible to the beam. The diffraction cones cut the film at various angles, 20.
reflections can be difficult to
Once the film is developed and laid flat a typical powder diffraction pattern,
observe.
as shown in Fig. 1.18b, is obtained. Each line of the film represents a lattice
spacing in the sample; at low angles the lines have distinct curvature as the
diffraction cone angle is small and a large part of the cone is intersected, at
angles near 90° only a small section of the cone intersects the film and the
line curvature is less marked. The position of the undiffracted beam can be
seen as the dark area surrounding the shadow of the beam stop. From the
radius of the camera and the distance along the film from the straight-through
position, the diffraction angle, 20 can be determined for each of the lines.

Fig. 1.18 The Debye-Scherrer

camera and the form of the
photographic diffraction pattern
obtained.

►
0 26 180
 Inorganic materials chemistry 13

One drawback of the Debye-Scherrer technique is that the defocusing of

the X-ray beam once it has left the collimator results in relatively poorly
resolved data, that is, the diffraction lines obtained on the film are quite
broad. For simple structures which have relatively few lattice planes, and
hence few lines in their diffraction patterns, this is of little consequence as
their positions can still be readily obtained. However, for more complex
structures with large unit cells and low symmetry the large number of
different lattice rf-spacings produce many diffraction cones. In such systems
broad lines may overlap and it becomes difficult to measure their exact
positions.
Other photographic techniques, principally the Guinier camera, use a crystal
monochromator to focus the X-ray beam through the sample on to the film
and this gives rise to much higher quality powder diffraction data.

The powder diffractometer

The powder diffractometer uses an X-ray detector, typically a Geiger-Muller

tube or scintillation detector, to measure the positions of the diffracted beams.
Scanning the detector around the sample along the circumference of a circle
cuts through the diffraction cones at the various diffraction maxima, Fig. 1.19.
The intensity of the X-rays detected as a function of the detector angle, 20,
for a typical material is shown in Fig. 1.20. The reflection geometry
frequently used for diffractometers is useful in that it is to some degree
focusing, leading to reasonably well resolved data. This can be seen in Fig.
1.21, the sample acts like a mirror and provided the geometry of the
equipment is correctly designed the diffracted beam can be focused on to the
detector.

Focusing circle^

Detector
Fig. 1.19 Schematic of a typical
powder diffractometer.
X-ray Tube

v
\
^
A
Measuring circle
14 Basic crystallography

Fig. 1.20 A typical powder pattern

obtained using a diffractometer.

Detector Comparison of the experimental techniques

The photographic methods have the advantage that very little sample is
needed, typically a few milligrams, and the equipment is relatively cheap.
Also, exposure times are reasonably short, generally <1 h, as all the diffracted
X-rays in the different diffraction cones are collected simultaneously by the
film. A disadvantage, in comparison with the diffractometer method, is that
the film requires developing and analysis of the data can be time consuming;
data collected on a diffractometer are readily stored in numerical form ( 20
and counts ) on a computer and analysis of this data in terms of determining
peak positions is readily achieved. The other advantage of the diffractometer
Focusing
circle method is that intensities of the various reflections can be quickly ascertained;
these can be extracted from photographic data using a densitometer but the
Fig. 1.21 Focusing in a powder task is time-consuming and less accurate.
diffractometer.

1.10 Problems

1.1 What are the lattice types of NaCl (Fig. 4.1), U02 (Fig. 6.6), Ti02 (Fig.
4.2) and K2NiF4 (Fig. 4.15).

1.2 Show that a body centred monoclinic lattice can be redrawn as, and is
thus equivalent to, a C centred lattice.

1.3 Calculate the d spacings for 101 and 10-1 planes in an orthorhombic cell
with dimensions a = 3 A (300 pm), b = 5 A (500 pm) and c = 4 A (400 pm).
Carry out the same calculation for the monoclinic cell with the same a, b and
c but with 6=110°. Show diagrammatically, using a projection down c, the
positions of these planes in the monoclinic system and, hence, explain why
for this system dn0 * d10_j.
2 The application and
interpretation of powder
X-ray diffraction data
S

2.1 General

The form of the powder X-ray diffraction data obtained from a material will
depend upon the crystal structure it adopts. This structure is delineated by the
lattice type, crystal class, unit cell parameters and the distribution of the
various ion and molecule types within the unit cell. The number and positions,
in terms of 20, of the reflections depend upon the cell parameters, crystal
class, lattice type and wavelength used to collect the data, while peak intensity
depends upon the types of atoms present and their positions.
As a result of the enormous range of different structures which materials
adopt, nearly all crystalline solids have a unique powder X-ray diffraction
pattern in terms of the positions of the observed reflections and the peak
intensities. In mixtures of compounds each crystalline phase present will
contribute to the powder diffraction pattern its own unique set of lines. The
relative intensity of line sets from mixtures will depend upon the amount
present and the ability of a structure to scatter X-rays.
The effectiveness of powder X-ray diffraction has led to it becoming the
major technique used for the characterisation of polycrystalline, solid
inorganic materials. Table 2.1 summarises some of the applications to which
the technique is put and the most important of these are discussed in more
detail in the following sections.

2.2 Identification of unknowns

Many of the powder diffraction data sets collected from inorganic,

organometallic and organic compounds have been compiled into a data-base
by the Joint Committee on Powder Diffraction Standards (JCPDS). This data
base which contains over 50 000 unique powder X-ray diffraction patterns
may be used to identify an unknown material from its powder pattern alone.
With simple materials this can be done by comparing two or three intense
lines from the powder diffraction pattern with listings of the most intense
lines in book form. For more complex materials a computer can be employed
to look for coincidences between the experimental data and the data-base. For
many solid materials this technique provides a rapid method of identification.
Drawbacks are that the compound must have previously been made and its
pattern entered in the data-base and that the compound must be crystalline.
16 Application and interpretation of powder X-ray diffraction data

Table 2.1 Applications of powder X-ray diffraction

Identification of unknown materials

Determination of sample purity

Determination and refinement of lattice parameters

Investigation of phase diagrams / new materials

Determination of crystallite size / stress

Structure refinement

Phase changes / expansion coefficients

2.3 Phase purity

In a mixture of compounds each crystalline phase present will contribute to

the overall powder X-ray diffraction pattern. In preparative materials
chemistry this may be used to identify the level of reaction and purity of the
product. The reaction between the two solids A1203 and MgO to form
MgAl204 may be monitored by powder X-ray diffraction. Fig. 2.1. At the start
of the reaction a mixture of A1203 and MgO will produce an X-ray pattern

Fig. 2.1 The reaction between

Al203 and MgO followed by
powder X-ray diffraction.
 Inorganic materials chemistry 17

(combining those of the pure phases. As the reaction proceeds, Fig. 2.1
(patterns (a) and (b), a new set of reflections corresponding to the product,
MgAl204, emerges and grows in intensity at the expense of the reflections
:from A1203 and MgO. On completion of the reaction the powder diffraction
ipattem will be that of pure MgAl204. A materials chemist will often use
.powder X-ray diffraction to monitor the progress of a reaction in this way.
The powder X-ray diffraction method is widely employed to identify
impurities in materials whether it be residual reactant in a product, as above,
or an undesired by-product. However, the impurity must be crystalline, even
a high level of amorphous phase would not be detectable except maybe as an
increase in the background of the powder pattern. Also, the ability to detect
a second, impurity phase will depend on how well the two components scatter
X-rays though normally it is possible to detect a few per cent of an impurity
phase. As X-rays are scattered by electrons, materials containing heavy Amorphous materials have no long
elements, e.g. second and third row transition metals, diffract X-rays strongly range order but short range
and will produce, in general, strong reflections in the powder pattern. correlations produce broad
diffraction maxima around positions
Materials containing light elements, e.g. organic compounds and first and
corresponding to the interatomic
second row oxides, will produce weaker patterns. As an example consider the
distances.
reaction of Li20 and Nb2Os to form LiNb03. It would be easy to identify a
few per cent of Nb2Os as an impurity in LiNb03 but Li20 would need to be
present at a much higher level for its pattern to be observed in the powder
diffraction data from a mixture of Li20, Nb2Os and LiNb03.

2.4 Determination and refinement of lattice parameters

If Miller indices can be assigned to the various reflections in the powder

pattern it becomes possible to determine the cell constants; these are directly
related to 20 and h,k,l through the relationships derived from the combination
of Bragg’s Law and the ^-spacing expression and summarised in Table 1.3.
This assignment may be readily achieved for cubic crystal systems with a
simple relationship between the diffraction angles and lattice parameters.
However, as the cell symmetry decreases, the number of lines in the
diffraction pattern increases rapidly as the relationship between 20 and the
cell constants becomes more complex. In these cases indexing the powder
diffraction data becomes a computational problem.

2.5 indexing powder diffraction patterns from cubic

systems

The relationship between diffraction angle and the lattice parameter is

obtained by combining the Bragg equation with the expression for d-spacing
in terms of lattice parameter and Miller indices. In the cubic crystal system
the following expression was obtained in Section 1.8:
18 Application and interpretation of powder X-ray diffraction data

sin20 = —— { h2 + k2 + l2 } (2-1)
4a2
which can be written as
sin20 = A { h2 + k2 + l2 } @.2)

as both a and X are constants. Sin20 values for the various reflections can be
readily obtained from the diffraction angle 20. These will all be related to
each other through a common multiplier A as h, k and / can only take integer
values. Note that this will only be true for cubic systems; the more complex
expressions for the d-spacings in terms of the lattice parameters in the other
crystal systems means that more than one multiplier will be required. This
argument may also be reversed in that if all the reflections in the powder
diffraction pattern can be related through their sin20 values by a single,
common multiplier then the crystal system must be cubic.
The process involved in indexing data from a cubic system and obtaining
the lattice parameter is, therefore, straightforward and is illustrated in Table
2.2. Once sin20 values are tabulated they can be inspected for the common
multiplier A; the ratio for each peak is the sum of the squares of the Miller
indices.

Table 2.2 Indexing of the powder diffraction data from a cubic material

20 sin20 Ratio Miller indices

22.983 0.03969 1.00 100

32.729 0.07938 2.00 110

40.372 0.11907 3.00 111

46.962 0.15876 4.00 200

52.908 0.19845 5.00 210

58.418 0.23814 6.00 211

68.595 0.31752 8.00 220

Note that there are some values which h2+k2+l2 cannot take. For instance
it is impossible to chose the integer values of the Miller indices such that the
sum of their squares equals 7 or 15. This means that there can be no peak in
the powder diffraction pattern corresponding to 7A or 15A.
Once the data have been indexed, derivation of the lattice parameter is
readily undertaken. Any peak may be chosen and the Miller indices,
wavelength and sin20 value for that peak substituted back into eqn 2.1 to give
a. For example, from Table 2.2 where the data were collected with
X = 154 pm

0.31752 = 1542.8/ 4a2

a = 386.5 pm (3.865 A)
 Inorganic materials chemistry 19

In general, if all lines have been measured with the same precision of 20
'then the use of a high angle line will produce a more accurate lattice
i parameter. In practice, a least squares refinement technique using all the
reflection positions is used by powder diffractionists to obtain the best value
for the lattice parameter.
■

> 2.6 Tetragonal and hexagonal systems

Combination of the expression for the d-spacings in a tetragonal unit cell with
the Bragg equation and replacement of constant terms by A and C gives the
following expressions

tetragonal sin20 = A ( h2 + k2 ) + Cl2 (2.3)

hexagonal sin20 = A(h2+hk+k2) + Cl2 (2.4)

where A and C incorporate the wavelength and the lattice parameters. For
these crystal systems sin20 values obtained from the powder diffraction will
be related through two multipliers. By consideration of the possible values Table 2.3 Sin20 relationships in the
which the Miller indices can take it is feasible to derive the relationships tetragonal system, h,k,l < 2
A
between the various sin20 values. This is done for a tetragonal system in Table
C
2.3. 2A
Obviously, assigning these ratios to particular peaks can be quite difficult. A+C
However, the method can be illustrated for a tetragonal system by 2A+C
consideration of a system where there is only a slight distortion from a cubic 4A
4C
unit cell; that is, the lattice parameters a and c are very similar. In this case
4A+C
the relationship between the reflections seen in the powder diffraction pattern A+4C
of a tetragonal material to those of a cubic phase is shown schematically in 5A
Fig. 2.2. The 100 reflection of a cubic material is made up of diffraction from 5A+C
4A+4C
the (100), (010), (001), (-100), (0-10) and (00-1) lattice planes which all
have an identical ^-spacing. The number of planes which contribute towards
a reflection is termed the multiplicity, in this case the multiplicity is 6. The
sin20 values for all these reflections are identical as they depend only on
(h2+k1+i1). In a tetragonal material with slightly different a and c lattice
parameters the constants A and C will be very similar. Hence, the sin20 values
for the (100) and (001) reflections will diverge only marginally and these
maxima will be observed at only slightly different 20 values in the powder
diffraction pattern. In Fig. 2.2 the peak doublet near 22° can be assigned as
deriving from the 100 reflection of the cubic material and these two
reflections indexed as 001 and 100. Lattice planes contributing to the (100)
peak will be the (100), (010), (-100) and (0-10) giving a multiplicity of 4,
those contributing to the (001) reflection will be the (001) and (00-1) giving
a multiplicity of 2. The multiplicity is reflected in the peak intensities which
have a 2:1 ratio, Fig. 2.2.
In this slightly distorted material, assignment of Miller indices to all the
other reflections has now become easy. Once the 100 and 001 reflections
20 Application and interpretation of powder X-ray diffraction data

Fig. 2.2 Comparison of the powder

X-ray diffraction patterns of cubic
and slightly tetragonally distorted
BaTiOj.

have been indexed, values of the constants A and C can be obtained using
eqn. 2.3 and the other observed sin20 values assigned using the possible ratios ;
shown in Table 2.3. Again, following full indexing of the data, the values of
A and C may be used to calculate the a and c parameters. For the compound
in Fig. 2.2, BaTi03, the values obtained are 399.5 pm (3.995 A) and 403.4
pm (4.034 A). BaTi03 is discussed more fully in Chapter 4.

2.7 Lattice type and systematic absences

In indexing the powder patterns it has been assumed that all the possible
reflections are observed, that is, scattering from each of the different lattice
planes is sufficiently intense to contribute to the diffraction profile. This is
normally so for a primitive lattice, but for body centred and face centred
lattices restrictions occur on the values that h, k and 1 may take if the
reflections are to have any intensity. This results in certain reflections not
being observed in the powder diffraction pattern and these are known as
systematic absences. The origin of these absences can be illustrated with
regard to Fig. 2.3. Consideration of diffraction from the 100 plane of a cubic
material with a body centred lattice shows that X-rays diffracted at the Bragg
angle from the 100 planes (the faces of the cube) will be, by definition, in
phase. However, halfway between the 100 planes, as a result of the body
centring, there is an identical plane of atoms shifted in the x and y directions
by (Vi,V4). X-Ray diffraction from this plane, the 200, will have a pathlength
difference relative to that from the 100 planes exactly X/2 out of
phase and will, therefore, destructively interfere with it. In a body centred
 Inorganic materials chemistry 21

Fig. 2.3 Origin of 100 absence in a

body centred lattice. The scattering
from the lattice point at the cell
centre is 180° out of phase with
that from the cell corners.

attice there is one lattice point in the 200 plane for every lattice point in the
100 plane and overall there will be total destructive interference for the 100
■eflection. Note that it does not matter where the atoms or ions actually are
in the unit cell, it is the fact that they always occur in pairs (at (x,y,z) and at
[x+V2, y+V2, Z+1/2)) generated by the body centring that causes the destructive
interference.
Extension of these considerations to other possible values of h, k and / in
a body centred structure show that for a reflection to be present then h+k+l
is even. Alternatively this may be expressed as the condition, h+k+l=2n+l is
:he systematic absence for I lattices. Similar considerations for a face centred
;ubic lattice show the following restrictions on h, k and l for a reflection to
be observed: h+k=2n, k+l=2n and h+l=2n. This restriction may also be
expressed in the terms, h, k and / must all be odd or all even. Table 2.4
summarises which reflections are observed for the various lattice types in
;ubic systems.
Using this information it is possible, once a data set has been indexed, to
determine the lattice type from the systematic absences. The existence of these
absences can be a complicating factor when attempting to index cubic
systems. If a material adopts a face centred cubic lattice then the common
ratio between the sin20 values may not be immediately obvious. For this F
lattice the first few peaks, if all are observed, will have sin20 values in the
ratio 1.00 : 1.333 : 2.666 : 3.666 : 4.00. However, there is still a single
bommon multiplier giving the integral ratios 3:4:8:11:12 for the first few
Reflections corresponding to the (111), (200), (220), (311) and (222)
Reflections.
In a body centred cubic system the systematic absences can also be mis¬
leading in terms of indexing. The sin20 values of the first few lines will be
22 Application and interpretation of powder X-ray diffraction data

Table 2.4 Reflection conditions imposed by lattice type

Lattice type

Miller indices Primitive, P Body centred, 1 Face centred, F

no systematic h+k+l=2n h,k,l all odd or
absence all even

1 0 0 ✓ X X

1 1 0 ✓ ✓ X

1 1 1 ✓ X /

2 0 0 / ✓ ✓

2 1 0 ✓ X X

2 1 1 ✓ ✓ X

2 2 0 ✓ ✓ /

221/300 / X X

3 1 0 ✓ ✓ X

3 1 1 / X ✓

2 2 2 ✓ ✓ ✓

3 2 0 ✓ X X

3 2 1 ✓ ✓ X

4 0 0 ✓ ✓ ✓

found to have the ratios 1:2:3:4:5:6:7:8 corresponding to the ratios of the suma
of the squares of the Miller indices. The presence of a line with a ratio ofi
seven, points to these lines originating from a body centred system as there
are no values of h, k and l which square and sum together to give seven. Thei
common multiplier must be adjusted, by being halved, for these lines to give
the ratios 2:4:6:8:10:12:14:16 corresponding to the Miller indices (110), (200)(
(211), etc. These are the first reflections found for a material adopting a body
centred cubic structure.
Other symmetry elements in crystal structures can also give rise ta
absences in the powder diffraction pattern and sometimes peaks are noc
observed experimentally, as their intensity is so low that they cannot bet’
distinguished from the background noise. One case where peaks have such z
low intensity as to give rise to apparent absences in the powder diffraction]
pattern can be illustrated by consideration of KC1. X-Rays are scattered by
electrons and isoelectronic species such as K+ and Cl“ will scatter X-rays to
almost identical degrees. Therefore, in the face centred cubic sodium chloride!
structure adopted by KC1, Fig. 2.4, the ion positions all scatter X-rays to an
almost equivalent degree, that is, the potassium and chloride ions are:
practically indistinguishable to X-rays. Hence, the lattice repeat observed by \
 Inorganic materials chemistry 23
i

G> K
:
Fig. 2.4 The structure of KCI
9 ci
y showing the primitive unit cell of
isoelectronic species.

X-rays is much smaller than the true one and is the smaller cube outlined in
Pig. 2.4. This unit cell is, if the corner atoms are considered identical as they
ire to X-rays, primitive and the powder X-ray diffraction pattern of KCI can
ie indexed using a primitive cubic unit cell of dimensions half that of the true
face centred cubic unit cell. In essence the scattering equivalence of K+ and
CT has made the reflections with odd Miller indices from the face centred
;ubic structure so weak that they are not observed. It is then possible to re-
index the peaks with even Miller indices (by halving them) using a smaller
:ommon ratio which corresponds to the smaller (half size) unit cell, Fig. 2.5.
The fact that the powder pattern of KCI, and other similar compounds with
isoelectronic species, for example MgO, can be indexed using a primitive unit
-ell is merely a consequence of the origin of the X-ray scattering which
makes such species indistinguishable.

