The f elements

Nikolas Kaltsoyannis and

Peter Scott

OxXcroORD SCIENCE PUBLICATIONS

 OXFORD GC Pim
Sai Re “eRe lMic nes
Physical Chemistry Editor Founding Editor and Organic Inorganic Chemistry Editor Chemical Engineering Editor
RICHARD G. COMPTON Chemistry Editor JOHN EVANS LYNN F. GLADDEN
Physical and Theoretical STEPHEN G. DAVIES Department of Chemistry Department of Chemical Engineering
Chemistry Laboratory The Dyson Perrins Laboratory University of Southampton University of Cambridge
University of Oxford University of Oxford

S. E. Thomas Organic synthesis: The roles ofboron P. D. Bailey and K. M. Morgan Organonitrogen
and silicon chemistry
D. T. Davies Aromatic heterocyclic chemistry C. E. Wayne and R. P. Wayne Photochemistry
P.R. Jenkins Organometallic reagents in synthesis C. P. Lawrence, A. Rodger, and R. G. Compton
M. Sainsbury Aromatic chemistry Foundations of physical chemistry
L. M. Harwood Polar rearrangements R. G. Compton and G. H. W. Sanders
J. H. Jones Amino acid and peptide synthesis Electrode potentials
C. J. Moody and G. H. Whitham Reactive P. B. Whalley Two-phase flow and heat transfer
intermediates L. M. Harwood and T. D. W. Claridge
G. M. Hornby and J. M. Peach Foundations of Introduction to organic spectroscopy
organic chemistry C. E. Housecroft Metal-metal bonded carbonyl!
R. Henderson The mechanisms of reactions at tran- dimers and clusters
sition metal sites H. Maskill Mechanisms of organic reactions
H. M. Cartwright Applications of artificial P.C. Wilkins and R. G. Wilkins /norganic
intelligence in chemistry chemistry in biology
M. Bochmann Organometallics 1; Complexes with J. H. Jones Core carbonyl chemistry
transition metal—carbon o-bonds N. J. B. Green Quantum mechanics 1; Foundations
M. Bochmann Organometallics 2: Complexes with I. S. Metcalfe Chemical reaction engineering:
transition metal—carbon m-bonds A first course
C. E. Housecroft Cluster molecules of the p-block R.H.S. Winterton Heat transfer
elements N.C. Norman Periodicity and the s- and p-block
M. J. Winter Chemical bonding elements
R.S. Ward Bifunctional compounds R. W. Cattrall Chemical sensors
S. K. Scott Oscillations, waves, and chaos in M. Bowker The basis and applications of
chemical kinetics heterogeneous catalysis
T. P. Softley Atomic spectra M. C. Grossel Alicyclic chemistry
J. Mann Chemical aspects of biosynthesis J.M. Brown Molecular spectroscopy
B. G. Cox Modern liquid phase kinetics G. J. Price Thermodynamics of chemical processes
A. Harrison Fractals in chemistry A. G. Howard Aquatic environmental chemistry
M. T. Weller /norganic materials chemistry A.O.S. Maczek Statistical thermodynamics
R. P. Wayne Chemical instrumentation G. A. Attard and C. J. Barnes Surfaces
D. E. Fenton Biocoordination chemistry W. Clegg Crystal structure determination
W.G. Richards and P. R. Scott Energy levels in M. Brouard Reaction dynamics
atoms and molecules _K. Brisdon /norganic spectroscopic methods
M. J. Winter d-Block chemistry . Proctor Stereoselectivity in organic synthesis
D. M. P. Mingos Essentials of inorganic chemistry 1 .M.A. Brett and A. M. O. Brett Electroanalysis
G. H. Grant and W. G. Richards Computational Ze
Gy _ J.B. Green Quantum mechanics 2: The tool kit
chemistry D. M. P. Mingos Essentials of inorganic chemistry 2
S.A. Lee and G. E. Robinson Process development: I. Fleming Pericyclic reactions
Fine chemicals from grams to kilograms
N.S. Lawrence, J. D. Wadhawan & R. G. Compton
C. L. Willis and M. R. Wills Organic synthesis Foundations of physical chemistry. Worked examples
P. J. Hore Nuclear magnetic resonance J. R. Chipperfield Non-aqueous solvents
G. H. Whitham Organosulfur chemistry T. J. Mason Sonochemistry
A.C. Fisher Electrode dynamics
J. McCleverty Chemistry of the first-row transition
G. D. Meakins Functional groups: Characteristics metals
and interconversions
E. C. Constable Coordination chemistry of
A.J. Kirby Stereoelectronic effects macrocyclic compounds
P. A. Cox Introduction to quantum theory and C. E. Housecroft The heavier d-hlock metals:
atomic structure Aspects of inorganic and coordination chemistry
—_—_e——_—__C_C
 HA 0196 504675 9001

la Loe =a
STILLORGAN LIBRARY
Inv/06 : K109 Price E16.05
Title: os elements
e | Class: Ub: ae
witr i as ee

Nikolas Kaltsoyannis
Lecturer in Chemistry, University College London

Peter Scott
Senior Lecturer in Chemistry, University of Warwick

BAINTE DEN STOC

WITHDRAWN FROM
DUN LAOGHAIRE-RATHDOWN COUNTY
LIBRARY STOCK

OXFORD
UNIVERSITY PRESS
 OXFORD
UNIVERSITY PRESS

Great Clarendon Street, Oxford OX2 6DP

Oxford University Press is a department ofthe 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 Dares 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
© Nikolas Kaltsoyannis and Peter Scott, 1999
The moral rights of the author have been asserted
Database right Oxford University Press (maker)
First published 1999
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 850467 5 (Pbk)
Bo SES VO SG
Typeset by the author

Printed in Great Britain

On acid free paper by
Antony Rowe Ltd
Chippenham, Wiltshire
Series Editor's Foreword
The f elements have tended to be relatively neglected in inorganic chemistry courses in
recent years, in spite of their intrinsic interest and technological importance. A contributing
cause has been the lack of an up-to-date text on this field, which has been studied less than the
d block elements by inorganic chemists in recent years. This Primer is aimed at that gap.
Oxford Chemistry Primers are designed to give a concise introduction to all chemistry
students by providing the material that would usually form an 8-10 lecture course. As well as
providing up-to-date information, this series expresses the explanations and rationales that
form the framework of current understanding of inorganic chemistry.
Nik Kaltsoyannis and Peter Scott have provided a timely synopsis of f element
chemistry, covering standard complex chemistry, their electronic and magnetic properties
and also organometallic chemistry. The complementary skills of these two authors have
given an authority to this wide scope. Companion books covering the Heavy Transition
Elements (OCP73) and the First row transition metals (OCP71) have preceeded this volume
and this suite gives excellent coverage of the whole oftransition element chemistry.
John Evans
Department of Chemistry,
University ofSouthampton

Preface
There are two rows of elements propping up the periodic table about which most graduate
chemists know very little. Some may have come across the lanthanide contraction and may
be aware of its consequences for the chemistry of the d block. Others may have met the sand-
wich molecule uranocene in courses on organometallic chemistry. It is also possible that the
unique magnetic and spectroscopic properties of these elements were encountered at the end
of a physical chemistry lecture series. This piecemeal treatment is unfortunate as the lantha-
nides are finding increasing technological and catalytic applications, and the concern over
the growing quantities of actinide waste littering the globe is well founded. Our graduates
ought to know more about the f elements.
In this text we have drawn together key concepts in f element chemistry to provide a
student-friendly introduction to this fascinating group of elements. The traditional approach
adopted in the few existing texts on the f elements is to treat each element in turn. The very
nature of the lanthanides and actinides as families of elements with similar properties means
that this involves considerable repetition. We have chosen instead to concentrate on the
trends within the two series and the comparisons between them and the other elements of the
periodic table. We hope that we have provided a collection of concepts to be applied rather
than facts to be memorised. The material is targetted at third or fourth year undergraduates,
and will put to use many ofthe skills developed in earlier years and indeed in other Primers.
We would like to thank Andrea Sella for casting a critical eye over the entire manuscript,
and Geoff Cloke, Bob Denning, Peter Moore, David Parker, and John Watkin for helpful
comments on various sections. We also want to thank J and N for their uniquely special
contributions.
NK PS
London Warwick
January 1999
Contents
Definitions
and origins

Properties of the atoms and ions

Relativity, electronic spectroscopy, and magnetism

Solid state compounds

Coordination chemistry

nn
Aa
NHN
B&B
&= Organometallics
Further reading

Index
1 Definitions and origins

1.1 Definition of the f elements

The modern periodic table is most commonly depicted as in Fig. 1.1, with
the two rows Ce—Lu and Th-Lr set apart from the other elements. These two
rows are collectively known as thef block orf elements, and are divided into
the lanthanides (Ce—Lu) and actinides (Th-Lr). The lanthanide elements lie
between lanthanum (atomic number 57) and hafnium (atomic number 72),
while the actinides are situated between actinium and rutherfordium (atomic
numbers 89 and 104 respectively). All of the f elements are metallic.

orm | Gi | Gd] te | Dy | Ho | er [im | vo | tu

Fig. 1.1 The modem periodic table. The f elements are situated in the two shaded rows, and are divided into the lanthanides (light
shading) and the actinides (dark shading). The ‘parent f elements, lanthanum and actinium, are also shaded appropriately.