(200)

Fig. 2.5 Powder X-ray diffraction

pattern from KCI. Indices in
parentheses refer to the true unit
cell while those directly above the
peaks are from the 'primitive'
subcell. The position of the zero
intensity (111) reflection from the
true unit cell is also marked.
24 Application and interpretation of powder X-ray diffraction data

2.8 Peak intensities

The peak intensities of the various reflections in an X-ray diffraction

experiment are controlled by the atom types and their distribution in the unit
cell. Measurement of these intensities allows the crystallographer to work out
the structure of the crystalline material which produced them. This structure
determination is, normally, carried out using the single crystal X-ray
diffraction technique. In a single crystal experiment the diffraction data is
collected using a detector or film covering two dimensions and the orientation
of the crystal being studied can also be adjusted. Using this technique the
intensities of all the individual reflections may be obtained and employed to
determine the full crystal structure.
In the powder X-ray diffraction experiment, because all the reflections
occur along the single 20 axis, reflections frequently overlap, particularly at
high 20 values. Extracting intensities for the individual reflections, essential!
for the structure determination, becomes very difficult. Structural information
may still be obtained from powder diffraction patterns using a technique
known as the Rietveld method. In this method, a trial structure (in which!
likely positions, in the unit cell, are assigned to the different atoms present):
is used to calculate intensities for the various reflections. These intensities:
may then be combined with the various factors controlling the form of the
powder pattern, e.g. lattice parameters, radiation wavelength, to generate a
'calculated' powder pattern based on the trial structure. Comparison of this
profile with the experimental one can then be undertaken and adjustment oh
the trial structure parameters carried out in order to obtain the best fit between:
the experimental and calculated data. This then provides reasonably accurate
information on the atom positions in the unit cell. However, as stated above,:
a trial structure is required before the technique can be employed.

2.9 Other applications of powder X-ray diffraction

Crystallite size
./
/ \V <a>
In order to observe sharp diffraction maxima in the powder X-ray diffraction
pattern, the crystallites need to be of sufficient size to ensure that slightly
away from the 20 maximum, destructive interference occurs. The requiremem
<e)
is for a sufficiently large number of planes that summation of the diffractec
40 41 42 43 44 45 waves, which are only slightly out of phase between successive planes-
29 eventually produces destructive interference. In materials with very smal
crystallites, with few planes, the diffracted X-ray intensity slightly away from
Fig. 2.6 The effect of crystallite
the Bragg angle does not totally destructively interfere and this leads tc
size on peak width in powder X-ray
diffraction pattern: (a) instrumental broadening of the reflections, Fig. 2.6.
broadening, (b) 1 pm particles, (c) The Scherrer formula relates the thickness of the crystallites to the breadth
100 nm, (d) 10 nm, (e) 5 nm. of peaks in the powder X-ray diffraction pattern:
 Inorganic materials chemistry 25

0.9A.
t (2.5)
/flis
where t is the crystallite thickness, A the X-ray wavelength, 0 the Bragg angle
[half the measured diffraction angle) and BM and Bs are the width in radians
of diffraction peaks of the sample and a standard at half height. The standard
peak width is obtained from a highly crystalline sample (with a diffraction
peak at a similar diffraction angle to that of the sample under investigation)
and represents the instrumental broadening.
The Scherrer method of determining crystallite size from powder data is of
wide application. For materials which are of plate-like habit, for example
days, the size of the crystallites in the different directions may be ascertained
by consideration of the peak widths of the various (hOO), (OkO) and (00/)
reflections.

Variable temperature studies

Combination of powder X-ray diffraction equipment with a furnace or cryostat

allows powder diffraction patterns to be collected over a large temperature
range. Studying the evolution of the pattern as a function of temperature
allows rapid identification of phase changes, which cause marked changes in
the pattern as the material adopts a new structure, and the determination of
thermal expansion coefficients. For the latter, the peak positions are monitored
as a function of temperature, allowing determination of the lattice parameters
and unit cell volume.

2.10 Problems

2.1 The following powder X-ray diffraction data (A=154 pm (1.54 A)) were
obtained from zeolite 4A:

20 (°) 7.144 10.121 12.405 14.302 16.005 17.563 20.300

Index the data and determine the lattice parameter and type. The structure of
zeolite 4A is shown in Fig. 7.4a; estimate the size of the main channel.

2.2 Plutonium metal undergoes a phase transition when heated to 750 K

changing from a face centred cubic structure with a = 463.7 pm (4.637 A) to
a body centred cubic structure with a = 363.8 pm (3.638 A). Determine the
peak positions expected in the powder X-ray diffraction pattern (A, =154 pm)
of plutonium metal above and below 750 K in the 20 range 0-75°. How
would the expansion of plutonium metal on heating above 750K affect the
form of the diffraction pattern?

2.3 Determine the expected appearance of the 310 reflection of cubic BaTi03
once the compound has transformed to the tetragonal phase.
3 The synthesis of inorganic
materials

The synthesis methods used in the preparation of inorganic materials, many

of which have extended lattices rather than discrete molecules, are quite
different from those used by organic, organometallic and coordination
chemists. Rather than altering a single functional group or ligand attached tc
a molecule, the materials chemist works with complete lattices, either building
or modifying them. The techniques used to do this and their application tc
inorganic materials synthesis are discussed in this chapter.

3.1 High temperature reactions

The most widely used method for the synthesis of inorganic materials follows
an almost universal route that involves heating the components together at
high temperature over an extended period. Many of the inorganic materials
discussed in this book are complex oxides, i.e. materials containing more thar
one metal in addition to oxygen. These compounds include ternary phases
such as BaTi03, and quaternary oxides, e.g. YBa2Cu307.*. In such cases the
direct high temperature reaction of the component oxides frequently yields the
desired complex oxide phase. This direct method of preparation is, however
equally applicable to other inorganic material types, for example the following
syntheses

Chlorides 3CsCl + 2ScC13 —> Cs3Sc2Cl9 (3.1:

Aluminosilicates NaA102 + Si02 —» NaAlSi04 (3.2;:

3.2 The process of the reaction between solids

In order to illustrate a typical synthesis in solid state chemistry the preparation

of the ternary oxide, SrTi03, will be used as an example. Simple binary
oxides are generally available commercially as polycrystalline powders with
typical particle dimensions of a few micrometres. In the synthesis of SrTi03
SrO and Ti02 would be ground together in the correct stoichiometric
proportions (1:1 molar ratio) using a pestle and mortar. A representative
reaction would then involve pressing the mixture into a pellet, transferring i;
to a crucible and placing this in a furnace at 900°C. The crucible is normally
constructed of an inert material such as vitreous silica, recrystallised alumins
or platinum.
 Inorganic materials chemistry 27

A schematic view of the reaction mixture, magnified many times, is

epresented in Fig 3.1a and consists of particles of the constituent oxides,
(eparated mainly by voids, but with some contact at particular crystal faces,
i’eaction between the binary oxides occurs by migration of ions between the
iure oxides across these interfaces and formation of the new structure of the
arget ternary oxide. A representation of a partially reacted mixture, derived
irom that in Fig. 3.1a is shown in Fig. 3.1b. In reactions between oxides the
lost mobile ions are normally the cations, as these are usually smaller than
ib oxide ions, and it is these ions that diffuse through the oxide and between
le reactant blocks.

a b

ig. 3.1 Schematic of reacting particles in a solid state reaction.

The need for the high temperatures required in the synthesis of mixed
xides can be rationalised by consideration of the energetics of the reaction
rocess. In comparison with organic and organometallic chemistry where a
pical reaction will involve the breaking of one bond and the formation
aother, sometimes simultaneously, the solid state reaction involves a
implete disruption of the structure. The cations in an oxide material are
ormally coordinated to a large number of oxide ions ranging from 4 for the
nallest ions, e.g. Li+, to 12, or more for ions such as barium. In order for
lese ions to migrate to the interfaces/grain boundaries and form a new
ructure, considerable energy is required to overcome the lattice energy,
lence, the typical reaction temperatures of oxide chemistry in the range
30-2000°C. At the lowest temperatures, ions with a low charge to radius
itio, for example Cs+ and other alkali metals, which have the weakest
teraction with the lattice, will undergo reaction. So, for example, the
action
Na202 + 2CuO —» 2NaCu02 (3.3)
28 The synthesis of inorganic materials

proceeds reasonably rapidly at 400°C. Small, strongly charged cations, whicl

interact more strongly with the oxide lattice, require higher temperatures t(
produce significant reaction. For example the reaction

MgO + A1203 -> MgAl204 (3.4!

would normally be carried out at temperatures above 1400°C. Increasing th*

temperature of a solid state reaction will speed up the process as the rates o
diffusion of the various ions increase. The limit of this behaviour is seei
when the mixture of solids melts and the ions become highly mobile
Occasionally the reaction to form a new complex oxide may be carried ou
in the melt; however, as discussed in Section 3.6, crystallisation of a moltei
reaction mixture will often not give the desired product.
Besides raising the temperature, other methods can be used to enhance th-!
reaction rate between solids. Consideration of the schematic reaction mixtur
in Fig. 3.1a helps illustrate these. To improve the reaction rate, the numbe
of interfaces at which the reaction can occur needs to be maximised. Thij
may be achieved by both reducing the particle size and pelletising thl
material. Thorough grinding of the metal oxide mixture will produce
homogenous distribution of the two oxides ensuring that each oxide particl
is surrounded, as far as possible, by oxide particles of the other type
Pelletising the material, normally carried out using a hydraulic press, reduce
the inter-particle void space (typically to less than 10%) producing many mor
interfaces over which the reaction can take place.
After the reaction has proceeded for some time a volume of the produa
builds up along the interface between the reacting crystallites. Furthe
formation of the ternary phase requires the cations to migrate over longe;
distances, though this distance is minimised by reducing the particle size i
the initial grinding. To facilitate further reaction the partially reacted powde
is normally reground at intervals, this has the advantage of introducin'
pristine reactant crystal faces to each other.

3.3 Precursor, solution and gel methods

The reduction in the particle size by grinding the reaction mixture, eves
mechanically, can only achieve a limiting level of grain diameter, at bes:
about 0.1 pm. However, chemical methods can be used to effectively reduc;
particle size still further and so enhance reaction rates at a particula
temperature. One valuable method is the use of precursors that decompose s
moderate temperatures to generate the reactant oxides; such precursors includs
metal carbonates, hydroxides and nitrates. When the crystallites of thes
precursors disintegrate with the loss of gaseous species they break down t
yield fine, very reactive particles resulting in faster reaction times. At
additional advantage of precursors is that they are normally air stable whil
many oxides are hygroscopic and/or pick up carbon dioxide from the air. *
commercial sample of BaO normally contains Ba(OH)2 and BaC03 and wii
continue to react with water/C02 in the air whilst being weighed out in th
 Inorganic materials chemistry 29

laboratory. This makes the stoichiometry of a reaction involving BaO difficult

to control. BaC03, however, is perfectly air stable, and is frequently employed
for reactions involving barium, though it needs to be heated strongly to ensure
complete decomposition to the oxide.
In order to produce very homogeneous mixtures the solid state chemist can
employ solutions where the reactant ions are dispersed on the atomic scale.
Precipitation from mixed ion solutions will produce a solid containing the
required ions, though care must be taken to ensure that the precipitate is of
’"the correct stoichiometry. A typical reaction would be

Zn2+(aq) + 2Fe2+ (aq) + 3C2042-(aq) -> ZnFe2(C204)3 i (3.5)

where the highly insoluble oxalates are precipitated from a 1:2 molar aqueous
solution of zinc and iron. Heating the mixed metal oxalate product in air
above 800°C results in its decomposition to yield the ternary oxide, ZnFe204,
in just a few hours.
Precipitation methods have considerable advantages, it is, however, difficult
to ensure that the precipitate has the desired stoichiometry. In the preparation
of complex oxides, such as YBa2Cu307, the correct conditions, for example
pH, to precipitate a 1:2:3 Y:Ba:Cu complex from solution are difficult to
determine and one ion may be left, to some degree, in solution leading to an
impure product. This may be overcome using gel methods. In this technique
rather than precipitating and filtering to produce a solid, the whole solution
is solidified through complexation and/or removal of water producing a gel.
Such techniques have been widely used for the synthesis of ceramics
including the high temperature superconductor, YBa2Cu307. Dissolution of
yttrium, barium and copper as their nitrate salts in water is followed by
addition of citric acid and ethanediol. Heating of this mixture and evaporation
of water results in the formation of a gel that may then be heated further in
a crucible to form the complex oxide. The final stage, due to the mixing of
various ions at the atomic level, can be carried out at temperatures as low as
700°C and requires only a few hours. This should be contrasted with a direct
reaction of oxides which typically requires temperatures above 900°C for
several days.

3.4 Sealed tubes and special atmospheres

The direct reaction discussed above is widely used for the synthesis of
complex oxides, fluorides, chlorides, phosphates, silicates and sulphides, and
is generally carried out in air. However, the reaction environment may need
to be controlled if a particular oxidation state or reaction stoichiometry is
targeted. In such cases the reactions are carried out in a controlled
atmosphere, using a tube furnace, where a particular gas can be passed over
the reaction mixture during heating. An example of such a reaction would be
the use of an inert gas to prevent oxidation, as in the preparation of the
thallium® compound TlTa03
30 The synthesis of inorganic materials

N2 / 600°C
T120 + Ta205 2TlTa03 (3.6)

Another example would be the reduction in hydrogen of V205 to V02

h2 / iooo°c
V205 -> V02 (3.7)

High gas pressures may also be used to control the reaction product
stoichiometry. For example the use of oxygen under high pressures, several
hundred atmospheres, permits the formation of Sr2Fe04 (Fe(IV)) from
mixtures of SrO and Fe203. This should be contrasted with the formation of
compounds containing Fe(III) under ambient pressures.
Another problem encountered in direct synthesis is the volatility of one or
more components of the mixture at the high temperatures required for
reaction. In such cases the reaction mixture is normally sealed in a glass tube,
under vacuum, prior to heating. Examples of such reactions are

500°C
Ta + S2 —^ TaS2 (3.8)

860°C
T1203 + 2BaO + 3CaO + 4CuO —> Tl2Ba2Ca3Cu40I2 (3.9)

where the sulphur and thallium(IH) oxide, respectively, are volatile at the
reaction temperature and would be lost from the reaction mixture in an open
vessel, leading to impure products.

3.5 Solution and hydrothermal methods

Many non-oxide inorganic materials may be synthesised by crystallisation

from solution. Whilst the methods are very diverse, the range can be
illustrated by the following;

3KF(aq) + MnCl2(aq) -+ KMnF3i + 2KCl(aq) (3.10)

Zr02 + 2H3P04 - Zr(HP04)2.H20 + H20 (3.11)

90°C

12NaA102 + 12Na2Si03.9H20 -+ Na12[Si12Al12048].nH20 (Zeolite 4A)

+ 24NaOH (3.12)

The application of solution methods is extended by using hydrothermal

techniques where the reacting solution is heated in a sealed vessel above its
Fig. 3.2 The tetrapropyl-ammonium boiling point. Such reactions are important in the synthesis of open structure
ion directing the channel structure aluminosilicates (zeolites). The preparation of zeolites typically involves
of ZSM-5. crystallisation from a strongly basic aqueous solution containing the
 Inorganic materials chemistry 31

tetrahedral structure building ions ( NaA102 and Si02) plus a templating ion
that is frequently an organic base. The shape of the templating ion directs the
crystallisation of the aluminate and silicate tetrahedra and determines the
structure of the zeolite product. The crystallisation process is frequently slow
and hydrothermal methods act to speed up the process. An example of zeolite
preparation is that of ZSM-5, a material discussed in more detail in Chapter
7. A mixture of silicic acid ( Si02.nH20), sodium hydroxide, aluminium
sulphate, water, n-propylamine and tetrapropylammonium bromide is heated
fn a hydrothermal bomb at 160°C for several days. The organic template ion
directs the formation of the structure as shown in Fig. 3.2 with the alkyl
groups filling the channels. The template ion may be removed from the cavity
by heating the product in air, oxidising the organic cation and leaving intact
the aluminosilicate framework.

3.6 Phase diagrams and synthesis

The direct reaction of components to form inorganic materials is generally

carried out below the melting point of the system despite the process being
extremely slow. Ideally the materials chemist would like to use molten phases
where reaction rates are much faster due to the more rapid ion diffusion.
However, cooling a molten mixture does not normally yield the desired
product in a pure form and to understand this behaviour a knowledge of phase
diagrams is required.
Phase diagrams are very useful to the solid state chemist; they illustrate the
behaviour of multi-component systems as a function of temperature by
delineating the ranges of stability of various phases in terms of the
stoichiometry of the mixture and temperature.
A simple two component system phase diagram is shown in Fig. 3.3 with Fig. 3.3 Simple binary eutectic
the x axis denoting the relative proportions of two components A and B and system.
the y axis the temperature. A and B do not react together, i.e. there is no
compound A^ in the system; an example of this behaviour would be CuO
and CdO. The diagram has four regions. At low temperatures both A and B
are solids and a mixture of the two consists of discrete particles of the two
solids. At higher temperatures A and B melt and, depending on the
stoichiometry of the mixture temperature/composition, regions occur where
solid A exists with a liquid, solid B exists with a liquid or the whole system
is molten. The composition/temperature point E, i.e. the stoichiometry which
is totally liquid at the lowest temperature, is known as the eutectic point and
this composition is the eutectic composition. The curve separating totally
liquid from solid+liquid regions is known as the liquidus.
The above case is of little interest to the synthetic chemist as no compound
is formed in the system. However, the simple eutectic system behaviour is
found in more complex phase diagrams such as that shown in Fig. 3.4. In this
case the components A and B react together and form a compound of
stoichiometry AB with a different structure. This new compound is
represented in the figure by a vertical line. A feature of this type of phase
32 The synthesis of inorganic materials

diagram is that if the compound AB is heated then on reaching the :

temperature Tx the compound melts to produce a liquid of the same
composition. This is. known as congruent melting. For this type of phase
diagram the overall behaviour is that of two simple, binary, eutectic phase
diagrams joined side to side. Hence, the left-hand side of the diagram is a i
binary eutectic between A and AB and the right-hand side one between AB
and B. In a binary system that adopts this type of phase diagram the synthesis ;
of AB from the components A and B can be undertaken very easily. A 1:1
stoichiometric mixture of A and B is melted; on cooling the melt, if fully
homogeneous, will crystallise to give pure AB.

3.4 3.5

Figs. 3.4 and 3.5 Two component phase diagrams forming a phase AB with congruent
and incongruent melting respectively.

The other type of phase diagram that occurs for a two component system
forming a compound AB is shown in Fig. 3.5, where the behaviour of the
compound AB on heating is somewhat different. In this case when the
compound AB reaches the temperature T2 it partially melts to yield some
solid A plus a liquid that will be richer in B. On further heating the solid A
dissolves in the liquid until at temperature T3 the system is completely molten
and the liquid is of the composition AB. This behaviour, where AB does not
melt directly to give a liquid of the same composition, but deposits one
component, generating a liquid richer in the other is known as incongruent
melting and has profound effects regarding synthesis. In this system if a
molten 1:1 mixture of A and B is cooled to temperature Tv solid A will
crystallise out and the liquid will get richer in B, its composition following
the liquidus curve. On reaching the temperature T2 the compound AB is now
stable and the reaction
 Inorganic materials chemistry 33

A(s) + B(l) —» AB(s) (3.13)

can occur. However, this reaction involves a solid and, as has been described
above, will be very slow, particularly so as solid A will have settled out on
the bottom of the reaction vessel. In addition as the particles of A start to
react back with the liquid, to form solid AB, the product will coat the solid
A inhibiting further reaction. If the system is cooled extremely slowly,
typically over many days then the reaction between solid A and liquid B will
6ccur and pure AB can be obtained. However, faster cooling causes the liquid
rich in B to solidify before the reaction with solid A is complete yielding a
mixture of AB and B plus some solid A which crystallised out initially.
Hence, for this type of phase diagram, with incongruent melting, heating
the system above T2, known as the peritectic temperature, the maximum
temperature at which the compound AB is stable, is unhelpful. In such
systems the preparation of AB is normally carried out below T2 where AB is
stable but both A and B are solid, that is, the typical, direct reaction of solids
discussed above in Sections 3.1 and 3.2.

Figs. 3.6 and 3.7 The MgO / Si02 phase diagram and the three component system
Cu0-Ba0-Y203 at 900°C respectively.