Although the ‘parent’ elements of the f block, lanthanum and actinium, are
really members of group three, they are often included in discussions of f
element chemistry. We will not, therefore, confine ourselves to the shaded
rows in Fig. 1.1, but will include certain aspects of the chemistry of
lanthanum and actinium where appropriate. Comparisons will also be made
between the lanthanides and scandium and yttrium.
Throughout this book we will use the the general symbols Ln and An to
refer to the lanthanide and actinide elements respectively.

1.2 Historical perspectives

Discovery of the lanthanides and actinides

In 1794 J. Gadolin, a Finnish chemist, isolated yttria from a mineral that had
recently been discovered at Ytterby, a village near Stockholm in Sweden.
The village of Ytterby has the honour
Gadolin believed that yttria was the oxide of a single new element, but of being the origin of the name of four
subsequent work revealed it to contain the oxides of no fewer than ten new elements, yttrium, terbium, erbium,
elements, yttrium, terbium, erbium, ytterbium, scandium, holmium, and ytterbium.
2 Definitions and origins

thulium, gadolinium, dysprosium, and lutetium! Shortly after Gadolin’s

discovery, M.H. Klaproth and, independently, J.J. Berzelius and W. Hisinger,
isolated another new oxide, ceria. This was later shown to contain the oxides
of cerium, lanthanum, praseodymium, neodymium, samarium, and europium.
Thorium, protactinium, and uranium are the only naturally occurring
actinide elements. In 1789 Klaproth showed that pitchblende, thought
previously to be a mixture of zinc, iron, and tungsten oxides, also contained
the oxide of a new element which he called uranium. 39 years later Berzelius
discovered thoria, a new oxide from which he subsequently isolated thorium.
Protactinium was not discovered until 1913, when K. Fajans and O. Gohring
identified **Pa as a member of the ***U radioactive decay series. This isotope
of protactinium is short-lived (t;, = 6.7 hours), but the more stable **'Pa,
identified in 1916 by O. Hahn, L. Meitner, P. Soddy, and J.A. Cranston, has
a half-life of 32 760 years.
None of the other actinide elements occurs naturally, and must be
synthesised by nuclear reactions. Table 1.1 lists the date and method of the
first synthesis of the transuranium elements.

Element Date and method of first synthesis

Neptunium (Np) 93 1940. Bombardment of 738U with }n
Plutonium (Pu) 94 1940. Bombardment of *35U with 7H
Americium (Am) 95 1944, Bombardment of *33Pu with }n
Curium (Cm) 96 1944. Bombardment of *33Pu with 4He
Berkelium (Bk) 97 1949. Bombardment of *4Am with }He
Californium (Cf) 98 1950. Bombardment of *§7Cm with 3He
Einsteinium (Es) 99 1952. Found in debris of first thermonuclear
explosion.
Fermium (Fm) 100 1952. Found in debris of first thermonuclear
explosion.
Mendelevium (Md) 101 1955. Bombardment of “ahs with }He
Nobelium (No) 102 1965. Bombardment of *33Am with !>N
Lawrencium (Lr) 103 1961-71. Bombardment of mixed isotopes of
act with "{B, *{Bandof 2Am with #0
etc.

Table 1.1 Discovery (synthesis) of the transuranium elements.

Position of the f elements in the periodic table

Although it may appear strange that yttria was initially regarded as being the
oxide of just a single element, instead of the ten eventually discovered, it
must be recognised that there was no guide available as to how many new
elements remained to be found. Indeed, it was not until the work of H.G.J.
Moseley in the early part of the 20" century that it was recognised that there
were 14 elements between lanthanum and hafnium. By 1907 all of these had
 The f elements 3

been identified bar the radioactive promethium, conclusive evidence for which
had to wait until 1947, when J.A. Marinsky, L.E. Glendenin, and D.C.
Coryell observed element 61 in the radioactive decay products of ?°U.
Shortly after Moseley’s work, N. Bohr realised that the 14 elements
between lanthanum and hafnium reflected the fact that the fourth atomic
primary quantum shell could accomodate 32 electrons, 14 more than the third.
These additional electrons are placed in the 4f orbitals, and the lanthanide
elements were recognised as forming a new family in the periodic table. For
obvious reasons the lanthanides are often referred to as the 4f series of
elements. That the actinides form a second f series - the Sf - was not realised
until the work of G.T. Seaborg during the Second World War. His suggestion
took thorium, protactinium, and uranium out of groups four, five, and six
respectively and into their rightful place at the start of a new family of
elements, many of which Seaborg was instrumental in synthesising.

1.3 Occurrence, extraction, and synthesis of the f

elements

Lanthanides
The term ‘rare earth’, which is often used to describe the lanthanides, is rather
misleading as many of the 4f elements are quite abundant. Cerium is the 26"
most abundant element on the Earth, neodymium is more abundant than gold
and even thulium (the least common lanthanide except for the radioactive
promethium) is more abundant in the Earth’s crust than iodine. Only two
lanthanide-containing minerals are important commercially; monazite, a
mixed lanthanide orthophosphate (LnPO,) and bastnaesite, a fluorocarbonate
(LnCO,F). The most common metals in both ores are (in order of decreasing
abundance) cerium, lanthanum, neodymium, and praseodymium, with
monazite also containing up to 10% ThO, as well as smaller quantities of the
later lanthanides. Monazite deposits occur in many countries, including India,
Brazil, Sri Lanka, South Africa, Australia, and Malaysia, while the principal
sources of bastnaesite are China and the Sierra Nevada mountains in the
western USA.
Monazite is typically processed by dissolution in ca. 70% NaOH solution
for several hours followed by addition of hot water to generate a slurry of
crude hydrous oxides. The slurry is then added to boiling HCl until a pH of
3.5 is reached, at which point crude hydrous ThO, is precipitated.
Stoichiometric amounts of BaCl, and Ln,SO, (three and one equivalents
respectively) are then added to the remaining solution of impure LnCl, to
precipitate BaSO, and RaSO,, leaving a solution of mixed lanthanide
trichlorides. This is also the end product of bastnaesite processing, although a
somewhat different route is employed. Large-scale separation of the individual
lanthanide chlorides is achieved using solvent extraction, typically with a
complexing agent such as tributylphosphate - ("BuO),PO - in an inert diluent
such as kerosene. Separation relies on the increased solubility of Ln(II) with
increasing atomic mass. Alternatively, high purity, small-scale separation
4 Definitions and origins

may be achieved using ion exchange chromatography, which is discussed in

Chapter Five, Section Four.

Occurrence and extraction of thorium, protactinium, and

uranium
The Earth’s crust contains 8.1 ppm of thorium, the principal source of which
is monazite. Following the precipitation of ThO, with boiling HCl (as
described above), nitric acid is used to dissolve the hydrous ThO,, and the
resulting thorium nitrate is purified by extraction into tributylphosphate
diluted with kerosene.
Protactinium is one of the rarest naturally occuring elements, and is found
in trace amounts in uranium ores. The majority of the World’s supply of
protactinium was produced by workers at the UK Atomic Energy Authority in
1960. They painstakingly extracted a total of 126.75 g of protactinium from
about 60 tons of sludge that had built up during the extraction of uranium
from UO, ores. Small amounts of this sample have been distributed to
laboratories all over the world, from which much of the chemistry of
protactinium has been elucidated.
Uranium is the heaviest naturally occuring element, with an abundance of
2.4 ppm in the Earth’s crust. It is found in many oxide minerals, the most
important of which are pitchbende or uraninite (U;O,) and carnotite
[K,(UO,),(VO,).3H,O]. The main sources of these minerals are Canada, the
USA, South Africa, Australia, France, and the countries of the former Soviet
Union. The details of the extraction of uranium depend upon the particular ore
being used, but typically the ore is leached with H,SO, in the presence of an
oxidising agent such as MnO, or NaClO, to ensure that all of the uranium is
present as an acidic solution of a uranyl (UO,’*) complex of sulphate or
chloride. Neutralisation is usually performed with NH, to yield a precipitate
of ‘yellow cake’, a mixture of salts and oxides of approximate composition
(NH,),U,0,. UO; is produced by heating the ‘yeliow cake’ at 300 °C. This
may in turn be reduced to UO, by heating at 700 °C in the presence of H,, and
the UO, may be converted to UF, by reaction with HF. Uranium metal is
obtained by reduction of UF, with magnesium.