More complex binary phase diagrams with several compounds of differing

stoichiometry, for example Fig. 3.6, are constructed from regions with
congruently and incongruently melting phases. In this figure Mg2Si04 melts
congruently whilst MgSiOj melts incongruently. An understanding of the
simpler binary phase diagrams discussed above is all that is required to
interpret this and even more complex looking diagrams.
In systems with more than two components the complexity of the phase
diagram increases. For three components triangular diagrams are used to
34 The synthesis of inorganic materials

represent the phases occurring and the temperature axis is not shown, for
example, Fig. 3.7. The information shown on the diagram represents a single
temperature-, as this changes then the phases that are stable and shown on the
diagram may alter. Each line across the diagram between phases would
represent a binary phase diagram of the type discussed above.

3.7 Low temperature methods

The synthesis methods described above are applicable to most solid state
reactions where a new structure is built from the reactants. This generation of
a new ion or bonding arrangement requires considerable energy, demanding
the high reaction temperatures common in solid state chemistry. However,
some reactions of solids can be carried out at much lower temperatures, at or
just above room temperature, and involve modification of a material’s
structure. These reactions include intercalation or insertion where an ion or
molecule is added to a compound leaving the basic structure of the solid
unchanged and ion-exchange when one ion within a material is replaced by
another.

Strongly bonded s a
layer

Fig. 3.8 Schematic of an

intercalation reaction.
® ft
Weak interaction
between layers

& ®
iPl

Intercalation in layer structures

Many materials that undergo intercalation reactions have layer structures of

the type shown schematically in Fig. 3.8. Such compounds have strongly
bonded two dimensional layers which only weakly interact with the other
layers above and below, typically through van der Waals type forces. The
simplest compound of this type is graphite shown in Fig. 3.9 in which
hexagonal rings of carbon form infinite layers separated by 350 pm. Adjacent
layers are displaced relative to each other in the layer direction so that the
repeat distance along the c direction, perpendicular to the layers, is twice the
layer separation. The weak van der Waals interactions between the layers
mean that ions and molecules may be readily incorporated or intercalated
between the layers. For example, the reaction

8C(s) +K(1) -> C8K (3.14)

 Inorganic materials chemistry 35

Fig. 3.9 The layer structure of

graphite.

Table 3.1 Some graphitic intercalation compounds

Compound Intercalate stoichiometry

Br2 C8Br, C28Br

h2so4 C96HS04, C^HSO,, C24HS04

K C8K, C24K, CggK, C^K

FeClg C12F6CI3, C7F6CI3

OsF6 C8OsF 6

takes place above the melting point of potassium, 64°C, with the driving force
for the reaction being electron transfer from the potassium to the graphite
layers. A better description of the product is Cg~K+. The potassium ions
occupy sites between the graphite layers as shown in Fig. 3.10. As the
carbon-carbon bonds in the layers remain intact throughout the reaction, the
energy required for this process is much less than typical values for reactions
involving solids, hence the low reaction temperature. Potassium is just one
example of an ion or molecule which can be inserted into graphite and a
selection of species is given in Table 3.1.
Other materials that adopt crystal structures consisting of weakly interacting
layers undergo similar reactions to those of graphite. Examples include
transition metal sulphides, such as TaS2 and VPS3, and FeOCl. The transition
metal disulphides have been found to intercalate some large and complex
species; examples include long chain alkyl amines C„H2n+1NH2 up to n = 18
and organometallics such as cobaltocene. The structure and properties of these
layer compound intercalates are discussed in more detail in Chapter 7.

Insertion compounds of metal oxides Fig. 3.10 The structure of C K.

8

Insertion reactions involving large molecules normally require a layer type

structure in order for sufficient room to be generated, by increased separation
of the layers, to accommodate the guest species. Very small ions, e.g. Li+,
36 The synthesis of inorganic materials

may be inserted into three-dimensional structures, such as oxides, provided

there is sufficient room. In the case of alkali metal insertion into transition
metal oxides this process occurs with reduction of the transition metal. One
example is the insertion of lithium into rhenium trioxide via the reaction

0.3LiI(s) + Re03(s) Li03ReO3(s) + 0.15I2(s) (3.15)

The structure of Re03 is shown in This reaction may be carried out by grinding the two solids, anhydrous
Fig. 4.3. lithium iodide and rhenium trioxide, in a pestle and mortar. The rhenium
trioxide structure is sufficiently open to allow rapid diffusion of lithium ions
to empty sites and the metal oxide framework remains intact throughout the
reaction. The rhenium is partially reduced from Re(VI) to Re(V). Other
transition metal oxides including W03, Mo03, V205 undergo similar insertion
reactions with lithium and other small ions such as Na+ and H+.

Ion exchange

Structures of the type described above, with strongly bonded layers or

frameworks containing weakly held guest molecules or ions, frequently
undergo facile reactions at low temperatures in which one weakly bonded
species is swapped for another. This ion exchange behaviour of zeolites is
well known. For example, the process

{2Na+/zeolite 4A} + Ca2+ (aq) —> {Ca27zeolite 4A} + 2Na+(aq) (3.16)

where sodium ions in a zeolite framework are replaced by calcium ions from
solution, is widely used in water softening processes.
Similar reactions can occur for other materials that have ions weakly
interacting with a strongly bonded framework. One example is LiTiNb05
which has a layer structure of the type schematically shown in Fig. 3.11. The
lithium ions may be exchanged with other ions such as sodium and potassium
by stirring the solid with a solution of the desired cation.
Fig. 3.11 The structure of LiTiNb05
with linked octahedra (NbOe and
TiOe) forming layers separated by 3.8 Problems
Li ions.

3.1 The reaction MgO + A1203 only occurs at a reasonable rate at

temperatures above 1400°C. How could MgAl204 be prepared at lower
temperatures?

3.2 A compound of the stoichiometry KNbTiOj may be prepared via two

methods (i) reaction of K2C03, Nb2Os and Ti02 at high temperature and (ii)
reaction of Li2C03, Nb205 and Ti02 at high temperature followed by stirring
the product with potassium nitrate solution at 50°C. The products have
different structures. Explain.

3.3 Explain what happens if a 40:60 mixture of A and B is cooled from T4 in

a system adopting the phase diagram shown in Fig. 3.5.
4 Transition metal oxides

4.1 Review of binary transition metal oxide structures

MO

'The first row transition metal monoxides TiO to NiO adopt the rocksalt
structure shown in Fig. 4.1. This would be predicted from the ionic radius
ratio, rM2+/rQ2_, between 0.414 and 0.732, and filling of the octahedral site in
a close-packed array of oxide ions. Many of these materials are non-
stoichiometric existing over a range of composition and this behaviour is
described in detail in Chapter 6. The electronic properties of these compounds
which range from metal to insulator are discussed in Chapter 5. CuO has a
structure formed from linked Cu04 square planes, a more typical geometry for
Cu2+ than the octahedral sites found in the rocksalt structure. Few second and
third row transition metals form monoxides.

® M 0 0

Figs. 4.1 and 4.2 The rocksalt structure and the rutile structure respectively.

mo2
The majority of transition metal dioxides crystallise with the rutile structure,
shown in Fig. 4.2, though in some materials the structure is slightly distorted.
Materials of this stoichiometry exist for Ti, Cr, V and Mn in the first row
transition metal series and for Zr-Pd and Hf-Pt in the second and third rows.
The rutile structure is again predicted for the majority of transition metal
dioxides on the basis of radius ratio rules and filling of sites in a close-packed
oxide structure.
38 Transition metal oxides

The rutile unit cell is tetragonal with the metal atoms lying on the cell 1
corners and body centre; however, because the oxide ions are not related i
through the body centred condition the unit cell is primitive. The co¬
ordination geometry around the metal ion is approximately octahedral, but t
four metal to oxide ion distances are, in general, slightly different from the :
other two.

mo3

Few transition metals adopt the +6 oxidation state in oxide and, therefore, ,
^ Re © O oxides of the stoichiometry M03 are rare. Re03 has the structure shown in i
Fig. 4.3 which may be easily constructed by linking Re06 octahedra through i
all vertices. This gives rise to a primitive, cubic unit cell. W03 adopts a i
Fig. 4 .3 The Fte03 Structure. One
of the Re06 octahedra is outlined. slightly distorted variant of the Re03 structure, but still contains solely corner
sharing octahedra.

4.2 Ternary oxides

The structure of many binary oxides can be predicted on the basis of the
relative sizes of the metal and oxide ions and filling of holes in a close
packed oxide lattice. Such predictions of structure are more difficult for
ternary phases. The combination of two or more metals in an oxide generates
a wealth of structural possibilities dependent on the relative sizes of the two
metal ions and the oxide ion. In addition the stoichiometry of the ternary
oxide may be changed by varying the proportions of the two component
oxides and, for transition and lanthanide elements, the oxidation state. For
example, at least twenty ternary oxide phases are formed between strontium
and vanadium including SrV206, Sr2V205, SrV03 and Sr2V04. The structural
chemistry of ternary and more complex oxides is, thus, an extensive subject.
There are, however, a few structures which are widely adopted by ternary
oxides and many materials from these structural classes have important
technological applications. These ternary oxide structures are also used as
units in building more complex oxides, such as the high temperature
superconductors, Chapter 8.

4.3 The perovskite structure

The perovskite structure, frequently adopted by materials of the stoichiometry

AB03 is, perhaps, the most widespread ternary phase. Perovskites are named
after the mineral CaTi03 which was identified by the Russian mineralogist
L.A. Perovski. Several hundred materials of stoichiometry AB03 adopt the
perovskite structure or a slighted distorted version. In the notation AB03 the
A cation is conventionally the larger of two metal ions so, for example, in
SrTi03 [ r(Sr2+)= 1.52 A and r(Ti4+) =0.745 A)] strontium would be denoted
as the A cation and titanium as the B cation.
 Inorganic materials chemistry 39

The perovskite structure is shown in Figs 4.4a and 4.4b with the A and B
cations at the unit cell origin respectively. The coordination geometry of the
smaller B cation is octahedral while that of the larger A cation is 12 fold.

Fig. 4.4 The perovskite structure with the origin at the B (a) and A (b) cations.

The relationship of the structure to that of Re03 is readily appreciated by

reference to Figs. 4.5 and 4.3. The B06 octahedra link all vertices forming the
same cubic framework as in Re03. The addition of the A cation at the centre
of the unit cell produces the AB03 perovskite stoichiometry.
The range of oxides adopting the perovskite structure is delineated by the
relative sizes of the cations A and B, provided that their overall charge sums
to six. The dimensions of the sites accommodating the A and B cations mean
that the perovskite structure is frequently found for ternary AB03 oxides
formed with one large and one small cation. Some examples are given in
Table 4.1
In some complex oxides the perovskite sites may be shared by two or more
cations. A complicated looking formula such as La2MgRu06 may be rewritten
as La(RuwMg,/2)03, which is a perovskite with La on the A site and a random Fig. 4.5 The perovskite structure
mixture of Ru and Mg on the B sites. using linked octahedra.

Table 4.1 Perovskite cation types

Charge on B Example
Charge on A

3+ LaCr03
3+

4+ SrTi03
2+
5+ NaW03
1+
40 Transition metal oxides

no

Fig. 4.6 The powder X-ray

diffraction pattern of SrTiOs.

The cubic perovskite unit cell has a primitive lattice and, therefore, gives
rise to the typical powder X-ray diffraction pattern with no general absence.
Fig. 4.6 shows the powder diffraction pattern obtained from SrTi03.

Tolerance factors in perovskites

For ternary and more complex oxides, including perovskites, the size and
coordination preferences of three or more ions need to be satisfied
simultaneously by the structure; it is, however, rarely possible that all these
preferences can be accommodated perfectly. This can be illustrated for the
perovskite structure by deriving a relationship between the radii of the various
ions. Consideration of Fig. 4.7 yields the expressions

a = 2 x B-O = 2 ( rB + rQ) (4.1)

and

Fig. 4.7 The relationship of ionic

a = -L x 2 x A-O = yjl x (rA +rQ ) (4.2)
radii in perovskites.
s/1

where a is the unit cell parameter and rA, rB and rQ are the ionic radii of A,
B and O respectively. Hence, for the perovskite structure to provide the ideal
contact distance for both the A and B cations

2 X ( rB + ro ) = \/2 X ( rA + rO ) (4 3)

However, ionic radii are not fixed values, but figures derived from
consideration of the coordination preferences of a particular ion in a wide
range of crystal structures. Some variation in the ion separations from (r+ +
rj can thus be countenanced in any crystal structure and this can be built into
the above relationship by introduction of a tolerance factor, t.
 Inorganic materials chemistry 41

2 x t x ( rB + rQ ) = ft, x ( rA + rQ) (4.4)

For a perfect fit for the two ions into the perovskite structure t= 1; increasing
i deviation of t from unity indicates increasing strain on the perovskite structure
: in terms of the ions being too big or too small to fit on their allotted site,
i Ionic radii normally used in the above expression are those of Goldschmidt
j as these provide the best predictive power in terms of defining the stability of
the perovskite structure for different cation sizes. However, other systems of
Ionic radii can be used to illustrate how well the two cations fit into the
structure and, for consistency, those of Shannon and Prewitt used elsewhere Note that these ranges only apply
in this primer will be employed. Taking the values for Sr2+ (r= 158 pm ), Ti4+ to Shannon and Prewitt radii: for
(r = 74.5 pm) and O2- (r = 126 pm) yields a value of 1.002 for t indicating other systems, e.g. Goldschmidt,
the perovskite stability ranges for t
that the coordination distance preferences for both strontium and titanium are
are quite different.
met. As the values for t depart from unity increased stresses are built up in
the perovskite structure as one or both of the cations occupies a site of non¬
ideal dimensions. This may be overcome by a distortion of the perovskite
structure and for t values in the range 0.85 <r<l .06 a perovskite structure, or
a distorted version thereof, is frequently adopted by compounds of the
stoichiometry AB03. Using the Shannon and Prewitt radii, for 0.9<t<1.0 the
structure is often cubic, above 1.0 and below 0.9 the structure is normally
distorted. For t values outside the range 0.85-1.06 the perovskite structure
cannot meet the coordination requirements of the two cations and other
structures are normally adopted. For example, LiNb03 and FeTi03
have t « 0.9, as the cations are of similar sizes, and the perovskite structure
is not formed; the structure of LiNb03 is described in Section 4.5.

BaTiOj

For BaTi03 a tolerance factor of 1.06 is obtained; this value, greater than
unity, indicates that the titanium ion is occupying a site that is larger than it
would prefer. Above 120°C BaTi03 adopts a perfect cubic perovskite
structure; at these high temperatures the extensive thermal motion of the
various ions makes packing considerations in terms of ionic radii less
important. On cooling, BaTi03 undergoes a series of phase changes as
summarised in Table 4.2.

Table 4.2 Phase behaviour of BaTiO,

Transition -80°C 5°C 120°C

temperature

Crystal Rhombohedral Orthorhombic Tetragonal Cubic

Structure a = 3.995 A a = 4.002 A
c = 4.034 A

These phase changes are a result of the structure distorting from the perfect
perovskite description so as to develop a more acceptable co-ordination
geometry around the titanium atom. This is shown in projection
42 Transition metal oxides

Fig. 4.8 Cation displacements

relative to the oxide sublattice in
tetragonal BaTi03.

0.06A

in Fig. 4.8 with the barium ions at the unit cell comers. In the tetragonal
phase the major positional shift is that the titanium ion becomes displaced
from the central site towards one of the oxide ions in the cell faces; smaller
displacements of the barium and oxygen ions also occur. These displacements
make the lattice directions inequivalent such that a=b*c, that is the cell
symmetry reduces to tetragonal, though the distortion is quite small with c/a
= 1.01. However, this distortion is readily observed in the powder diffraction
pattern, Fig. 2.2. The indexation of the powder pattern from BaTi03 was
carried out in Section 2.6.
One effect of this displacement of titanium from the central site of BaTi03
is on the charge distribution in the unit cell. If charges are formally attached
to the various ions as Ba2+, Ti4+ and 02~ then the centres of negative and
positive charge in the cubic form of BaTi03 coincide at the cell centre.
However, in the tetragonal phase, owing to the large displacement of titanium
Fig. 4.9 The centres of positive ion, the unit cell charge centres are no longer coincident, Fig. 4.9. This
and negative charge in the polarisation of charge gives the unit cell a high dipole moment.
tetragonal unit cell of BaTi03. The displacements of the titanium ions in the large number of unit cells
which make up part of a crystal or domain of tetragonal BaTi03 can be
arranged such that in neighbouring unit cells the displacements are in exactly
the same directions. The result of this is that the individual unit cell dipoles
all line up in the same direction, Fig. 4.10a, and this behaviour means that
tetragonal BaTi03 is termed a ferroelectric. Ferroelectric crystals are non-
centrosymmetric and display a long range ordering of the dipoles below a
Centrosymmetric structures have characteristic temperature known as the Curie temperature. In the absence of
an inversion centre at the origin, an external electric field, the ordering of the dipoles within the various
that is for every atom on a site domains (sections of the crystal) produces a spontaneous polarisation of the
x,y,z in the unit cell there is an
crystallite. Following the application of an external electric field the direction
equivalent atom at -x-y-z. In
ferroelectric crystals, therefore, no of polarisation in the various domains can order so that they all line up with
inversion centre is present leading the external field. The polarisation of the crystal results in a very high
to a relatively low crystal dielectric constant for BaTi03 and it is widely used in capacitors.
symmetry. Charge separation in distorted perovskite structures is a common feature
though the interaction of ion displacements in neighbouring unit cells may
 Inorganic materials chemistry 43

occur in a manner different from that observed in BaTi03. In the tetragonal (a) Ferroelectric
phase of PbZr03 the lead ion displacements in adjacent unit cells are in
opposite directions, Fig. 4.10b, and this material is an antiferroelectric. □ □□□□□□
BaTi03, in common with a number of other ferroelectrics shows what is
(b) Antiferroelectric

□nnnnnn
known as piezoelectric behaviour. A piezoelectric material develops an
electric polarisation on the application of stress. Compression of a crystal of
BaTi03 causes further distortion of the tetragonal BaTi03 unit cell, a further
separation of the charges and a higher degree of polarisation of the charge. (c) Ferrielectric

This may be used in a number of ways to convert mechanical energy into

electrical energy; examples include applications as record player needles and
DODDDDD
in gas lighters. Fig. 4.10 Ordering schemes for
dipole moments.

4.4 Insertion compounds and perovskite structures

In Section 3.7 the reaction of Lil with Re03 to give Li03ReO3 was discussed.
The perovskite and Re03 structure are identical except for the filling of the
A sites in the former. The reason for the ease of this reaction can now be
readily seen in that the structure of the product, Li0 3Re03, is a perovskite with
a proportion, 30%, of the A sites filled. Owing to the small size of Li+ the
perovskite structure distorts slightly by tilting of the Re06 octahedra so as to
bring oxygen ions closer to it. The insertion of lithium ions into Re03 merely
involves diffusion of Li+ into the Re03 structure through the faces of the unit
cell cube. The Re-0 bonds forming the framework are unaffected allowing
this solid state reaction to progress at room temperature. The powder X-ray
diffraction patterns of Li03ReO3 and Re03 are almost identical, as they are
dominated by scattering from the rhenium (due to the high number of
electrons), except for a small shift in peak positions caused by a decrease in
lattice parameter.
Tungsten trioxide is another material which undergoes insertion reactions
of this type. W03, with a slightly distorted version of the Re03 structure,
forms the compound Na10WO3, adopting the perovskite structure. H06WO3
synthesised by inserting protons into W03, has a structure of the perovskite
type, but the hydrogen atoms, instead of sitting isolated on the perovskite A
sites, form OH groups with the oxide ions along the unit cell edges.

4.5 LiNb03
Lithium niobate whilst having the AB03 stoichiometry does not adopt the
perovskite structure. The ionic radii of Li+ and Nb5+ are too similar to fit the
perovskite structure, producing a very small tolerance ratio; in addition lithium » Li
normally occupies sites with low coordination numbers, 4 or 6. The structure
Fig. 4.11 The LiNb03 structure
adopted by LiNb03 is shown in Fig. 4.11 where the lithium ions are
showing concerted lithium ion
coordinated to six oxygens. The lithium ions may be displaced within the displacements of the ferroelectric
octahedral sites and LiNb03 is a ferroelectric. LiNb03 is also electro-optic in phase. Polyhedra outlines
that the application of an external electric field aligns the lithium ion represent NbOe octahedra.
44 Transition metal oxides

displacements and also changes the refractive index of the material. This
behaviour has led to important applications in optical devices.