Synthesis of the transuranium elements

As indicated in Section 1.2, none of the transuranium elements occurs
naturally. Neptunium—fermium have been synthesised by neutron
bombardment, a process which relies on the fact that neutron capture by a
heavy atom is often quickly followed by the emission of a B particle to
produce a new atom with its atomic number increased by one. The principal
problem with this route is that the yields of the heavier transuranium
elements are very low, as their production relies on successive neutron
capture. For example, continuous irradiation of kilogram quantities of
plutonium by neutrons in a nuclear reactor will produce only milligram yields
of californium over several years.
A second problem is that neutron capture cannot be used to synthesise
elements beyond *°’Fm, as the product of the next neutron capture (**Fm)
undergoes spontaneous fission. To obtain the elements beyond *°’Fm another
 The f elements 5

approach must be used, involving the bombardment of an actinide target by

very light nuclei. This method, some examples of which are given in Table
1.1, produces extremely low yields (the new elements are produced one atom
at a time!) and is limited by the availability of suitable actinide targets.
The creation, manipulation, and characterisation of the later actinides is
extremely difficult. Not only are the yields extraordinarily small, but the
reactions are generally not clean and the desired products must be separated
from other actinides as well as lighter fission products (which include
lanthanides). Furthermore, the intense radioactivity of most of the actinides
necessitates the use of radiation shielding, and most manipulations must be
cared out by remote control. However, the radioactivity has the advantage
that it allows the detection of tiny quantities of material which would be
unmeasurable by conventional chemical detection methods.
 2 Properties of the atoms and
ions

2.1 Electronic structures of the f elements

The electronic configurations of free lanthanide and actinide atoms are often
difficult to determine owing to the complexity of their electronic spectra (and,
in the case of the later An, the difficulties in producing sufficient numbers of
atoms). Table 2.1 presents the generally accepted ground electronic
configurations of all of the f elements, from which it may be seen that the
principal change as the two series are crossed is the gradual filling of the 4f
(Ln) and 5f orbitals (An).

Lanthanides Actinides
Element Symbol Atomic Electronic Element Symbol Atomic — Electronic
number configuration number configuration
Cerium Ce 58 [Xe]4f!5d!6s? Thorium Th 90 [Rn]6d?7s?
Praseodymium Pr 59 [Xe]4f%6s" Protactinium Pa 91 [Rn]5f°6d'7s"
Neodymium Nd 60 [Xe]4f*6s" Uranium U 92 [Rn]5f°6d'7s?
Promethium Pm 61 [Xe]4f6s" Neptunium Np 93 [Rn]5f*6d'7s"
Samarium Sm 62 [Xe]4f%6s" Plutonium Pu 94 [Rn]5f°7s"
Europium Eu 63 [Xe]4f’6s" Americium Am 95 [Rn]5f’7s*
Gadolinium Gd 64 [Xe]4f’5d'6s" Curium Cm 96 [Rn]5f’6d'7s"
Terbium Tb 65 [Xe]4f6s" Berkelium Bk 97 AIST
Dysprosium Dy 66 [Xe]4f'°6s’ Californium Cf 98 {Rn]5f'°7
Holmium Ho 67 [Xe]4f!'6s? Einsteinium Es 99 [Rn]5sf'7¢
Erbium Er 68 [Xe]4f!*6s" Fermium Fm 100 [Rn]5f'?7s?
Thulium Tm 69 [Xe]4f'6s? Mendelevium Md 101 (Rn]5f'*7s"
Ytterbium Yb 70 [Xe]4f!*6s? Nobelium No 102 (Rn]5f'*7s?
Lutetium Lu as Xe]4fl5d'6s*__ | Lawrencium Ly 103 [Rn]5f'*6d'7s"

Table 2.1 Ground electronic configurations of the lanthanides and actinides.

f orbitals
The 4f and 5forbitals play a central role in determining the physicochemical
properties of the lanthanides and actinides respectively. Unfortunately there is
no unique way of representing f orbitals. Figure 2.1 presents one of the most
common sets of f orbitals - the cubic set - which is appropriate for molecules
in which the x, y, and z axes are symmetry related, e.g. those belonging to
the O, or 7, point groups.
 Thef elements 7

£,3 (= fy222-3:2-a2))- £,3 (= faex2-ay2-a22)) ANA £3 (= fyiay2-a22.a,2)) are similar and are oriented
along the x and y axes respectively.

Fig. 2.1 The cubic set of f orbitals.

f.2-y2)> fxy2-22) aNd fy-2.,2) are similar.

xyz
8 Properties of the atoms and ions

When dealing with non-cubic geometries we use another set of f orbitals -

the general set - which consists of combinations of the orbitals shown in Fig.
2.1. One approach to understanding why there is no unique way of
representing f orbitals is as follows. In O, symmetry, the f orbitals split into
three sets, which span the a,,, t,,, and t,, irreducible representations. If the
symmetry is now lowered to Dy, by, for example, distortion along the z axis,
the f orbitals transform as b,,, a),+e,, and b,,+e,, i.e. the formerly triply
degenerate sets of orbitals have each split into a non degenerate orbital and a
doubly degenerate e, set. As the two e, orbitals have the same symmetry they
can mix with one another, and it turns out to be most useful to allow them to
do so and to work with the new combinations.

2.2 Trends in ionization energies and electrode potentials

and their relationship to oxidation states

Lanthanides
The chemistry of the lanthanides is dominated by the +3 oxidation state. One
of the principal reasons for this is illustrated by the ionization energy data in
Table 2.2, which reveal that in all cases the fourth ionization energy, /, [the
energy associated with the process Ln** (g) — Ln* (g) + e (g)] is greater than
the sum of the first three ionization energies. The extra energy required to
remove the fourth electron is so great that in most cases it cannot be
recovered through chemical bond formation, and thus the +4 oxidation state is
largely inaccessible.

Element 1, iD ic Label, iE
Ce 527 1047 1949 3523 3547
Pr 523 1018 2086 3627 3761
Nd 530 1035 2130 3695 3899
Pm 524 1052 2150 3726 3970
Sm 543 1068 2260 3871 3990
Eu 547 1085 2404 4036 4110
Gd 593 1167 1990 3750 4250
Tb 565 1112 2114 3791 3839
Dy 572 1126 2200 3898 4501
Ho 581 1139 2204 3924 4150
Er 589 1151 2194 3934 4115
Tm 597 1163 2285 4045 4119
Yb 603 1176 2415 4194 4220
Lu 524 1340 2022 3886 4360
Table 2.2 lonization energies of the lanthanides (kJ mot’).
 The f elements 9

As shown in Table 2.1, the valence electrons of the neutral lanthanide

elements are distributed in the 4f, 5d, and 6s orbitals. As electrons are
removed from the neutral atoms all of the orbitals are stabilised, but the 4f,
5d, and 6s levels do not experience the same degree of stabilisation, with the
4f being stabilised most and the 6s least. Once three electrons have been The order of orbital stabilisation 4f >
removed the additional stabilisation of the 4f orbitals over the 5d and 6s is so 5d > 6s with successive removal of
electrons from the neutral lanthanides
large that no electrons remain in the latter two orbitals. Furthermore, the
reflects the relative ability of these
remaining 4f electrons are so tightly held as to be chemically inaccessible. orbitals to penetrate the inner core of
Although the +3 oxidation state is by far the most common, five electrons to the region of space around
lanthanides also have a tetravalent chemistry. For neodymium and dysprosium the nucleus. This in tum is related to
this is confined to solid state fluoride complexes, while praseodymium and the number of radial nodes in the 4f,
5d, and 6s wavefunctions; zero, two,
terbium also form the tetrafluoride and dioxide. The most extensive Ln*
and five respectively.
chemistry is that of cerium, for which a variety of tetravalent compounds and
salts are known (see, for example, Chapter Four, Sections One and Two).
That Ce* is chemically accessible is due to the high energy of the 4f orbitals
at the start of the lanthanide series, such that they are not sufficiently stable
in Ce* to prevent the loss of another electron.

/kJ
|,
mol"
8

Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er

Fig. 2.2 The variation of the third ionization energy (/,) of the lanthanides.

The variation in the third ionization energy, /;, of the lanthanides is shown
in Fig. 2.2. The most striking feature of this plot is the very high values for
europium and ytterbium and the very low values for the elements immediately
following them, gadolinium and lutetium. This may be explained by
consideration of the electronic configuration of the Ln’* ions that are being
ionized to form the corresponding Ln**. Yb** has the [Xe]4f'* configuration.
Lu?*, however, has an additional electron in the 5d orbitals, which are less
stable than the 4f and therefore easier to ionize. The situation is less clear for
Gd**, because while there is little doubt that the electronic configuration of
Eu** is [Xe]4f’, it is not certain if that of Gd** is [Xe]4f’5d' or [Xe]4f*. If it is
the former, the low ionization energy may be rationalised in the same way as
10 Properties of the atoms and ions