4.6 The spinel structure

The spinel structure is widely adopted by ternary oxides and sulphides of the
stoichiometry AB2X4, X = 0,S. The overall charge on the cations is eight and
the possibilities for A and B are summarised in Table 4.3. Note that the label
B applies to the cation type which is twice as abundant in the structure. The
two cations are of a similar size in contrast to the perovskites.

Table 4.3 Possible charges on A and B in spinels.

Cation A charge Cation B charge Example

2+ 3+ MgAI204

4+ 2+ TiMg204, ZrCu2S4

6+ 1+ Na2W04t

* Distorted

The spinel structure is rather more complicated than those discussed earlier
though one helpful representation is shown in Fig. 4.12. The structure is based
upon a cubic close packed arrangement of anions in which a proportion of
tetrahedral and octahedral sites is occupied. With n close packed anions, 2n
tetrahedral holes and n octahedral holes are generated. In the spinel structure
one-eighth of the tetrahedral holes and one-half of the octahedral

O-j X)

?x> rx>.

• B

O 0

Fig. 4 .12 The spinel structure.

 Inorganic materials chemistry 45

holes are occupied in an ordered manner throughout the structure. The A

cations normally occupy the tetrahedral sites and the B cations the octahedral
ones giving the stoichiometry AM8Bn/2Xn, which is equivalent to AB2X4.
The unit cell in Fig. 4.12 shows the tetrahedral and octahedral co-ordination
of the A and B cations. Rather than considering the structure as the filling of
holes in a close packed anion array one alternative is to draw the structure as
a face centred cubic array of A ions into which smaller cubes of stoichiometry
B4X4 and AX4 are added. The AX4 and B4X4 cubelets alternate in the structure
and the unit cell contains four of each giving the unit cell stoichiometry
4(AX4) + 4(B4X4) + 4A ( from the cell corners and faces) = A8B16X32
corresponding to the formula AB2X4. The structure is face centred cubic;
consideration of any ion in the unit cell will show other ions in identical
environments with translations of +(V2,V2,0), +(1/2,0,1/2) and (Od/VA). The
powder X-ray diffraction pattern of a spinel, for example Fig. 4.13, shows the
typical general absences for a material crystallising with a face centred cubic
structure.

Fig. 4.13 The powder X-ray

diffraction pattern of a face centred
cubic material such as a spinel.
Owing to the high symmetry of the
spinel lattice some of the
reflections, e.g. 200, are normally
absent.

This arrangement of the A and B cations over the tetrahedral and

octahedral sites in the spinel structure, that is A‘B2°X4, where the superscripts
refer to the coordination geometry, is known as the normal arrangement and
the structure as the normal spinel structure. The B ion, which exists at twice
the level of the A cation, occupies only octahedral sites within the structure.
In some compounds adopting the spinel structure, known as inverse spinels,
a different arrangement of cations exists corresponding to Bl(AB)°X4, where
all the A cations occupy octahedral sites and the B cations are distributed
equally over the two coordination geometries.
Why some materials adopt the inverse spinel structure rather than the
normal one is a result of a number of factors including the ion charges. The
46 Transition metal oxides

following discussion applies to the most common formulation (A2+B23+)X2“4

but the arguments can be equally well extended to the (A4+B22+)X2'4 system.
Ionic size is one parameter, though flexibility in the position of the anions
in the spinel structure permits a range of cation sizes to be coordinated at
their preferred contact distances. The higher charged B cations are generally
smaller than the A ions, whilst the tetrahedral sites in a close packed anion
array are significantly smaller than the octahedral holes. This would point
towards the smaller B cations preferring to occupy the tetrahedral sites i.e. the
inverse formulation Bl(AB)°X4. However, the oxide ions are not fixed and
positional displacements are possible, to some extent, to accommodate
differently sized ions.
The ion size factor is generally outweighed by lattice energy calculations;
probably the most important parameter in determining the cation distribution.
Calculation of the lattice energy of spinels with the same cations, but with the
two different distributions, shows that the normal spinel often has a slightly
higher lattice energy than the inverse description. The simple reason for this
is that lattice energies are determined through summation of all the Coulombic
interactions of the ion pairs in the lattice. That is.

UL oc X z+z_ / r+_ (4.5)

where UL is the lattice energy, z+ and z_ cation and anion charge, r+_ their
separation and the summation is carried out over the whole lattice for all ion
pairs. Placing a highly charged cation on a high coordination site contributes
strongly to the lattice energy and produces a greater lattice energy than other
arrangements. This lattice energy consideration indicates that the majority of
(2+,3+) spinels should adopt the normal structure. In practice for non¬
transition metal systems, where the third factor, discussed below, is non¬
operative, inorganic spinels, for example MgTi204, are generally found to
adopt the normal spinel structure.

At
Fig. 4.14 d-Orbital distributions in
octahedral and tetrahedral fields.
Filling of orbitals is for a d3 ion.
'Octahedral Free Tetrahedral
Ion
a b
 Inorganic materials chemistry 47

In spinels which contain a first row transition metal, a third factor, partial
occupancy of the d-orbitals, is of importance in determining the cation
distribution. This is the crystal field stabilisation energy of the various ions
in the octahedral and tetrahedral coordination geometries. The occupancy of
the d-orbitals on Cr3+ ions in a spinel on both tetrahedral and octahedral sites
is shown in Figs. 4.14a and 4.14b. In an octahedral site the d-orbitals are split
linto two groups, the t2g set ( d^, dV2 and dt,), which point between the ligands,
land the eg set (dx2_y2, dz2), which point towards the ligands. The t,g set are
! stabilised relative to the eg set by an amount A0. In the tetrahedral case the
arrangement is inverted and the separation of the t2 and e sets is A, where A,
~ 4/9A0. In oxide spinels the splitting of the t^ and eg levels is fairly small as
oxide is a weak field ligand. This leads to oxide spinels being high spin
materials where the electrons are distributed in the d-orbitals with the
maximum number of parallel spins.
With Cr3+, a d3 ion, on an octahedral site only the t2g set of orbitals are
occupied, Fig. 4.14a; owing to the lower energy of these orbitals this
octahedral configuration is stabilised relative to the case where the ion is non-
coordinated to a level known as the crystal field stabilisation energy (CFSE).
In terms of A0 this CFSE is 3x2/5xA0 = 6/5xA0. This may be compared to the
case where chromium occupies a tetrahedral site, Fig. 4.14b; here with a high
spin electron distribution, e2t2\ the CFSE is 2x3/5At-2/5At = 4/5xA, (= V3AJ.
This is much less than the octahedral case and the CFSE considerations imply
that Cr3+ would very much prefer to occupy an octahedral site than a
tetrahedral one in a spinel. In a spinel, such as MgCr204, the CFSE
considerations reinforce the lattice energy support for the normal spinel with
the cation distribution MgtCr2°04.
The difference between the CFSE in octahedral and tetrahedral geometries
is known as the excess octahedral stabilisation energy and the values for
various first row transition metal 2+ and 3+ ions are given in Table 4.4.

Table 4.4. Excess octahedral stabilisation energies (kJ mol ')

O
+

V3*
Mn2+ Fe2+ Ni2+ Cu2+ Ti3* Cr3* Mn3* Fe3*
o

0 17 31 86 64 29 54 158 95 0

The high excess octahedral CFSE for Cr3+ of 158 kJ mol-1 (corresponding
to 6/5xA0 - 4/5xA, ) was demonstrated above and supports these 3+ ions on
an octahedral site. If another spinel, Fe304, is considered, the use of CFSE’s
to stabilise an inverse spinel can be demonstrated. Fe304 can be rewritten as
Fe2+Fe3+204, fitting in with the general spinel formula. The excess octahedral
CFSE for Fe3+ is zero; as a d5 ion in a high spin environment the distribution
of the d-orbitals, which each contain one electron, produces no CFSE in any
coordination. However, for Fe2+, a d6 ion, there is a small excess octahedral
CFSE of 17 kJ mol-1; this indicates that this ion would rather occupy an
octahedral site in preference to the tetrahedral site that this 2+ ion would be
allocated in a normal spinel structure. In Fe304 the excess octahedral CFSE
48 Transition metal oxides

for Fe2+ is sufficient to overcome the lattice energy mandate and the Fe2+ ions
In most spinels other than Fe304 occupy octahedral sites giving the inverse spinel (Fe3+)l(Fe2+,Fe3+)0O4. With
the excess CFSE would need to be
Fe3+ and Fe2+ occupying identical sites in the lattice, electrons may readily hop
somewhat larger than this to
overcome the lattice energy from one iron site to the next and Fe304 is a good electronic conductor.
preference for a normal The zero excess octahedral CFSE for Fe3+ means that many transition metal
arrangement. spinels containing this species are inverse, as most 2+ transition metals have
some preference for octahedral sites. Hence, CoFe204 and NiFe204 are also
inverse spinels. Other inverse spinels in 2+, 3+ systems are rather rare as
there must be a significant CFSE driving force to overcome the lattice energy
preference for normal spinels. One ion which does frequently form inverse
spinels is Ni2+ which has a high excess octahedral CFSE of 86 kJ mol-1. In
NiAl204 this demand of Ni2+ for an octahedral site produces the ion
arrangement Al‘(NiAl)°04, an inverse spinel.

4.7 K2NiF4

The K2NiF4 structure is shown in Fig. 4.15 and consists of sheets of NiF6
octahedra sharing four vertices. These layers are separated by K+ ions in nine
fold coordination to fluorine. The structure is body centred with the NiF6
octahedron at the unit cell centre displaced by (Vi,Vfc,V6) from that at the cell
origin. This structure is widely adopted by ternary oxides of the stoichiometry
A2BX4 where one cation, A, is much larger than the other, B. This contrasts
with the spinel structure which has the same stoichiometry but occurs where
A and B have similar ionic radii. Examples include Sr2Ti04, La2Ni04, Cs2U04
and the superconducting phase (La; gjBao 15)Cu04 discussed in Chapter 8.

Fig. 4.15 K2NiF4 structure shown

with NiF6 octahedra and potassium
ions as circles. 4.8 Problems

4.1 Calculate a tolerance ratio for perovskite PbTi03 taking the ionic radii as
1.68 A (Pb2+), 0.745 A (Ti4+) and 1.26 A (02~). This material is a ferroelectric.
Explain this behaviour in terms of likely ion displacements.

4.2 By consideration of CFSE factors predict whether the following spinels

would be expected to adopt the normal or inverse structures, (a) FeCr204 and
(b) NiGa204.
5 Electronic, magnetic and
optical properties of
inorganic materials

rhe electronic properties exhibited by solids are crucial in a large number of

inorganic materials applications. These unique electronic properties result from
the extended structures adopted by many inorganic materials, where strong
interactions between the atoms, ions or molecules occur throughout the lattice,
[n terms of conductivity, behaviour ranges from insulating through
semiconducting to metallic and superconducting. Similarly, the interaction
between the unpaired electrons in inorganic material structures gives rise to
magnetic properties characteristic of the whole solid rather than the individual
atoms. Finally, the potential for concerted ion displacements in neighbouring
unit cells produces ferroelectrics and optoelectric materials. This chapter
provides an introduction to these co-operative phenomena.

5.1 Introduction to band theory

In simple molecules the interactions between the atomic orbitals on the

various atoms forming the compound give rise to molecular orbitals. The
number of molecular orbitals generated in a molecule is identical to the
number of atomic orbitals used in forming them. This may be illustrated by
consideration of the K orbitals formed in a series of long chain alternate
alkenes, Fig. 5.1.

Fig. 5.1 n Molecular orbitals of

alternate alkenes. The dotted line
separates the bonding and
at antibonding orbitals.

C.H C.H.

In ethene, Fig 5.2a, the pz orbitals on the two carbon atoms interact to form
one bonding and one antibonding configuration. If the number of carbon
atoms and n orbitals in the chain is increased to four, producing 1,3 butadiene,
Fig. 5.2b, the four carbon 7t orbitals can interact to give rise to two bonding
and two antibonding orbitals. Further extension to an eight
50 Electronic, magnetic and optical properties of inorganic materials

Fig. 5.2 Combinations of p orbitals

giving rise to n molecular orbital in
ethene (a) and 1,3 butadiene (b). (a) (b)

carbon atom chain produces four bonding and four antibonding molecular
orbitals. With continued extension of the chain, every pair of carbon pz
orbitals added to the molecular orbital diagram provides one bonding and one
antibonding combination. The molecular orbitals, within the bonding and
antibonding groups, become closer and closer in energy, and this continues
until such time as the chain has an almost infinite number of orbitals, when
they form a practical continuum. In such cases, rather than drawing all these
energy levels as separate lines, they are replaced in the energy level diagram
by a block representing a continuous band of energy levels. Hence, in the case
of an infinitely long chain of 7t orbitals, corresponding to (CH)X, the energy
level diagram consists of two bands, one bonding in character and one
antibonding in character.
This behaviour of atomic orbitals in infinite structures, forming continuous
bands of energy levels, is frequently found in inorganic materials where there
is overlap between the atomic orbitals. The simplest case is that of an
elemental metal such as sodium. The electronic structure of a sodium atom
is ls22s22p63s'; the core levels with principal quantum numbers one and two
are held close to the nucleus and extend poorly into the body centred Cubic
lattice adopted by sodium. However, the 3s orbital extends out from the
nucleus and can overlap with similar orbitals on neighbouring sodium atoms.
In crystalline sodium metal, with an almost infinite lattice and the number of
orbitals approaching Avogadro’s Number 6xl023, the overlap of these 3s
orbitals gives rise to a band. Other atomic orbitals on sodium which are
sufficiently diffuse, and overlap with similar orbitals on neighbouring atoms,
will also give rise to bands, for example, the 3p levels. The band structure of
sodium metal derived from the 3s and 3p levels is shown in Fig. 5.3. The
width of the individual bands derived from the 3s and 3p orbitals is such that
they also overlap.
Remembering that each atomic orbital gives rise to one molecular orbital
and, hence, one level in the continuous band, the four orbitals on each sodium
atom ( 3s, 3p^, 3pv and 3pz) will, within the final band structure, give rise to
four levels. Each level can contain two electrons as with the original atomic
orbitals on the sodium atom. In an infinite lattice of N sodium atoms 4N
energy levels are generated. Each sodium ion contributes only one electron to
 Inorganic materials chemistry 51

this band structure, shown in Fig. 5.3, which can hold eight and, therefore, the
band is only partially filled. It is this partially filled band which gives rise to
the electronic properties of metals such as high electrical conductivity and
lustre. Any inorganic material which has an arrangement of atoms, ions or
molecules that produces partially filled bands would be expected to
demonstrate the physical properties of a metal.

5.2 Band structure and electronic properties

Fig. 5.3 The band structure of
Metals
sodium metal.
In all solids with effectively infinite structures there will be an interaction
between atomic orbitals with the potential to give a band structure. The
electronic properties of a solid depend upon the nature of these bands in terms
of their width, separation, the number of electrons they can hold, and the
number of electrons in the system. The width of the bands depend on the
degree of overlap of the orbitals which interact to give rise to them. Strongly
overlapping atomic orbitals give rise to wide bands whilst weakly interacting
orbitals give rise to narrow bands. This can be illustrated by reference back
to molecular orbitals. Weakly overlapping atomic orbitals give rise to bonding
and antibonding molecular orbitals with energies close to those of the original
atomic levels, that is with little dispersion in energy terms; with strongly
overlapping atomic orbitals the bonding and antibonding combinations are
more spread out along the energy axis. The electrons in the wide band can be
considered to be dispersed over the whole lattice whilst the electrons
occupying a narrow band are more strongly associated with the nuclei.
Band structure diagrams are frequently drawn to represent the number of
electrons which may occupy a band. In the band structure derived from the
t
(CH^ chain there will be a large number of energy levels near the centre of
Energy
the band and much fewer near the edges. Evidence for this build up of energy
levels at the mid point of the energy scale can be seen in the molecular orbital
diagram of C16Hlg. The information on the number of energy levels at any
particular point in a band can be represented in a density of states diagram
such as Fig. 5.4 where N(E), the density of states (or energy levels) is plotted
as a function of energy. Filling of the bands by electrons in such a diagram
can be denoted by shading of the band. The level to which a band structure
is filled is denoted as the Fermi level, EF. If this Fermi levels falls in the
middle of a band then a material can be expected to exhibit metallic Fig. 5.4 Density of states diagram
properties. The band structure of magnesium metal is shown in Fig. 5.5. for (CH)„.

Insulators

If the Fermi level lies at the top of a band it is normally drawn mid way
between this band and the next, which will be empty, Fig. 5.6. This band
structure is that of an insulator which crystallises with an infinite lattice;
diamond is a good example. The separation of the filled and empty bands, the
band gap, is large, typically greater than 3.0 eV.
52 Electronic, magnetic and optical properties of inorganic materials

Semiconductors

With semiconductors, a similar band diagram to that of insulators is obtained;

however in these materials the band separation is much smaller, typically less
than 3.0 eV. Silicon crystallises with the same structure as diamond and the
band structure is very similar. Differences arise because the atomic orbitals
on silicon are higher in energy than those on carbon and also more diffuse;
this modifies the band structure to produce wider bands and a smaller
2s 2p
separation between the filled (valence) band and the empty (conduction) band.
In silicon and similar materials with a small separation of the filled and empty
E band there is sufficient thermal energy at room temperature to promote some
Fig. 5.5 Band structure of electrons from the valence band to the conduction band. These electrons, and
magnesium metal as a density of the holes in the valence band that they leave behind, give rise to the moderate
states diagram. conduction properties of semiconductors. The smaller the band gap and the
higher the temperature then the greater the number of electrons which can be
excited into the conduction bands. This results in higher conductivity in
materials with small band gaps and the increasing conductivity seen in
semiconductors with increasing temperature.

5.3 Simple applications of band theory to inorganic

materials

First row transition metal monoxides

Fig. 5.6 Density of states and
Fermi level for an insulator. The colours of the majority of transition metal monoxides are quite different
from those normally found for the same metal ion in octahedral coordination

5.7

Figs. 5.7 and 5.8 Rocksalt structure and the overlap of the metal t2g orbitals in the unit t
cell faces.
 Inorganic materials chemistry 53

omplexes. This is particularly so for the early transition elements, for

xample, V(H20)62+ is violet whilst VO is black, but for nickel the oxide is
be same colour, green, as Ni(H20)62+. For the early transition metals this
vould indicate that the electronic structure of the oxides is quite different
rom discrete octahedrally coordinated 2+ ions.
The first row transition metal oxides from TiO to NiO adopt the rocksalt
tructure shown in Fig. 5.7, though many of these compounds are non-
toichiometric as discussed in Chapter 6. The coordination geometry of the
netal ion is octahedral and the d-orbitals on the metal ion will be split
nergetically into two sets as in typical coordination complexes. The eg levels,
orresponding to the d^.2 and dz2 orbitals, point towards the oxide ions in the
tructure and the t2g levels, formed from the d^, dt, and dyz orbitals on the
netal, point between the oxide ions. Looking at these directional properties
)f the t2g levels in relation to the structure of the monoxides, Fig. 5.8, shows
hat these orbitals point towards identical orbitals on neighbouring metal
itoms.
In TiO the Ti2+ d-orbitals are very diffuse as the effective nuclear charge
elt by these levels is low. This results in good overlap between the t2g orbitals
dong the (110) directions in the crystal. Hence, rather than considering
liscrete t2g levels in solid TiO, the electronic structure should be described in
erms of a band. The bands formed from the overlap of the t2g orbitals will be
Typical ranges for conductivities
lble to hold up to six electrons from each contributing titanium ion; Ti2+ is a
are metals 10'-105(Q cm)-1,
l2 ion and these two d-electrons will partially occupy the t2g band, Fig. 5.9a. semiconductors 10-5-101 (O cm)-1
riO, therefore, exhibits metallic behaviour with a high electronic conductivity and insulators <10-12(O cm)-1
if about 103 (Q. cm)-1. Similar electronic behaviour occurs with other early
ransition metal monoxides, for example VO.