for Lu’*. The [Xe]4f* configuration, however, may also be expected to ionize
more easily than the [Xe]4f’. In the latter, all seven f electrons occupy
different f orbitals with the same spin. In [Xe]4f*, however, one f orbital must
contain two paired electrons, and the increased repulsion between this pair
will destabilize the ion and make it easier to ionize. Furthermore, there is no
loss of exchange energy on ionization of the [Xe]4f* configuration, as the
electron that is removed has opposite spin to the other seven f electrons. By
contrast, there is a significant exchange energy loss on ionization of the
[Xe]4f’ configuration, and hence more energy is required to remove an electron
from it.
Given the data in Table 2.2 and Fig. 2.2 it should come as no surprise that
the lanthanides with the most extensive divalent chemistry are europium,
ytterbium, and to a lesser extent, samarium and thulium (i.e. those with large
I,). We shall come across many examples of divalent lanthanide compounds
throughout this book.

Actinides
The actinide elements display a much greater range of oxidation states than
the lanthanides, particularly in the early part of the series. Figure 2.3 shows
the oxidation states adopted by the actinides, from which it may be seen that
some of the lighter actinides resemble the transition metals in their range of
possible oxidation states, while the later actinides are more like the
lanthanides in favouring trivalency. This change of character as the series is
crossed is a general theme in actinide chemistry, one which we will return to
many times in this book.

+7 Oo. ©

+6 e oo oOo oO

2
S45 e o eo oO
6
S 44 0 OF —-O* Oh * « Os — -6— O
6
E +3 Oo oF OF 08F 0O06oe e060 0 06e 6 6 oO @®
LL

+2 oO oO o> «0- .O.

Th Pa U Np Pu Am Cm Bk Cf Es Fm Md No Lr

Fig. 2.3 The oxidation states adopted by the actinide elements in their compounds. The most stable oxidation state in aqueous solution
is represented by the black circles. Open circles indicate other oxidation states adopted and squares indicate that the oxidation state is
found only in solids.
 The f elements 11

Unfortunately it is not possible to use ionization energy data to rationalise

Fig. 2.3 as they are not available for all of the An"'. We can make progress,
however, by considering the standard electrode potentials, E’, for the An** + &
— An* and An* + e — An* couples, which are plotted in Fig. 2.4.
Although they are not as direct a measure as ionization energies, these data
may nevertheless be used as an approximate guide to the stability of An’*,
An*, and An*.

Th Pa U Np Pu Am Cm Bk Cf Es Fm Md No Lr

Fig. 2.4 Electrode potentials for An**/An** and An**/An?*.

The trend in the An**/An** couple is one of decreasing stability of An* as

the series is crossed. Thus, while +4 is practically the only oxidation state
available for thorium, and is perfectly stable for protactinium, uranium, and
neptunium, Am** and Cm** are found in solution only as fluoride complexes.
The drop in An**/An** potential at berkelium is presumably a reflection of the
small J, of this element (Bk* has the (Rn]5f’ configuration). The trend in
An**/An”* potential is similar to that in An**/An**, with the +2 oxidation
state becoming increasingly stable as the series is crossed. Note that the
discontinuity in the plot appears one element earlier than in the An**/An**
data; Cm** has the half-filled Sf shell and hence /, is small for this actinide.
The greater range of oxidation states exhibited by the early actinides in
comparison with their lanthanide counterparts indicates that the valence
electrons are less tightly bound and therefore more available for bonding. All
of the valence electrons may be used 1n bonding for all the actinides up to and
including neptunium, whereas only cerium can achieve its ‘group valence’ in
the 4f series. This is partly due to the single radial node in the 5f atomic
orbitals, which makes them less able to penetrate the core electrons than the
4f, which do not have a radial node. Relativistic effects (Chapter Three,
Section Two) also have an important role in determining the energies of the
valence electrons of the actinide elements.
12 Properties of the atoms and ions

2.3 Trends in metallic and ionic radii - the lanthanide and

actinide contractions
The variations in the radii of the metallic lanthanides and Ln** are shown in
Fig. 2.5. Both series display a gradual reduction in radius with increasing
atomic number but whereas the Ln** radii decrease uniformly from Ce** to
Lu**, there are marked breaks in the metallic radius trend at europium and
ytterbium. Figure 2.6 plots the analogous data for the metallic actinides and
the An**, An*, and An® ions. As with the lanthanides, the ionic radii display
a gradual reduction with increasing atomic number, while the trend in metal
radius is more complicated.

xs ‘ > z
A

a 20 8 +—e_ ee «©
160 Metallic Ln

(pm)
Radius

Ce Pr Nd Pm sm Eu Gd Tb Dy Ho Er Tm Yb Lu

Fig. 2.5 Variation of metal radius and +3 ionic radius for the lanthanide elements.