Fig. 5.9 The band structure of TiO

(a) and the corresponding levels in
NiO (b).

The behaviour of the latter transition metals is quite different. As the first
•ow is crossed, increasingly poor screening of the d-electrons by each other
-esults in a rapidly increasing effective nuclear charge. This causes a marked
contraction of the d-orbitals and consequently the overlap between the t2g
levels decreases significantly. Rather than forming a band, the t2g levels on the
lickel ions within NiO should be considered to be discrete levels. In addition,
he increased number of d-electrons for nickel, Ni d , means that the t2g
54 Electronic, magnetic and optical properties of inorganic materials

levels are filled and it is the eg metal orbital set which is partially empty, Fig;
5.9b. These eg orbitals point towards the oxide ions and cannot overlap with
those on neighbouring nickel ions to form a band structure. NiO therefore
shows no metallic properties. It is an insulator and the d-orbital energies are
very similar to those in the hexa-aquo ion, leading to similar d-d transitions;
and, hence, visible spectrum.

Trioxides

The contrasting electrical properties of W03 and Re03 can also be rationalisec
in terms of their band structures. W03 is a pale yellow, insulating solid whils
Re03 is a beautiful, lustrous, red metal. Both materials crystallise with the
Re03 structure described in Section 4.1 though that of W03 is slightly
distorted by twisting of the W03 octahedra. The octahedral co-ordination oi
the metal produces the same d-orbital splitting observed in the monoxides
producing the t2g and eg sets. The p-orbitals on the oxide ions are also dividec
into two sets; one p-orbital points towards the metal ion whilst the other two;
px and pv, are perpendicular to the M-O-M direction. The t2g orbitals on the
metal ions and the px, py orbitals on the oxide ion overlap as shown in Fig i
5.10 producing a band which can contain up to six electrons. In W03, W6+ d°
this band is empty but in Re03, Re6+ d1, the one d-electron partially fills the
band giving rise to the metallic characteristics of this material. At room
temperature, Re03 is as good a conductor of electricity as copper metal.

Fig. 5.10 Re (t^) and O (p) orbital

7t overlap and the band structure of
Re03.

The insertion compounds of W03, for example NatW03 or in electronic

terms Na/[W03(6~*)+], contain the electron donated by the inserted sodium
atom in the n band discussed above. These materials, therefore, demonstrate
metallic characteristics and Na^WOj is a gold coloured solid.

Graphite intercalates

The band structure of pure graphite is shown in Fig. 5.11a. A particulai

feature of this diagram is that the valence and conduction bands touch bu!
there are no energy levels at the Fermi level. Another way of considering
5

Inorganic materials chemistry 55

graphite is as a zero band gap semiconductor. Due to thermal energy a few

ilectrons are excited from the lower valence band into the conduction band
md the Fermi level in graphite, above absolute zero, lies slightly into the
conduction band. As a result graphite is a moderate conductor of electricity.
However, removing or adding electrons to the graphite band structure will
produce a Fermi level well inside either the conduction or valence band and
this can be achieved chemically through the insertion chemistry of graphite.
^ Reaction of graphite with potassium to yield C8K transfers electrons from
potassium metal on to the graphite layers; these occupy the valence band and
the density of states diagram in Fig. 5.11b results. Similarly reaction with
bromine to form CgBr transfers electrons from the graphite layers on to
bromine to form bromide and the valence band is partially emptied, Fig.
5.11c. Both CgK and CgBr are true metals and exhibit higher electrical
conductivity than pure graphite

Fig. 5.11 Band structures of (a)

graphite, (b) C8K and (c) CaBr.
5.4 Magnetism in extended structures

As with the electronic properties of solids the important magnetic properties

of inorganic materials result from the interaction between the atomic centres
in the extended lattice.

Ferromagnetism, ferrimagnetism and antiferromagnetism

Transition metals in the majority of their compounds have partially filled d-

orbitals and frequently a proportion of the electrons occupying these orbitals
are unpaired; this gives rise to the magnetic behaviour. Where discrete metal (a) Paramagnetic
centres exist, for example solutions of transition metal compounds and in y S
* "»>Z •4 1/5 >2

many simple, molecular compounds, the unpaired electrons are normally

s s X
orientated randomly on different metal centres. This random orientation of (b) Ferromagnetic
electron spins and their associated magnetic fields is known as paramagnetic
behaviour.
However, in solid materials with ordered lattices, the unpaired electrons on
(c) Antiferromagnetic
individual metal centres can interact, resulting in particular alignments.
Paramagnetic behaviour, Fig. 5.12a, is observed in the majority of solid
inorganic materials at high temperatures where any interaction of the electrons
(d) Ferrimagnetic
between metal centres is removed by thermal disordering.
The ordering of the magnetic moments in solids may occur in a variety of
ways. If the unpaired electrons align such that their magnetic moments all
orientate in a parallel manner the material is ferromagnetic, Fig. 5.12b. One Fig. 5.12 Spin arrangements in
alternative is that the magnetic moments may align on neighbouring metal paramagnetic, ferromagnetic,
antiferromagnetic and ferrimagnetic
centres in opposite directions or antiparallel, Fig. 5.12c, with the overall
materials.
magnetic moment averaging to zero: such behaviour is known as
antiferromagnetism. A third type of behaviour, ferrimagnetism, is where the
alignment of magnetic moments produces an overall magnetic moment, but
not all the individual electron spins are parallel; Fig. 5.12d shows one
example.
56 Electronic, magnetic and optical properties of inorganic materials

The magnetic interaction between the various centres in a solid is normally)

(a) Paramagnetic relatively weak and is readily overcome by thermal energies. Hence, many v
materials which have magnetically ordered structures at low temperatures s
become paramagnetic on heating.
The temperature dependence of the magnetic susceptibility of paramagnetic,,
ferromagnetic and antiferromagnetic materials is shown in Fig. 5.13. Thee
magnetic susceptibility is a measure of the magnetisation of a material when n
placed in a magnetic field. The positioning of a paramagnetic material with i
randomly-orientated, individual, magnetic moments in an external magnetics
(£0 Ferromagnetic
field results in their partial alignment parallel to the field; this acts against I
thermal effects which tend to disorder them. High alignment of the individual 1
magnetic moments in a sample parallel to the external field produces a high i
magnetic susceptibility. Hence, as a paramagnetic material is cooled the 2
energy available for thermal disordering of the individual magnetic moments sj
is reduced and such nuclei align more readily parallel to the external magnetic 2
field; Fig. 5.13a results with an increasing susceptibility with decreasing?
temperature.
If the material undergoes a transition to a state where all the individual 1
(c) Aitiferromagnetic
magnetic moments align, that is a ferromagnetic material, there will be ai
surge in the magnetic susceptibility as all the electrons attempt to line up in 1
the same direction. The temperature of this transition is known as the Curie 2
temperature, Tc ; the temperature dependence of the magnetic susceptibility /
of a material which becomes ferromagnetic on cooling is shown in Fig. 5.13b..
Note that above the Curie temperature the compound acts exactly as a 1
paramagnetic material.
If a material undergoes a transition to an antiferromagnetic state the 2
Fig. 5.13 Temperature dependence
susceptibility drops markedly, Fig. 5.13c, as the individual magnetic moments ;
of the magnetic susceptibility of
paramagnetic (a), ferromagnetic (b) try to line up half parallel and half antiparallel to the external field. The 2
and antiferromagnetic materials (c). temperature at which a material becomes antiferromagnetic is known as the 2
Neel temperature, TN.

5.5 Magnetic properties of oxides

Superexchange

One mechanism which causes magnetic ordering in inorganic materials is;

superexchange and is illustrated in Fig. 5.14. Many inorganic systems, for
example oxides and sulphides, contain metal centres separated by an anion. .
A good example is MnO which adopts the rocksalt structure where overlap >
of the metal eg d-orbitals with the p-orbitals on the anion can occur along ;
each of the cell axes. Mn2+ has five d-electrons in a high spin configuration 1
and two of these occupy the eg levels. An unpaired electron occupying an eg
orbital on Mn2+ can interact with an electron in the oxide p-orbital forming a 1
pair of antiparallel spins. The other electron in the oxygen p-orbital is
necessarily of opposite spin orientation and this can in turn interact with an 1
Fig. 5.14 Schematic of eg electron on the next manganese ion. This chain of interactions can run 1
superexchange via an oxide ion. throughout the structure and will produce alternating spin orientations on all
 Inorganic materials chemistry 57

O # Mn
Fig. 5.15 Antiferromagnetic
ordering of Mn2+ ions in MnO below
® O 122 K.

leighbouring manganese ions. At low temperatures the strength of these

mperexchange interactions through oxide ions is such that this magnetic
Drdering process occurs and MnO is antiferromagnetic below 122 K, 77 Fig.
5.15. Above the Neel temperature the thermal energy is greater than the
interaction energy between the unpaired electrons and the electrons on
adjacent manganese ions become randomly oriented.
Similar behaviour to that of MnO is seen with other late first row transition
metal monoxides; Fe, xO, CoO and NiO, all adopting the rocksalt structure,
with 77s of 198, 293 and 523 K respectively. The increasing Neel
temperature across this series may reflect the decreasing lattice parameters and
increasing overlap between the metal and oxide ion orbitals.

Spinels

The ferrite spinels of composition MFe204, M = Fe, Ni, Zn, have important
magnetic properties which have led to their use in transformer cores and in
magnetic recording media. The distribution of cations in the spinel structure
has been discussed in Section 4.6 and spinels containing Fe3+ because of its
zero excess octahedral CFSE are often inverse. This results in structures
which contain Fe3+ on both tetrahedral and octahedral sites. In Fe304 the iron
ion distribution is that of an inverse spinel, (Fe3+)'(Fe2+,Fe3+)°04, and all the
metal ions have unpaired electrons. The superexchange interactions between
the various sites is complex but the overall effect, in the magnetically ordered
phase, is to align the magnetic moments of all the octahedral sites in one
direction and those on tetrahedral sites in the opposite. Fe304 is, therefore, a
ferrimagnet.
58 Electronic, magnetic and optical properties of inorganic materials

5.6 Optical properties of solids

Non-linear optical materials

The interaction of laser light with the electronic charge distribution in non-i
centrosymmetric crystals gives rise to observable non-linear effects which may;
be used for the amplification, modulation and conversion of the laser;
frequencies. Frequency conversion is particularly useful as it extends the range;:
of available laser frequencies. In designing inorganic materials as non-linean
optic materials the relationship between the charge distribution, dictated by then
structure, and the optical properties is paramount.
The interaction of light with the electronic charge distribution around an ioni
induces oscillations of the electron cloud. With light of relatively low intensity;
the induced polarisation, the displacement of the electron cloud or dipolei
moment per unit volume, P, is directly proportional to the magnitude of the:
electric field of the light wave, E. That is

P = XE (51)
where % is the linear optical susceptibility, a function of the refractive index)
of the material. In materials with a crystal symmetry below tetragonal three:
values of % are required to define the optical susceptibility in the differenn
lattice directions. Laser light generates very intense electric fields and gives:
rise to non-linear optical effects and the expression for P must be modifiedc
to

P - x{1)£ + X(2)£2 + x(3)£3 + ... (5,2)

where x(2) and x(3) are constants, the second and third order susceptibilities s
respectively. The second term, including x(2), is responsible for the generationr
of second harmonic radiation with a wavelength half that of the impinging);
laser light. A large x(2) represents a material which will produce a higher level;
of this frequency doubled radiation; in order to compare different materials,;
values of x(2) are often quoted relative to a standard such as potassium!
dihydrogen phosphate, KH2P04 with x(2)=l- As with x> in low^

Fig. 5.16 The structure of KTiOP04

shown as linked TiOe octahedra
and P04 tetrahedra forming
cavities containing potassium ions
(circles).
 Inorganic materials chemistry 59

mmetry materials %(2) values should be specified for the different crystal
rections.
Non-linear optical crystals are generally insulators transparent to the laser
;ht of interest. Non-linear optical effects occur in non-centrosymmetric
aterials with highly polarisable crystals which also have a high refractive
dex; a number of non-linear optic crystals develop polarisation as a result
the displacement of ions as in ferroelectrics.
Two important non-linear optic materials are LiNb03 and KTiOP04. The
ructure of lithium niobate was described in Section 4.5. This material has
high refractive index and a high polarisability as a result of the lithium ion
splacements in the ferroelectric phase and %(2), relative to KH2P04=1, is 13.
pplications of lithium niobate in optical devices include frequency doublers
id waveguides. KTiOP04, has the structure shown in Fig. 5.16, and is
instructed from linked TiOs octahedra and P04 tetrahedra. The potassium
ns occupy cavities within this framework. In the ferroelectric phase the
)tassium ions are displaced from the centres of these cavities producing a
gh dipole moment. This material also has a high refractive index. The
imbination of these properties produces a high value for %(2) in eqn 5.2, of
5, relative to KH2P04=1, and potassium titanyl phosphate is rapidly
sveloping in non-linear optic applications.

,7 Problems

1 Re03 forms the insertion compound Li0 3Re03. Describe the band structure
: this material and the expected electron conduction properties.
Fig. 5.17 Edge linked MOe
octahedra along the c direction
2 The rutile structure consists of edge sharing M06 octahedra as shown in
forming the rutile structure. Using
ig. 5.17. Using this description, which d-orbital on the metal can be used to the metal centre as the origin the
>rm a band through overlap with similar d-orbitals on neighbouring metal cartesian x, y and z axes point
oms? Would you expect Ti02 or V02 to exhibit metallic properties? towards the oxide ions as shown.
 6 Non-stoichiometry

6.1 Point defects

All crystalline materials contain a certain number of defects, that is thee

structures are not perfect. The description of unit cells used in Chapter 1 we;
based around a perfect, ordered array of atoms and molecules. However, ret;
materials do not have structures with a flawless arrangement of atoms and thl
reason for this can be illustrated by thermodynamic considerations. Considd
a crystal into which a number of defects, examples would be a misplaced iot
or a vacant site, are gradually introduced; the free energy of the system (ACC
will fall initially before rising as shown in Fig. 6.1. The reason for thn
behaviour is the competing effects of the enthalpy (AH) and the entropy (AA
Energy

of introducing the vacancy. Removing ions and molecules from a structun

will, typically, reduce the lattice energy; lattice energies are dependent upoc
the total of the electrostatic attractions between all atom pairs. Hence, thl
incorporation of defects into a structure will have an enthalpy penalty and Aa
climbs steadily as the number of defects increases. Introducing defeci:
Equilibrium randomly into the structure will produce disorder, so AS will be positive. A;
number of
defects for each defect is large at first, as the perfect lattice is strongly disrupted, bo
as more and more defects are introduced further disruption of the lattice
Fig. 6.1 Energetics of defect less marked. The overall variation in AG, as a function of the number ci
incorporation. defects, shows that there will be a certain number of defects corresponding t
the minimum in the plot.
For materials such as NaCl and MgO, with high lattice energies and thin
a large enthalpy penalty for defect formation, AH rises very rapidly and thl
minimum in AG lies at a relatively low number of defects. At root)
temperature the number of defects in NaCl is about 1 ion in 1015, though a;
the temperature rises TAS becomes more significant and the number of defeci:
increases rapidly. In other materials the lattice energy costs of introducim
defects may be much lower and the equilibrium number of defect:
consequently rather higher.
Two types of defect are commonly found in ionic inorganic solids am
because they do not affect the stoichiometry of the compound they are know.)
as intrinsic defects.

Schottky defects

Schottky defects, Fig. '6.2, are vacancy pairs: ions are removed from the;
lattice sites and, in order to maintain charge balance, this must be carried or
for cation/anion pairs. The pairs of vacancies may associated with each otha
on neighbouring sites or be quite well separated. This type of defect
commonly found at low concentrations in alkali metal halides.
 Inorganic materials chemistry 61

■ Vacancy
6.2 ■ Vacancy 6.3
• Interstitial

Figs. 6.2 and 6.3. Schottky and Frenkel defects in a rocksalt structure.

Frenkel defects

Frenkel defects, Fig. 6.3, are interstitials, that is, a displaced atom, ion or
molecule that occupies a site within the structure that is not normally filled;
the displaced species leaves behind a vacant site. In order for an ion to fill an
interstitial site it needs to be small and, because it is quite often close to ions
of the same charge, polarisable. Typical materials exhibiting Frenkel defects
are the silver halides where Ag+ forms the interstitial. Again the defects can
occur in pairs, that is, a vacancy and interstitial close to each other, or they
may be separated by larger distances.

Split interstitials

In the description of Frenkel and Schottky defects no account has been taken
of what happens to the lattice around a defect once it has been introduced. In
fact the lattice relaxes to accommodate the vacancy and/or interstitial. In the
case of an interstitial the ions surrounding the additional ion will move, in so
far as it is possible within the constraint of the crystal structure, to fit around
it. In some materials the interstitial and an atom displaced from its normal site Fig. 6.4 A split interstitial in a face
occur as a pair; this is known as a split interstitial, Fig. 6.4. centred cubic metal lattice.

6.2 Non-stoichiometry

In non-stoichiometric materials defects of the type described above occur but

to a much larger extent. In addition, the vacancies and interstitials which exist
are not formed in pairs. For example, cation vacancies are not balanced by an
equal number of anion vacancies as in Schottky defects so, for an overall
charge balance, one component of the lattice, normally a metal ion, must be
62 Non-stoichiometry

oxidised or reduced. Introduction of additional anions on to interstitial sites

in U02 to form U02+, with x=0.1 requires the uranium oxidation state to
increase from 4 to an average of 4.2+. The occurrence of non-stoichiometry
4.31 -
is, therefore, restricted to compounds containing a species which is readily
oxidised or reduced. Non-stoichiometric behaviour is most frequently found
Q/A
for transition metal compounds though it is also known for some lanthanide,
4.29 - actinide and B metal compounds.
Non-stoichiometric materials are characterised by two criteria. From a
thermodynamic standpoint, the free energy of the system depends upon both
the composition and temperature. A more useful criterion in that the behaviour
4.27 - is readily identified by experiment, is that the lattice parameter of the system
-1-1 i-r
0.94 0.90 0.86 varies smoothly as a function of composition, e.g. Fig. 6.5. Thus, as defects
x in Fel xO are introduced into a structure, the lattice parameter will change gradually
between two end members of the non-stoichiometric system. This means that
Fig. 6.5 Variation of the lattice across the full composition range of the non-stoichiometric material the same
parameter of Fe,_,0 as a function structure is adopted and that the defects are randomly distributed throughout
of x. the material. This behaviour should be contrasted with a series of phases of
different fixed stoichiometry that may be formed between two elements, e.g.
V203, V02 and V205, all having different structures and, hence, lattice
parameters.
The methods by which defects, either vacancies or interstitials, can be
introduced in to a general binary compound AB are summarised in Table 6.1.
All these types of non-stoichiometric material are known and examples of
each type are discussed in the following sections.

Table 6.1 Scheme for the formation of non-stoichiometric compounds from AB

AB

Oxidise metal Reduce metal

Excess anions Metal vacancies Excess metal Anion vacancies

AB, +, A,,B A, +,B AB,.,

U02+„ CeH2+, Fe,.„0 Zn1+jrO TiO,,

Systems in which the metal is oxidised are by far the more common and
the level of non-stoichiometry is frequently much larger. The reasons for this
are thermodynamic. Oxidation of the metal results in a much smaller enthalpy
penalty for the inclusion of defects as, with an increased metal charge and
(consequently) smaller metal ions, the lattice energy of the solid increases to
partially outweigh the energy required to form the defect. Hence, in Fig 6.1
the AH line rises less rapidly than in the case where the metal is reduced and
the equilibrium number of defects is, in general, much larger.
 Inorganic materials chemistry 63

6.3 Structures and non-stoichiometry

uo2+x

Stoichiometric U02 crystallises with the fluorite structure shown in Fig. 6.6.
U02 will react with oxygen to form the non-stoichiometric materials U02+i
with 0 < x < 0.25. U0225 or U409 has an ordered (non-random) arrangement
ef both uranium and oxygen and is thus a stoichiometric phase forming the
end member of the series. Between U02 and U409 variable amounts of
oxygen are added to the lattice, with a resultant gradual change in the lattice
parameter.

® O
Fig. 6.6 The fluorite structure of

a u U02.

Inspection of the U02 structure for large, vacant sites which could Note that the fluorite structure is
accommodate an additional oxide ion indicates a possible position at the unit face centred cubic and sites
cell centre (Vfc,V6,Vi). The immediate environment of this site shown in Fig. 6.7 equivalent to that at (Vz.Vz.Vz) occur
along the cell edges at (V^,0,0),
consists of eight existing oxide ions at the comers of a cube. The distance
(0,%,0) and (0,0,14).
from the cell centre to these oxide ions is about 2.00 A which is somewhat
shorter than twice the ionic radius of O2- indicating that even this site is too
small to accept, directly, an oxide ion.
Investigations of the structure of U02 using powder diffraction techniques
have located the oxygen positions in the unit cell of U02+x. Two observations
were made: vacancies existed on the normal oxide sublattice and two types
of additional oxide ions were located close to (V4,*4,V4) but displaced in the
110 and 111 lattice directions. Fig. 6.8. shows the two interstitial oxide
positions, termed O' and O" respectively, which, by consideration of the
ionic radii, must be associated with vacancies on the adjacent normal oxide
ion sublattice. Interstitial oxygen may, therefore, be accommodated in the
fluorite lattice but at the expense of oxide vacancies. In order for an overall
increase in oxygen stoichiometry these interstitials and vacancies must cluster Fig. 6.7 Oxide ion sublattice in the
together. One possibility is shown in Fig. 6.9 with two vacancies being U02 unit cell.
64 Non-stoichiometry

6 HSt
f- -f f
111 no
a a_/,
“
53 ?- 9 >3"
J -0A
Figs. 6.8 and 6.9 interstitial oxide --4 9
ion positions in U02 and a defect
cluster formed from them. The /
boxes are the unit-cell, oxide-ion jSe-f—
4 >rm
sublattice taken from Fig. 6.7.
1

O" # O'
r T

ft —49

6.8 6.9

associated with two O' type and two O" type interstitials. The overall I
composition of this structural block, consisting of eight U02 units ('Ug016') I
is, therefore, increased by two oxide ions giving U409. This agglomeration of'
interstitials and vacancies is termed a defect cluster and represents an i
energetically favourable way of incorporating large numbers of defects in a i
non-stoichiometric compound. Other possible clusters are possible; for •
example extension of Fig. 6.9 by an additional vacancy would permit two i
more O' type interstitials to be incorporated into the defect cluster.
The observation that the end member of the non-stoichiometric range is ;
U409 results from a structure which contains one additional oxygen in each 1
unit cell. The additional oxide ions are now regularly distributed, rather than 1
randomly as in U02+x, 0<x<0.25.
The widespread adoption of the fluorite structure by materials with the AX2
stoichiometry and its open nature leads to frequent observation of non-
stoichiometric materials similar to U02+J. Other examples include CeH2+x and
materials such as Ca^Y^F^ where the fluorite structure of CaF2 has been
doped with YF3. Calcium and yttrium share the cation positions in the fluorite
structure and the additional fluoride ions occupy similar sites to those of
oxygen in U02+r

The Fe,_xO system is probably the most extensively studied non-stoichiometric

system despite the fact that Fe^O is metastable at room temperature. This
means that the phase, formed only at high temperature, requires quenching
(cooling quickly) to room temperature in order to leave the iron and oxygen
« Fe a O atoms locked into the Fe^O structure. With increasing temperature the range
of non-stoichiometry in this system increases; at 600°C, x can only adopt
Fig. 6.10 The rocksalt structure of values to give the range Fe093O to Fe094O, at 1000°C the range becomes
•FeO'. Feo.8?0 to Fe096O.
 Inorganic materials chemistry 65

In common with most first row transition metal monoxides Fe^O adopts
the sodium chloride structure shown in Fig. 6.10, though obviously a large
number of defects need to be incorporated into the lattice to produce the
stoichiometry Fe093O. The method of incorporating the iron defects is not
simply a matter of removing iron from the normally fully occupied sublattice.
Powder diffraction measurements on Fe,_xO have shown that some iron atoms
occupy interstitial sites within the sodium chloride structure and that the level
of iron vacancies on octahedral sites is greater than would be required by the
compound stoichiometry alone. The occupied interstitial positions within the
lattice were found to be tetrahedral sites such as (14,14,14).
Magnetic measurements have shown that the tetrahedral iron atoms have
a charge of 3+. The basic mechanism by which defects are incorporated into
'Fe^O' is illustrated in Fig. 6.11. A number of vacancies are created on the
normally occupied octahedral sites but over the whole structure this number
is greater than required by the compound stoichiometry. The additional iron
required occupies tetrahedral sites adjacent to the vacancies. In order to
Fig. 6.11 A 4:1 defect cluster in
maintain charge balance the interstitial iron is oxidised to 3+ and a number
Fe,_xO.
of iron atoms surrounding the defect but on normal lattice sites must also be
oxidised to Fe3+. Note that Fe3+, as a high spin d5 ion, has no preference for
octahedral or tetrahedral sites.

• Interstitial iron
e Oxygen
s Normal iron site
■ Iron Vacancy

Fig. 6.12 The Koch-Cohen cluster of Fe,_xO.

This defect cluster with four vacancies and one interstitial is termed a 4:1
cluster and is believed to occur for low concentrations of iron vacancies, i.e.
small x in Fe,_xO. At higher defect concentrations a larger cluster consisting
of 13 vacancies and 4 interstitials illustrated in Fig. 6.12 has been suggested.
This cluster is named after its proposers as a Koch-Cohen cluster. Once again
66 Non-stoichiometry

this cluster must be surrounded by a sheath of Fe3+ ions occupying normal

octahedral sites, to provide charge balance.

Zn1+I0

As discussed above, non-stoichiometric systems in which the metal is reduced

have a much more limited range of stoichiometry and are rather rare. One
O Zn O
example of a reduced metal system is Zn1+JtO but x is very small, =10-5, and
Zn moderate values are found only at high temperatures where the formation of
i
Zn Zn defects is favoured through entropy considerations. When zinc oxide is heated
the compound loses oxygen from the surface forming Zn1+/); charge balance
requires the zinc atoms to be reduced to Zn+ or Zn°. These ions migrate from
O the surface to interstitial positions within the lattice as shown in Fig. 6.13. As
the concentration of these interstitial ions is so low, structural studies have
been unable to locate their position, but consideration of the wurtzite structure
Zn
adopted by ZnO and the larger size of these ions, Zn+/Zn°, would indicate that
they would probably occupy octahedral holes. The formation of lower charged
'/2Q, zinc ions/atoms permits an electron transfer process between the various zinc
O Zn
* centres. The energy of these transitions occurs near to the visible region of the
spectrum giving rise to the yellow colour of hot ZnO. As the number of
interstitial zinc ions decreases markedly at low temperatures the colour of
Fig. 6.13 Zn1+xO. ZnO at room temperature is white.

TiO,
Titanium monoxide shows an enormous non-stoichiometry range; at 1400°C
the full range of materials from TiO07 to TiO, 25 can be prepared. Over this
whole composition range the material adopts structures based on that of
sodium chloride. At low oxygen content vacancies exist on the oxide
sublattice, whilst in TiOj 25 it is the titanium sublattice that has vacancies and

O Ti O O Ti O Ti O

Ti Ti O Ti O Ti Ti

Fig. 6.14 Vacancy distribution in O Ti O

O Ti O Ti O
■TiO'.

Ti O Ti Ti O Ti O ji

O O Ti o Ti O O

Ti O Ti O Ti

10
 Inorganic materials chemistry 67

this compound can be rewritten as Ti0 gO. Even the 1:1 stoichiometry material,
TiO, has a structure containing defects with equal numbers of vacancies on
both the titanium and oxygen sublattices.
The structure of the limiting composition TiO07 has not been determined
in detail and indeed the structural chemistry of the TiOx system is very
complex. At high temperatures, above 900°C, the vacancies in TiO are
randomly distributed and the material is cubic. Below 900°C ordering of the
vacancies can occur and in 'TiO' 50% of cations and anions are alternately
absent in every third 110 plane. Fig. 6.14.

6.4 Non-stoichiometry in more complex systems

The four non-stoichiometric systems discussed above in terms of their

structures, illustrate the major ways in which large compositional variations
can be incorporated into materials through defects and interstitials. Non¬
stoichiometry is a very widespread phenomenon in solid state oxide, sulphide
and halide chemistry, but it is also a feature of intercalation chemistry by the
nature of the synthesis process.
In more complex oxides, for example ternary phases, and insertion
compounds, a detailed level of structural information, as discussed above for
a number of binary systems, is often not available and the structures of
compounds are often described in terms of an average distribution of ions or
molecules. To illustrate this a good example is provided by the ternary oxide
system Sr3Fe207_y, 0<y<l.
The structure of Sr3Fe207 is shown in Fig. 6.15 and can be considered to
be blocks of perovskite separated by layers of SrO with the rocksalt structure.
The oxygen stoichiometry of this material may be reduced drastically by
gradual removal of oxide ions from the site within the perovskite block.
Complete removal of oxygen from this site results in the composition
Sr3Fe206 with reduction of the iron, from Fe4+ in Sr3Fe207, to Fe3+. The
coordination geometry of the iron decreases from six in an octahedral
arrangement to five forming a square based pyramid. For the intermediate
compositions in this system, Sr3Fe207_r 0<y<l, the structure would be
described in terms of partial filling of the oxide ion site between the two iron
atoms. Throughout the structure there would be a random distribution of
linked Fe06 octahedra and Fe05 square pyramids.
Similar changes in the oxygen content of materials, through partial filling
of a site, is important in the chemistry of the high temperature
superconductors discussed in Chapter 8.

6.5 Elimination of defects, crystallographic shear

In the structural chemistry of U02+x and Fe^O the formation of defect

clusters has been observed though these clusters are still randomly distributed
throughout the structure over the full stoichiometry range. Some materials
68 Non-stoichiometry

Fig. 6.15 The structures of

Sr3Fe207 and Sr3Fe206. FeOe
octahedra and FeOs square
pyramids are shown as polyhedral
outlines and the Sr ions as circles.

assimilate a large number of defects into their structures by incorporating

them as an embedded portion of a different structure. This effectively
eliminates the defects, forming a new ordered array and generating a more
complex structure. This process is known as crystallographic shear and is
illustrated schematically in Fig. 6.16.
Crystallographic shear is well characterised in systems derived from comer
sharing octahedra and the defects are eliminated by increasing the
coordination number of oxygen. The Re03 structure, Fig. 4.3, is formed from
octahedra sharing all corners. Oxygen deficiency introduced into this structure
can be accommodated through the crystallographic shear process as illustrated
in Fig. 6.17. Removal of oxygen (labelled 6 and 7) along the line of corner
sharing octahedra is followed by a translation along and perpendicular to the
line of missing oxygen by the lower set of octahedra.
The coordination number of the oxide ions along the crystallographic shear
plane (e.g. the oxide ion labelled 9) is increased to three while that of the
metal atom is maintained at six, consistent with the decrease in oxygen level.
The method of incorporation of this crystallographic shear into the full Re03
lattice is illustrated in Fig. 6.18.

Fig. 6.16 Rearrangement of

defects in crystallographic shear.

Fig. 6.17 Oxygen defect

elimination in a portion of the Re03
structure.
 Inorganic materials chemistry 69

Fig. 6.18 Crystallographic shear

along 310 in the ReOa
structure.The left hand side shows
the perfect Re03 structure prior to
elimination of defects along the
310 plane (diagonal line).The right
hand side shows the sheared
structure produced with edge
sharing octahedra.

The overall stoichiometry of the resultant compound depends upon the

frequency of occurrence of the crystallographic shear plane throughout the
structure: indeed by changing this frequency the stoichiometry can be varied.
If the crystallographic shear shown in Fig. 6.18 is separated by ten normal
M06 octahedra the compound stoichiometry becomes M10O29 and this
structure is adopted by W20O58. A separation of nine normal MOtl octahedral
blocks would result in the stoichiometry M9026.

6.6 Intercalation compounds — staging

Whilst many intercalation compounds are truly non-stoichiometric with a

random distribution of intercalated species over a proportion of sites, the
insertion compounds of some layer systems exhibit a phenomenon known as
staging. The potassium intercalates of graphite clearly illustrate this behaviour.
In CgK, the highest potassium intercalate or first stage compound, the
potassium ions occur between all the carbon layers on sites shown in Fig.
6.19. u u
The other compounds in the potassium-graphite system have the
stoichiometries C24K , C36K and C48K. In C24K rather than removing a portion
of the potassium ions from all the layers, alternate potassium layers are
removed completely and one third of the potassium ions from the remaining
strata. The potassium ions are disordered within the layer and this compound
is known as a second stage intercalate, Fig. 6.19. The C36K and C48K
compounds, third and fourth stage intercalates, Fig. 6.19, have potassium ions
every third and fourth graphite layer respectively. This behaviour is known as
staging and is clearly demonstrated in the powder X-ray diffraction patterns.
The layer structure of graphite leads to a strong 001 reflection which is
readily observed in the powder X-ray diffraction pattern. The position of the
001 reflection for some of the staged compounds is shown in Fig. 6.20 and
corresponds to the c lattice parameter marked in Fig. 6.19.
70 Non-stoichiometry

Fig. 6.19 Staging in CnK. The c lattice parameter is marked for each phase.

03
0-
Fig. 6.20 Powder X-ray diffraction o

patterns from graphite, C8K and

C24K.

6.7 Problems

6.1 The 111 reflection from CeH2 (?i=1.54 A) occurs at 27.96°. Taking the
ionic radius of H" as 2.08 A calculate the dimensions of the largest sphere
which could occupy the (V4,V6,Vi) site without removal of hydride ions from
their normal sites. The powder X-ray diffraction pattern of CeH215 is identical
to that of CeH2 except for a shift of peak positions to slightly lower 20 values.
Explain.

6.2 How would you expect the lattice parameters of Sr3Fe207_v to vary as a
function of y?
7 Zeolites, intercalation in layer
materials and solid
electrolytes
0

The preceding chapters have dealt mainly with the structure and properties of
metal oxides, an extremely important group of compounds which illustrate
many of the aspects of inorganic solid state chemistry. This chapter provides
an introduction to some other widely studied inorganic materials which have
significant applications and properties.

7.1 Zeolites

The zeolites are a group of compounds, many of which are naturally occurring
minerals, named after their property of evolving water when heated ( Greek,
zeo to boil, lith stone). These materials are widely used for their ion exchange,
absorption and catalytic properties.
The compounds are characterised by open structures which may incorporate
a range of small inorganic and organic species. The frameworks forming the
channels and cavities are constructed from linked tetrahedra and many
elements, which form T04 groups in their structural chemistry, can take the
part of the building blocks in zeolites. Examples include A104, Si04, P04,
Be04, Ga04, Ge04 and Zn04. The most common zeolitic materials are based
on silicon and aluminium M04 tetrahedra linked together. The different ways
in which these tetrahedra are connected in three dimensional space give rise
to the multitude of different zeolites. At present over 200 aluminosilicate
framework structures are known; about 40 of these are naturally occurring
minerals.
A general formula for zeolites may be obtained by starting from pure silica Framework
which in its various structures, quartz being one example, consists of Si04
tetrahedra sharing all vertices. Replacement of some of the Si04 tetrahedra by
A104 tetrahedra can occur provided that charge balance is maintained by
simultaneously incorporating cations into the structure, that is 'Si02' is
replaced by 'MA102' or 'M05AlO2' where M is a mono or divalent cation.
The cations introduced alongside the aluminate tetrahedra are frequently
hydrated leading to a general formula for zeolites of

{[ n,- [ ™h2o ]} {[ A102 ]x [ Si02

with the three components, cavity cations, absorbate (water) and framework.
/N Solvent # Cation
The final {} grouping describes the tetrahedral species forming the zeolite molecule
framework whilst the first large bracket describes the species present in the
cavities. In this description the absorbed species, coordinating to the cation, Fig. 7.1 Schematic of the
is water, though as will become apparent, other small molecules are readily components of a zeolite.
72 Zeolites, intercalation in layer materials and solid electrolytes

Fig. 7.2 Four, six and eight

membered rings formed from
linking tetrahedra.

absorbed into the zeolite pores.

Fig. 7.1 can be used to represent schematically a zeolite with the cavities
and pores, formed from the tetrahedra, containing the hydrated cations.
Following dehydration of a zeolite, readily accomplished by heating under
The nomenclature of zeolites is vacuum, the cations migrate to the edges of the cavities and channels,
rather unsystematic. Some interacting more strongly with the framework oxygen atoms. The cavities in
materials, e.g. sodalite and
zeolites, particularly once dehydrated, may absorb small molecules other than
faujasite, are named after minerals,
others e.g. ZSM-5 and VPI-5 were water. The larger the cavities and pores formed by the framework the larger
named by researchers during a the molecules which may be absorbed inside the zeolite.
programme of work to synthesise
new frameworks. Thus ZSM was
the fifth product of the Zeolite
7.2 Zeolite structures
Socony-Mobil project. Zeolites are
often described in the form M- The linking of the T04 tetrahedra may occur in numerous ways to produce the
[zeolite] where M refers to the huge variety of zeolite structures, but a number of structural features are
cation in the particular [zeolite]
common to these materials. Linking of the tetrahedral species may take place
framework.

Fig. 7.3 A sodalite or R cage and the sodalite structure where the 13 cages share six-
membered rings. Each line connects the silicon/aluminium centres and one six and one
four membered ring, corresponding to those in Fig. 7.2, are outlined.
 Inorganic materials chemistry 73

to form four, six and eight membered rings as shown in Fig. 7.2 and these
units are then used to form the three dimensional zeolite structures. The
wealth of zeolite structures is derived not only from the multitude of ways in
Note, the framework structures of
which the tetrahedra can be linked, but also from the flexibility of the T-O-T zeolites are frequently drawn
bond angle which can assume values in the extensive range 120-180°. showing just the tetrahedral cation
Figure 7.3 shows a B or sodalite cage produced from four and six centres and connections between
membered rings. This cage frequently occurs itself as a building block in the them. Hence in Fig. 7.3 each line
represents T-T. The oxygen atoms
^structures of other zeolites such as zeolite A, Fig. 7.4a, faujasite, Fig. 7.4b
which normally lie slightly off this
and VPI-5, Fig. 7.4c. In this series of materials the sodalite cages are line are omitted.
separated by increasing distances. In sodalite the B cages form a body centred
array sharing six membered rings, in zeolite A the B cages are separated by
four membered rings, in faujasite by six membered rings and in VPI-5 by
double four membered rings. This has the effect of opening up the structures

Fig. 7.4 Portions of the framework structures of zeolite A (a), faujasite (b) and VPI-5 (c)
showing the manner of 6 cage connection and maximum pore size.
74 Zeolites, intercalation in layer materials and solid electrolytes

Fig. 7.5 Channel dimensions as a

function of ring size. The aperture
sizes are determined after allowing
for the van der Waals radius of the 410 pm
oxygen atoms around the ring. 540 pm 740 pm

and increasing the sizes of the cavities and pores. Hence, in the series
sodalite, zeolite-A, ZSM-5, faujasite, VPI-5 the largest channel diameter
increases as 260 pm, 410 pm, 540 pm, 740 pm and 1020 pm (2.6-10.2 A) as
the number of tetrahedral units forming the outside of the channel increases
from 6 through 8, 10 and 12 to 18, Fig. 7.5.
In materials such as zeolite A and faujasite, which are cubic, the channels
run through the crystallites in each of the lattice directions. Other zeolites, for
example ZSM-5, Fig. 7.6, are of lower symmetry, in this case orthorhombic,
and large channels run parallel to the b lattice parameter and are intersected
by smaller tunnels.

7.3 The properties of zeolites

Absorption

The open nature of the zeolite structures allows small molecules to be

absorbed into their structures; the size and shape of molecules absorbed will
depend upon the geometry of the pores. Zeolite A with relatively small pores
can absorb molecules such as water and oxygen but larger species, for
example ethanol, are barred from entering the cavities. This is illustrated in
Fig. 7.7. Note that the type of cation present in the pores can also affect the
dimensions of the molecules which can be absorbed. Replacement of sodium

Fig. 7.6 The structure of ZSM-5

viewed down the large pores
formed from the 10 membered
rings. These large channels are
intersected by smaller channels
running in the ab plane.
 Inorganic materials chemistry 75

in zeolite A by the smaller calcium increases the effective pore dimension

allowing methane to be incorporated.
The ability of an aluminosilicate zeolite to absorb water is related to the
nature of the ions forming the framework. With a high number of aluminium
atoms in the framework and, therefore, a correspondingly high concentration
of charge-compensating cations in the zeolite channels, the structures are
highly hydrophillic. Materials, such as sodium-zeolite A, with a silicon to
aluminium ratio of 1:1 and a high level of sodium ions in the cavities are
widely used to dry gases and solvents. The dehydrating capacity may be
regenerated periodically by heating the zeolite and driving off the absorbed
water.
Materials such as silicalite, a derivative of ZSM-5 where the framework
consists almost totally of Si04 tetrahedra, have few cations in the cavities and
are hydrophobic. These materials readily absorb non-polar and weakly polar
molecules into their cavities.

Ion exchange

The cavity cations in zeolites interact weakly with the framework and may,
therefore, undergo ion exchange reactions readily at room temperature. The
Fig. 7.7 Pore sizes and molecules
sodium ions in Na-Zeolite A rapidly exchange with calcium in aqueous
which may be absorbed for Na-
solution zeolite A, Ca-zeolite A and Na-
faujasite.
Na-Zeolite A + ViCa2+(aq) -» Ca^-Zeolite A + Na+(aq) (7.1)

The particular ion exchange characteristics of a zeolite are determined by

the sizes of the cages/pores and the coordination environments present within
the zeolite. Ion exchange processes are widely used in water softeners and the
new 'micro' detergents. In the latter, zeolite A has replaced the
polyphosphates previously used, with the environmental advantages of
decomposition to soil-like minerals.

Catalysis
The acid derivatives of zeolites, H-zeolite, are excellent catalysts and are
widely used industrially. This form of the zeolite may be obtained by direct -S-. + H H Al
*P \
ion exchange with acids. More often, because many zeolite frameworks are
Al ) H OS'
slowly attacked by aqueous acids, they are obtained by exchange with \ +di /
Sr^^AI-
ammonium ions followed by heating to 500°C, driving off ammonia, and
\ I
leaving the proton

NHl 500° C
Na-zeolite - NH^zeolite - H-zeolite -sSo
! \ fo Al"
o +Si
The catalytically active acid zeolite can either be in the Bronsted form with
protons attached to the framework tetrahedron (Fig.7.8) or Lewis form sr.° m -

obtained by dehydration of the Bronsted acid.

The acid zeolites, once they have absorbed molecules into the cavities, Fig. 7.8 Bronsted and Lewis acid
catalyse reactions typical of very strong acids. The major reaction types are forms of a zeolite.
76 Zeolites, intercalation in layer materials and solid electrolytes

dehydrations and rearrangements. A special feature of zeolites which makes

them such superb catalysts is their shape selectivity. Zeolites are crystalline
materials with controlled channel geometry and fixed environments for the
active sites. The shape selectivity may arise in three ways: reactant selectivity,
product selectivity and, probably of lesser importance, transition state
selectivity. Reactant selectivity arises from the ability of only certain
molecules to be absorbed into the zeolite cavities and thus reach the active
acid sites. Product selectivity is derived from the fact that only certain
products are of the correct dimension to escape rapidly from the zeolite along
the channels once they have been formed. This is illustrated for the
isomerisation reactions of dimethylbenzenes in Fig.7.9. The 10 ring channels
of ZSM-5, with a dimension =5.4 A, allow rapid diffusion of 1,4-
dimethylbenzene, but not the 1,2 and 1,3 analogues; the diffusivity of the 1,4
derivative is about 104 times greater than its isomers. This difference is a
result of the effective diameters of these species; the 1,2 and 1,3 placements
of the methyl groups make these molecules about 0.6 A wider than the 1,4
arrangement and diffusion through the 10 rings is hindered. Once a mixture
of dimethylbenzenes has entered the ZSM-5 structure, protonation of the
benzene ring allows migration of the methyl groups around the ring,
equilibrating the isomers. However, the greater mobility of the 1,4
dimethylbenzene arrangement permits this isomer to escape from the zeolite
while the other isomers are likely to undergo further protonation and
transformation.
Transition state selectivity relies upon the fact that certain intermediates,
which are required to be formed during a chemical pathway at the active site,
will not fit in the channel/cavity; such a reaction process is barred from
occurring and the reaction will proceed along a different route to a different
product.

Fig. 7.9 1,4-Dimethylbenzene inside ZSM-5 channels and a comparison of the

dimethylbenzene geometries.
 Inorganic materials chemistry 77

7.4 Pillared clays

The largest pore sizes synthesised in zeolites are of the order of 10 A. In an

attempt to increase this further, and allow larger molecules to be absorbed into
inorganic structures, chemists have turned to pillared clays.
Clays such as bentonite and montmorillonite have layer structures such as
those shown in Fig. 7.10. The layers are constructed of vertex and edge
^sharing octahedra and tetrahedra. Atoms normally forming the layers are
silicon and aluminium plus small mono- and divalent species such as
magnesium and lithium. As with zeolites this framework layer has an overall

Fig. 7.10 The structure of a clay

showing layers formed from linked
octahedra and tetrahedra and
separated by cations.

negative charge and in clays this is balanced by incorporation of cations,

typically alkali metals, between the layers. These inter-layer cations can be
readily ion-exchanged.
In the pillaring of clays the species exchanged into the inter-lamellar region
is selected for size. Ions such as alkylammonium ions and polynuclear
hydroxy-metal ions may replace the alkali metal as shown schematically in
Figs. 7.11a/b. The most widely used pillaring species are of the polynuclear
hydroxide type and include Al1304(0H)2g3+, Zr4(OH)16_„"+ and Si80,2(0H)8; the
former consists of an A104 tetrahedron surrounded by octahedrally

(b) (c)
(a)
78 Zeolites, intercalation in layer materials and solid electrolytes

coordinated aluminium. This pillaring reaction is readily followed by powder

X-ray diffraction; expansion of the inter-lamellar spacing, corresponding to the
c lattice parameter, results in a marked shift of the 001 reflection to a lower
20 value.
Once an ion such as A11304(0H)283+ has been incorporated between the
layers, heating the modified clay results in dehydration and linking of the ion
to the layers, Fig. 7.11c. The resulting product is a pillared clay which has
excellent thermal stability to at least 500°C. The expanded inter-layer region
can now absorb large molecules in the same way as zeolites. However,
because the distribution of the pillaring ions between the layers is difficult to
control, the pillared clay structures are less regular than zeolites. Despite this,
pillared clays have been widely surveyed for their potential as catalysts and
they act in a similar way to zeolites promoting isomerisation and dehydration.

7.5 Two-dimensional intercalation chemistry

Intercalation between the layers of graphite has been described in Section 3.7
but graphite is only one member of a wide range of compounds which have
layered structures and are, therefore, able to incorporate guest species. A
number of these materials are summarised in Table 7.1. The most widely
studied system has been that of the intercalation compounds of transition
metal disulphides.
Intercalation reactions of metal disulphide layer compounds generally
involve a redox process where the metal sulphide layer is reduced and the
intercalating species oxidised during the reaction. Hence, many species which
are readily reduced or can donate electrons can be inserted between the
sulphide layers; this includes alkali metals, amines and electron rich transition
metal complexes.

Alkali metal insertion in transition metal sulphides

Lithium can be incorporated into TaS2 over the full range Li^TaS2 0<cc<1.0
occupying octahedrally coordinated sites between the layers; as x increases

Table 7.1 Examples of layer compounds which undergo intercalation reactions

Layer compound Layer type

Graphite, oxychlorides e.g. FeOCI Uncharged layers

Transition metal disuphides e.g. TaS2

LiCo02. Clays: montmorillonite, hectorite. Negatively charged layers separated by cations or with
Complex titanates and niobates e.g. LiNbTi05. pendant protons
Zirconium hydrogenphosphate Zr(HP04)2.nH20

Hydrotalcites, brucite Positively charge layers separated by anions

e.g Zn2Fe(0H)6.(C03)V4.nH20
 Inorganic materials chemistry 79

he c cell parameter which is perpendicular to the layers increases smoothly

xom 5.7 A to 6.2 A (57-62 pm). For the other alkali metals only certain
impositions, e.g. K^TaSj, Ko18TaS2 and KxTaS2 0.3 < x < 1.0, can be
prepared as single phases; this behaviour is related to staging, Section 6.6, and
probably results from the energy requirement to overcome the van der Waals
nteractions between layers. The Stage IV compound K01TaS2 results from
filling of sites between every fourth layer. The reason for the staging
behaviour seems to be energetic; partial filling of the sites between all layers
would have an energy penalty associated with breaking the van der Waals
interactions between all layers instead of one in every four as for the staged
material.
The alkali metal intercalates of TaS2 may be hydrated by subsequent
exposure to water, forming materials such as Na0 5(H2O)„TaS2 where hydrated
nations exist between the metal sulphide layers. Other solvents may also be
used to generate materials such as NaDMSO.TaS2, DMSO = (CH3)2SO.
TTTTTt
i i I i * i
Amine intercalation
TTTTTT
The intercalation of primary amines into TaS2 has been achieved for a wide i I I i i i
•'--
range of chain lengths from methylamine to C18H37NH2, in the latter case
forming (C18H37NH2)066.TaS2. In this compound the amine groups exist as a
bilayer between the metal sulphide strata, as shown schematically in Fig.7.12. Fig. 7.12 Long chain amine
As the number of carbon atoms in the amines increases then the inter-layer intercalate of a metal disulphide.
separation increases smoothly.

Organometallic intercalation

Intercalation into TaS2 generally occurs through a redox process and, provided
an organometallic species is readily reduced, incorporation of the resulting
cation between the metal sulphide layers may be achieved. Cobaltocene,
Co(r|-C H )2, as a nineteen electron compound is readily oxidised, and forms
5 5

an intercalate of the stoichiometry TaS2.[ Co(r|-C5H5)2]025 by direct reaction

in a sealed tube, Fig. 7.13. The eighteen electron compound ferrocene is,
however, much more difficult to oxidise and is not incorporated in tantalum
disulphide though other, more oxidising, layer compounds such as FeOCl may
act as hosts for this compound. Another organometallic which has been Fig. 7.13 Cobaltocene intercalated
inserted into layer disuphides is dibenzenechromium, Cr(r|-C6H6)2. in TaS2.

7.6 Fast ion conduction in solids

The diffusion of ions through solids is generally very slow, as is apparent in

the slow reaction rates of solids (Section 3.1). This is a result of a lack of
pathways in solid structures along which ions can migrate; for example for a
sodium ion to diffuse through the perfect sodium chloride structure it would
have to move through the solid between the sodium and chloride ions on their
fixed lattice sites using interstitial positions. Such a migration would have a
high energy penalty in disrupting the lattice. One possible pathway for ion
80 Zeolites, intercalation in layer materials and solid electrolytes

diffusion through a material with the sodium chloride structure would be by

making use of defects in the structure such as Schottky defects (Section 6.2).
Sodium ions may diffuse through the structure using the vacancies, Fig. 7.14.
As the number of these defects increases so the rate of ion diffusion increases
and just below its melting point sodium chloride is a reasonable ion
conductor, with a conductivity of 10'3 £2-1cm-1. Similar behaviour may be
found with anions and Zr02 is a good conductor of oxide ions above 600°C.
The diffusion of ions through solids is thus facilitated by the presence of
suitable vacant sites to which a diffusing ion may migrate. In some materials
the presence of a large number of suitable sites within the structure for ion
migration, even at low temperatures, allows facile diffusion of ions and these
materials are known as fast ion conductors. Their conductivities reach values
in the range 10-3—10* Q“lcirf1 comparable to ionic solutions. Typical structural
and chemical features exhibited by fast ion conductors are vacant sites to
which ions may migrate and small cations with a low charge; these interact
weakly, in electrostatic terms, with the rest of the structure and form low
Fig. 7.14 An ion conduction
pathway in NaCI using a Schottky coordination number environments. Some of the best ion conductors are,
defect. Once one ion has migrated therefore, based on silver and lithium.
to the empty site the vacancy left One example of a fast ion conductor is a-Agl. At 146°C silver iodide
behind may be filled by a different adopts a structure based on a body centred array of iodide ions shown in
sodium ion. Ion conduction thus
Fig.7.15. The silver atoms are able to occupy a range of sites throughout this
occurs by a series of hops.
iodide matrix close to the face centres and movement between these positions
is facile.
Some of the best fast ion conductors consist of a stable framework
constructed from octahedral or tetrahedral metal oxo species which form
channels or pathways along which cations can migrate. These channels are
somewhat smaller than those in zeolites as only small ions, typically alkali
metals and silver, need pass along them. Examples of fast ion conductors of
this type are NASICON (Na Superionic CONductor) which is Na3Zr2PSi2012,
and Li14ZnGe40,6 (a lithium ion conductor) and a number of derivatives of the
two dimensional material 6-alumina.
The NASICON structure, Fig. 7.16, has channels, formed from Zr06
octahedra and Si04 tetrahedra, containing the sodium ions. There are more
possible sites for sodium to occupy in the channels than sodium ions present
and by hopping between these positions sodium ion diffusion along the
channels is possible.
The 6-aluminas are a class of compounds of the general formula
M20.«A1203 where n=8-ll and M is a very wide range of cations which
includes Na+, K+, Rb+, Ag+, NH4+, H30+, NO, Mg+ and Ln3+. The structure,
Fig. 7.17, consists of blocks which have structures similar to that of the
spinels separated by oxide ion deficient regions in which reside the cations,
M. The sections containing the M cations provide a large number of possible
Fig. 7.15 The iodide ion framework
of a-Agl and the regions occupied sites and migration of these ions is rapid within these layers. At 300°C the
by silver ions. conductivity of sodium 6-alumina is about 0.1 Q-'cnr1.
 Inorganic materials chemistry 81

Conduction Channels

Fig. 7.16 The structure of

NASICON showing the channels
containing the mobile sodium ions.

Applications of fast ion conductors include batteries, where the solid cation
conductor acts as the electrolyte, fuel cells and oxygen gas sensors.

7.7 Problems
i—i i i i i i i—r
7.1 The powder diffraction pattern collected from ZSM-5 (A=1.54 A, 154 pm) Spinel Block
has reflections at 7.868°, 8.767° and 22.110°, which can be indexed as the 5 Ft
101, 200 and 050 reflections respectively. Calculate the orthorhombic unit cell C
} m c
) ® c5 9
parameters. Using Fig. 7.6, estimate the size of the 10 membered ring L T 1 i i i i
ZC a . i II II
.1,1,1, ill ii
apertures in ZSM-5.
+ 9 c
> 5 9
7.2 The 001 reflection collected from TaS2.C18H35NH2 (X,=1.54 A, 154 pm) is 1 1 1 ■ ■ ~i i i r
1 1 1 1 III
at 20 = 1.57° whilst the corresponding reflection in TaS2 is at 17.58°. | 11 I I i, i, i; i
Calculate a length for this amine chain assuming that they are perpendicular
to the sulphide layers. How would the position of the 001 reflection change Fig. 7.17 Schematic of the B-
if the amine groups were tilted at an angle relative to the TaS2 layers? alumina structure.
 8 Some recent developments
in inorganic materials
chemistry

Many of the recent advances in inorganic chemistry have come in the area of
new materials. The chemistry of high temperature superconducting phases, in
terms of the synthesis and structure of these compounds, lies within the area
of this text. More recently the structural chemistry of Cm in its ionic
compounds is also relevant, though the chemistry of this species is developing
rapidly in many areas

8.1 High temperature superconductors

Superconductivity was first observed by K.Onnes in 1911; on cooling a

sample of mercury below 4.2 K the resistivity of the metal suddenly decreased
to an immeasurably small value, Fig. 8.1. Since then many metals and alloys
have been found which have zero resistance below a certain critical
Fig. 8.1 Resistiance of a sample of temperature, Tc. The progress in terms of raising Tc, in metals and alloys since
mercury as a function of 1911, is summarised in Fig. 8.2.
temperature.
Above Tc superconducting materials show a finite resistance in common
with other compounds. This electrical resistance results from the interaction
between lattice vibrations and the conduction electrons as they move through
the structure. This interaction or scattering of the conduction electrons
strengthens as the atomic vibrations increase in amplitude, that is, as the
temperature of the system increases. Flence, the resistance of metals rises as
they are heated. In superconducting metals and alloys it is believed that the
electrons move through the lattice in a concerted fashion with the lattice
25- Nb3Ge
vibrations, resulting in no electron scattering and zero resistance.
□
Not all superconducting materials are pure elemental metals and their
20
a NbjSn
alloys. By 1985, zero resistance at low temperatures had been observed in
NbN □ many compounds, for example metal oxides Li2Ti04, Tc = 13.7 K (spinel type
15-

structure), sulphides PbMo6S8, T = 15.2 K and organic metals formed from

10 charge transfer compounds (up to 13 K); all of these systems exhibit metal¬
□
Pb □ Nb like behaviour in terms of their conductivity above their Tc. In general though
5
□ the critical temperatures of these compound types are fairly low, mostly below
Hg
i-r
10 K.
1900 <920 1940 1960 1980 In 1986, J.G.Bednorz and K.A.Miiller published work describing the
observation of anomalously low resistance in a mixed metal oxide system
Fig. 8.2 Increases in Tc since
1911.
containing barium, lanthanum and copper. The compound was later shown to
be La2JBaxCu04, with a critical temperature of 35 K, and Bednorz and Muller
 Inorganic materials chemistry 83

were awarded the Nobel Prize for their discovery. Since then progress in
developing new inorganic materials with higher Te’s has been extremely rapid.
These new materials are all complex copper oxides and, at the time of
writing, the critical temperature in bulk materials has been increased to 134
K with reports of even higher Tc’s under high pressures and in thin films. The
structures and chemistry of the most important of these superconducting
phases are discussed in the following sections.

La2_xMxCu04, M=Ba, Sr

La2Cu04 has a structure similar to that of K2NiF2 described in Section 4.7

though the elongated Cu06 octahedra are tilted relative to the 001 plane of the
unit cell. Partial substitution of the trivalent lanthanum by divalent barium,
whilst maintaining the oxygen content, produces La, 8Ba02CuO4, Fig. 8.3,
which has the perfect K2NiF4 structure. This material has a Tc of 35 K, the
Fig. 8.3 The structure of
strontium doped analogue, La, 85Sr015Cu04, has a Tc of 40 K. La, gBaojCuCV

YBa2Cu307_5, 0 < 5 < 1

lOO

The non-stoichiometric YBa2Cu307_5 system has been the most widely studied
of the superconducting materials. This is a result of its ease of synthesis and
moderately high critical temperature, 93 K, which lies reasonably above the
boiling point of nitrogen (77 K). One complicating factor in this system has
been the control of the oxygen stoichiometry which is important in
determining the critical temperature. The best superconducting properties are
observed when 5=0, YBa2Cu3O69_70 with Tc = 93 K, as 5 increases then Tc 0.0 0.4 0.8
8 in YBa2Cu30 7_5
falls rapidly as shown in Fig. 8.4. In the synthesis of YBa2Cu307 it is,
therefore, important to control the oxygen stoichiometry which requires a two Fig. 8.4 Variation of Tc in
stage reaction process. The direct high temperature reaction of BaC03, Y203 YBa2Cu307_5 with 8.
and CuO at 940°C over a period of days followed by quenching to room

® Yttrium

(d)

Fig. 8.5 Derivation of the YBa2Cu307 structure from a simple perovskite cube.
84 Some recent developments in inorganic materials chemistry

temperature gives rise to material deficient in oxygen with a stoichiometry of

about YBa2Cu3067. A material with a similar stoichiometry may be prepared
using the gel methods described in Chapter 3. In order to produce the
optimum oxygen content these oxygen deficient materials must be annealed
in pure oxygen, at temperatures of 400-500°C, followed by slow cooling to
room temperature. Under these conditions the material takes up additional
oxygen and 5 falls close to zero.
The structure of YBa2Cu307 is derived from that of perovskite. Stacking
three perovskite unit cells, of the stoichiometry AB03, directly on top of each
other, as in Fig. 8.5a, gives rise to a material of stoichiometry A3B309 (3 x
AB03). If the A type cations are replaced by 2 x Ba and 1 x Y in the
sequence Ba-Y-Ba-Ba-Y-Ba-Ba-Y in the tripled perovskite, and the B cations
are designated as copper then the compound stoichiometry, shown in Fig.
8.5b, becomes YBa2Cu309. Oxygen is then removed from the (OOV ) site in the
2

structure adjacent to yttrium and the site in the basal plane at (V2OO) giving the
YBa2Cu307 structure shown in Fig. 8.5d. The cell parameters are based on
those of the original perovskite cube with a~390 pm in YBa2Cu307, c is
approximately treble a and, because of the selective removal of oxygen from
the sites in the basal plane, a and b are different. The crystal system of
YBa2Cu307 is thus orthorhombic with the lattice constants a= 385 pm, b-388
pm and c = 1140 pm.
Consideration of Fig. 8.5d shows that YBa2Cu307 contains copper in two
different coordination environments, those with fractional coordinates near
(0,0,Vs) in square pyramidal sites and those at the origin in square planar sites.
A polyhedral description of the structure using the copper coordination
geometry, Fig. 8.6, makes this clearer. The linking of these sites leads to them
being referred to as forming chains and planes.

Fig. 8.6 The structure of

YBa2Cu307 showing the copper
oxygen polyhedra. Square planar
sites link together forming one
dimensional chains along the a
axis whilst square pyramidal sites
link together to form a two-
dimensional layer.
 Inorganic materials chemistry 85

o T1
2234

• Ba

• Ca
2223
• Cu

Q ° jq ®

O
?%2?
O

s'
©® 0°©3°
2201
©V ?%2?
^0_„o
o°cpO ®
Cu°

Cu°
CuO

•Pt&r Cu°
o°SP° CuO ©
CuO © •c^f CuO
CuO

3°CP° •C&f CuO

8°^
0 o-0
0 o-0

cuo
o^o0 no
TIO

0%° no
Ba
no Ba
Ba
Ca

3CP @ - o pu Ca Ca

ocfb o^o Ca

ig. 8.8 The structures of T^BajCa^CuA™* with n = 1, 2, 3 and 4.

’he oxygen stoichiometry in YBa2Cu307_5 may be varied with 0<8< 1 and in

5rms of the structure this is achieved by removing oxygen from the other
iasal plane site, (0,1/2,0). Complete removal of this oxygen yields YBa2Cu306,
?ig. 8.7. The coordination of the 'chain sites' drops to linear, typical of Cu+.

ri2Ba2Can_1Cun04n+2, n = 1, 2, 3, 4

dost of the highest Tc’s observed in bulk materials have been found in the
ystem of thallium barium calcium cuprates. A series of materials with
liffering numbers of Cu02 layers have been prepared in the homologous
eries Tl2Ba2Can_,Cun04n+2 with n = 1, 2, 3, 4 and the structures of these Fig. 8.7 The structure of
naterials are shown in Fig. 8.8. The structure of Tl^CaCuA (2212) YBa2Cu306.

onsists of two layers of Cu02 square nets separated by a double

hallium-oxygen layer, T1202. The copper^alcium-oxygen layer may be
:onsidered as being derived from a perovskite block, with oxygen vacancies,
vhilst the T1202 layer has an arrangement of ions similar to that of a portion
,f the rocksalt (NaCl) structure. As n increases in the formula
n Ra Ca ,Cu O, , then additional Cu02 layers are incorporated into the
Picture,""thus" Tl2Ba2Ca2Cu3O10 and Tl2Ba2Ca3Cu4012 have three and four
:uprate layers respectively. All these materials have tetragonal unit cells with
i=3.8 A (380 pm) but different c parameters; as the number of Cu02 layers
ncreases so c increases.
A series of materials with only a single thallium-oxygen layer separating
he Cu02 sheets is also known. For example, the structure of TlBa2Ca2Cu3Q9
86 Some recent developments in inorganic materials chemistry

may be derived from that of Tl2Ba2Ca2Cu3OI0 by removal of one of the TIC

layers.
The critical temperature behaviour of the Tl2Ba2Can_,Cun04n+2 system is
very complex and depends upon the exact compound stoichiometry, synthesis
conditions and subsequent annealing treatments in different atmospheres. One
trend is apparent, Tc rises as n increases from 1 to 3, 80 K in Tl2Ba2Cu06
105 K in Tl2Ba2CaCu2Og and 125 K in Tl2Ba2Ca2Cu3O10; for the next membei
of the series, Tl2Ba2Ca3Cu4012, however, Tc falls back to around 115 K.

BljSrjCViCXO^, n=1, 2, 3 and HgBa2Ca2Cu3O,0

A series of compounds analogous to the thallium-barium-calcium-coppei

oxide superconductors, but containing Bi202 rather than T1202 layers, has been
synthesised. As the Bi-0 distance is slightly larger than the Tl-O bond length,
the Bi202 layers fit poorly between the cuprate sheets causing some structural
distortions; the compounds adopt orthorhombic unit cells and critical
temperatures are somewhat lower than the thallium analogues.
The compound HgBa2Ca2Cu3O10 was reported in early 1993 with a Tc ol
134 K. Its structure is similar to that of TlBa2Ca2Cu309, but with a single
HgO layer separating three Cu02 planes. The compound is difficult to make
as a pure phase, but illustrates that further upward progress in Tc is still
possible; indeed under pressure the Tc can be raised to 160 K.

Superconducting cuprates — general points

Consideration of the several superconducting systems described above shows

two features in common. They all contain planes of the stoichiometry CuO-
formed by linking Cu04 square planes (elongated Cu06 octahedra in
Lat gBa^CuO,!) at all vertices, Fig. 8.9, and they all have an average copper
oxidation state in excess of 2+. These Cu02 sheets are separated by ‘charge
reservoir’ layers, e.g. CuO chains in YBajCujC^ and the T1202 rocksalt layers
in Tl2Ba2Ca„_1Cu,l04n+2, which seem to ensure the correct charge or electron
concentration on the superconducting sheets.

/ / / / / //////

Fig. 8.9 Schematic of the structural

elements present in a high
temperature superconducting
cuprate.
 Inorganic materials chemistry 87

Consideration of the numerous superconducting cuprates which have been

ynthesised in the last five years shows these features to be recurring. The
:earch for new superconducting materials is thus concentrating on compounds
vhich show these structural characteristics.

3.2 Fullerenes and Fullerides

rhe chemistry of C^, Fig. 8.10, has crossed many of the conventional borders
)f chemistry and physics with interest in this molecule covering many areas,
rhese fields of study include the organic and organometallic chemistry of C60
is a ligand. The structural chemistry of the solid fullerene C60 and the M„C60
'ulleride derivatives are covered here.

'■'60

rhe first crystals of were grown from benzene solutions and contained
solvent molecules. With the correct purification methods, using sublimation
:o eliminate the solvent molecules, pure Cm crystals may be grown. The
compound adopts a face centred cubic array of the C60 molecules shown in
Fig. 8.11, which is equivalent to the close packing of C60 molecules. At room Fig. 8.10 The structure of a
molecule.
emperature the thermal energy available to the molecules allows them to
*otate freely on their lattice sites and they can be considered to be spherical,
rhe powder X-ray diffraction pattern of polycrystalline C60, shown in Fig.
3.12, is typical of the face centred cubic lattice though due to the high Note that one of the most efficient
symmetry of the spherical molecules, reflections with Miller indices M)0 have ways of packing spheres, in terms
Lero intensity and are not seen. The lattice parameter of 14.17 A obtained of filling space, is cubic close
from the powder data yields a contact distance of 14.17/V2 or 10.02 A for the packing. Given the shape of the
Cgo molecule it is not surprising
2^ spheres, which is made up of the diameter of a molecule, 7.06 A, plus
that they adopt the face centre
i van der Waals distance of 2.96 A between the molecules. This van der cubic lattice.
Waals distance is slightly smaller than the value in graphite, corresponding to
:he inter-layer separation, of 3.35 A.
88 Some recent developments in inorganic materials chemistry

Fig. 8.12 The powder X-ray

diffraction pattern of solid C^.

The intercalation compounds of Cwl

Reaction of C60 with potassium vapour and other alkali metals forms
compounds of the type AXC60, the stoichiometry of the product depending on
the availability of the reactant. With excess alkali metal, compounds of the
stoichiometry A6C60, A = K, Rb, Cs are formed. The structure of K6C60 is
body centred cubic; C60 molecules occupy sites at the cell corners and body
centre whilst the potassium atoms fill a number of sites near the centres of the
faces. Of most interest of the other stoichiometries are the compounds A3C60
which have been found to be superconducting at reasonably high temperatures.
The powder X-ray diffraction pattern of KjC^ is similar to that of the parent
C60 in Fig. 8.12. The peaks are shifted to lower 20 values indicating an
expansion of the C60 face centred cubic lattice. The stoichiometry K3C60 is
obtained by filling all the tetrahedral and all the octahedral holes in the close
Fig. 8.13 The structure of KjCgo- packed C60 lattice and is shown in Fig. 8.13.
becomes superconducting at 19 K, though gradual replacement of
potassium by the larger alkali metal ions raises Tc, so that Rb3C60 has a Tc of
29 K and CsRb2C60 a value of 33 K.

8.3 Problems

8.1 The compound La2SrCu206 has the same structure as Sr3Fe206, shown in
Fig. 6.15, but with lanthanum replacing some of the strontium. Would you
expect this compound to be a high temperature superconductor?

8.2. The 111 reflection (^=1.54 A, 154 pm) of K3C60 is at 10.75°. Calculate
the shortest potassium-potassium distance in this compound using the
description of the structure given above.
Answers to problems

1.1. F, F, P, I.
1.2. Draw the unit cell in projection down b. The short diagonal of the
parallelogram becomes the face of the face centred lattice.
1.3. Orthorhombic; dm=dl0_{ = 2.40 A (240 pm). Monoclinic; dm= 1.956 A
(195.6 pm), dx„_!= 2.752 A (275.2 pm).
2.1. a = 12.358 A (1235.8 pm), lattice type P. Maximum aperture, wall to
wall is about 9 A (90 pm).
2.2. F lattice 33.43 (111), 38.70 (200), 56.03 (220), 66.84 (311) and 70.23
(222); I lattice 34.83 (110), 50.09 (200), 62.46 (211) and 73.55 (220). The
peaks would move to lower 20 values as the metal expands.
2.3. Peak splits giving 3 reflections. 310,301,103. with similar intensities.
3.1. Use precursors/precipitation methods.
3.2. The metal-oxygen framework is determined in the high temperature
reaction. KNbTi05 has a different structure from LiNbTiOs; the ion exchange
of Li for K leaves the [NbTi05]“ framework essentially unchanged.
3.3. The liquid cools until, just below T2, the liquidus is reached; solid AB
precipitates and the liquid becomes richer in B. This process continues with
further cooling until the eutectic is reached and the remaining liquid solidifies
to give further AB + B.
4.1. f=1.037. Similar behaviour to BaTi03.
4.2. FeCr204 normal spinel Cr3+ has very strong preference for octahedral
sites. NiGa204 inverse spinel Ni2+ has a preference for octahedral sites.
5.1. Li03ReO3 is metallic; the 7t-level has 0.3 extra electrons per rhenium, as
compared with Re03, but is still only partially filled.
5.2. The orbital points along the M-M direction. In TiOz a band formed
from the overlap of these orbitals is empty and Ti02 is an insulator. In V02
it is partially filled and V02 demonstrates metallic properties.
6.1. =0.6 A, hence need for vacancies on normal sites.
6.2. The a parameter decreases smoothly ( a non-stoichiometric system ) from
Sr3Fe206 to Sr3Fe207 as the Fe3+ is oxidised to the smaller Fe4+. The c
parameter initially gets larger as the FeOs square pyramids are pushed apart
to make room for the oxide ion then it also decreases smoothly.
7.1. a=20.15 A, b=20.08 A and c =13.51 A. The wall to wall distance is
=8.5 A taking no account of the radius of the oxygen atom.
7.2. The interlayer spacing is 56.2 A in the amine derivative and 5.03 A in
TaS2. Hence the amine chain length must be =25 A. If the amine chains were
tilted the interlayer spacing would be reduced and the 001 reflection would
move to larger 20 values.
8.1. La2SrCu206 is not a superconductor. The structure contains Cu02 planes
but the copper oxidation state is 2.0 exactly.
3.2. 6.17 A (617 pm)
Bibliography

1. Inorganic solids. Adams, D. M. (1974). Wiley, New York.

2. Modem powder diffraction: reviews in mineralogy volume 20. Bish, D.

L. and Post, J. E. (ed.) (1989). The Mineralogical Society of America,
Washington DC.

3. Inorganic materials. Bruce, D. W. and O’Hare, D. (ed.) (1992). Wiley,

New York.

4. Solid state chemistry: techniques. Cheetham, A. K. and Day, P. (ed.)

(1987). Oxford University Press.

5. Solid state chemistry: compounds. Cheetham, A. K. and Day, P. (ed.)

(1992). Oxford University Press.

6. Electronic structure and chemistry of solids. Cox, P. A. (1987). Oxford

University Press.

7. Transition metal oxides. Cox, P. A. (1993). Oxford University Press.

8. An introduction to zeolite molecular sieves. Dyer, A. (1988). Wiley, New

York.

9. Ionic crystals, defects and non-stoichiometry. Greenwood, N. N. (1968).

Butterworths, London.

10. Inorganic structural chemistry. 2nd edition. Muller, U. (1992). Wiley,

New York.

11. Solid state chemistry: an introduction. Smart, L. and Moore, E. (1992).

Chapman and Hall, London.

12. Structural inorganic chemistry. 5th edition. Wells, A. F. (1984). Oxford

University Press.

13. Solid state chemistry and its applications. West, A. R. (1984). Wiley,
New York.
ndex

'here chemical formula are indexed they are provided at the end of each letter section.

fences 20-21 (CH), 50, 51 high pressure 30

'sorption in zeolites 74 C8K 35, 55, 69 hydrothermal synthesis 30
alumina 80 Cu02 layers 85-86 HgBa2Ca2Cu309 86
jminosilicates 71
norphous materials 17 d-orbitals 47 indexing powder diffraction data 17-19
igstroms 9 d-spacing 6-8 insertion compounds 34, 43
itiferroelectric 43 Debye-Scherrer camera 12 insulators 51
itiferromagnetism 55-56 defect cluster 64-65 intensities (in X-ray diffraction) 24
iplications of powder X-ray density of states 51 intercalation 34, 69, 78
diffraction 16 diffraction 10 intrinsic defects 60
diffractometer 13 inverse spinel 45
inds 50 1,4-dimethylbenzenes in ZSM-5 76 ion exchange 36, 75
ind theory 49 domain 42 ionic conductivity 79
;ntonite 77 ionic radii 41
nary phase diagrams 31 eutectic 31 ion size (in spinels) 46
>dy centred lattice 4, 21, 22 excess octahedral crystal field
ragg’s equation 10, 17 stabilisation energy 47 JCPDS powder diffraction file 15
ragg’s law 10
ravais lattices 5 face centred lattice 4 Koch-Cohen cluster 65
ronsted acid 75 faujasite 73 KC1 (powder X-ray pattern ) 22
i2Sr2Can.1Cun04+2n 86 Fermi level 51 KH2P04 58
iTiO, 20, 41 ferrimagnetism 55-56, 57 K2NiF4 structure 48
ferroelectric 42 KTiOP04 58
pacitors 42 ferromagnetism 55-56 KAo 88
italysis 75 fluorite structure 63
11 parameters 2 fractional coordinates 2 lattice 4 <
ays 77 Frenkel defects 61 lattice energy 27, 46, 60
induction band 52 fullerenes 87 lattice parameter 17
ucible 26 FelxO 62, 64-65 lattice planes 6
ystallite size 24 Fe304 47, 57 lattice types 4, 20
ystal class 2, 3 layer structures 34
ystal field stabilisation energies 47 gel 28 Lewis acid 75
ystallographic shear 67 graphite 34, 55, 78, 69 liquidus 31
ystal system 2, 3 graphitic intercalation compounds 35, La2.xBaxCu04 48, 83
ibic materials (indexing) 18 54, 69, 78 LiNbOj 17, 41, 43, 59
Guinier camera 13 LiTiNbOj 36
iprates 86
jrie temperature 42
hexagonal systems 19 magnesium metal 51
iTi03 38
high temperature superconductors magnetic susceptibility 56
so 87-88
82-86 magnetism 55
,Br 35, 55
92 Index

metal 51 pillared clays 77 template 31

metal oxides 35, 37-48 point defects 60 ternary phase diagrams 34
Miller indices 6,18 pore dimensions in zeolites 74 tetragonal systems 3, 19, 42, 85
molecular orbitals 49 powder diffraction 11 tetrahedral holes 44
monoclinic 2, 3 precursors (in synthesis) 28 tolerance factors 40
montmorillonite 77, 78 primitive lattice 4, 40, 22 translational symmetry 4
multiplicities 19 projections 2 trigonal 3
MO 37,52 (see also TiO, MnO, Fe, O tube furnaces 29
and NiO) quenching 64, 83 TaS2 78-79
M02 37 TiO 37, 53, 66
M03 38 (see also Re03 and W03) reactions between solids 26 TljBajCa^.pUnO^ 30, 85-86
MgAl204 16, 28, 44-45 reflection conditions 22
MgCr204 Rietveld method 24 unit cell 1
Mg2Si04 33 ring sizes in zeolites 72 U02+x 62-64
MnO 56 rocksalt layers 85-86
rutile structure 37 variable temperature X-ray diffraction)]
NASICON 80 ReO, 36, 38, 39, 43, 54, 68 25
Neel temperature 56 VPI-5 73
non-linear optics 58-59 Scherrer formula 24-25
non-stoichiometry 60-72 sealed tubes 29 W03 38, 43, 54
normal spinel 45 selectivity 76
NiO 37, 53, 57 semiconductor 52 X-ray diffraction 8-14
NiAl204 Schottky defects 60, 80 X-ray generation 9
shape selectivity 76 X-ray spectrum 9
octahedral holes 44 sodalite cage 72 X-ray tube 9
optical properties 58 sodium metal 50 xylenes in ZSM-5 76
orthorhombic 3 spinel 44-47, 57
oxalates 29 split interstitial 61 YBa2Cu307 y 3, 29, 83-84
staging 69
paramagnetism 56 superconductivity 82 Zeolites 71-76
pelletising 28 superexchange 56 Zeolite-A 30, 36, 73
perovskite structure 38-41, 84 synthesis 26-32 ZSM-5 30,72,74-76
phase diagrams 31 systematic absences 20 Zn1+xO 66
phase purity 16 SrTiO, 38, 40 ZnS 3
piezoelectricity 43 Sr3Fe207.y 67 Zr02 80
 DATE DUE
DATE DE RETOUR

- J i

MAR 2 3 2005

CARR MCLEAN 38-296 ~

 \ ‘
H M I R Y
I M

STEPHEN G. DAVIES This series of short texts provides accessible accounts of a range of essential topics in
chemistry and chemical engineering. Written with the needs of the student in mind,
ilC
the Oxford Chemistry Primers offer just the right level of detail for undergraduate
JOHN EVANS study, and will be invaluable as a source of material commonly presented in lecture
courses yet not adequately covered in existing texts. All the basic principles and facts
LYNN F. GLADDEN
in a particular area are presented in a clear and straightforward style, to produce
concise yet comprehensive accounts of topics covered in both core and
specialist courses.

The chemistry of solid inorganic materials has become a central theme in research
and in the teaching of chemistry. This book is a clearly illustrated and comprehensive
introduction to solid state materials which also addresses the major experimental
technique used to study them, powder X-ray diffraction. Experimental data from
this technique is used to illustrate topics throughout the book. This tying together
of an experimental method and the chemistry reflects the true needs of the chemist
and the method by which chemistry is taught.

Mark T. Weller is a Reader in the Department of Chemistry, University of

Southampton.

ISBN 0-19-855798-

OXFORD UNIVERSITY PRESS 9 780198 5579 82