The reduction in the lanthanide metal and Ln** radii with increasing atomic
number is often referred to as the lanthanide contraction, and the
corresponding effect in the 5f series as the actinide contraction. These
contiactions arise from the poor ability of f electrons to screen the other
valence electrons from the nuclear charge. Although the additional unit of
nuclear charge on moving one element to the right is exactly balanced by the
opposite charge of an extra f electron (such that the total charge on the atom
or ion does not change), the poor screening of the nucleus by the additional f
electron means that the effective nuclear charge experienced by all of the
valence electrons increases slightly, and the atom/ion contracts.
The poor screening ability of f electrons is primarily a consequence of their
high angular nodality. However, relativistic effects (see Chapter Three,
Section Two) also play a part. Calculations reveal that 4f and 5f electrons are
expanded and destabilised with respect to ficticious atoms in which the effects
of relativity are not included, and are even poorer at screening the nuclear
charge than would be anticipated purely on the basis of their angular nodality.
 The f elements 13

The lanthanide contraction has important consequences for the chemistry of

the third row transition metals. It might be anticipated that these elements
would be larger than their second row counterparts by an amount similar to
the increase in size on moving from the first to the second transition series.
However, the reduction in radius caused by the poor screening ability of the 4f
electrons means that the third row transition metals are approximately the
same size as their second row congenors. Indeed, it has recently been shown
that the covalent radius of gold is actually less than that of silver (1.25 vs
1.33 A). Thus the lanthanide contraction is responsible for the many
similarities in the chemistry of the second and third row transition metals.

190

a ye ee
he ~—— Metallic An

(pm)
Radius

Th Pa U Np Pu Am Cm Bk Cf Es

Fig. 2.6 Variation of metal radius and +3, +4, and +5 ionic radius for the actinide elements.

The effects of the actinide contraction on the size and chemistry of the 6d
series 1s difficult to ascertain as the transactinide elements may only be made
in very small quantities and decay radioactively with very short half-lives.
Nevertheless, calculations suggest that the actinide contraction may have an
even greater effect than the 4f equivalent. For example, element 111 (which
lies under gold) is calculated to have a radius even smaller than copper!
Calculations also reveal that the role of relativity in the actinide contraction is
appreciably greater than in the lanthanide contraction.
All of the lanthanide metals except europium and ytterbium may be
considered to consist of Ln** ions with three electrons per atom devoted to
metallic bonding. Europium and ytterbium, however, are best regarded as Ln’*
(which are larger than Ln**) with only two electrons per atom involved in
metallic bonding. This accounts for the significantly greater metallic radii of
europium and ytterbium with respect to the other lanthanides and is discussed
in more detail in Chapter Four, Section Three. A similar effect 1s believed to
be the cause of the trend in the metallic radius of the actinide elements (Fig.
2.6). As shown in Fig. 2.3, the most stable actinide oxidation state increases
14 Properties of the atoms and ions

from +4 to +6 between thorium and uranium, and it is likely that there is a

concomitant increase in the number of electrons devoted to metallic bonding.
After plutonium the metallic radius increases significantly again, which may
well reflect the increasingly lanthanide-like behaviour of the later actinides in
that the preferred oxidation state is +3 with only three electrons per atom
available for metallic bonding.

2.4 Applications of the radioactivity of the actinides

All of the known isotopes of the actinides are radioactive, with a trend toward
decreasing half-life wth increasing atomic number. Table 2.3 indicates the
half-lives and radioactive decay mechanisms of the longest-lived isotopes of
each actinide element.

Nuclear weapons and nuclear power

Nuclear fission
Both the military and peaceful uses of actinide-based nuclear energy have as
their basis the same physical process, nuclear fission. Nuclear fission occurs
when a large nucleus splits into two smaller ones, a process which also
releases one or more neutrons which may collide with further nuclei causing
them to split and generate yet more neutrons. This is shown schematically for
the fission of a **°U nucleus in Fig. 2.7.

wy Ya
=e = — =e

O*
t Ry \\

Fig. 2.7 Schematic representation of the fission of a *°U nucleus, showing one of its many fission patterns. This process releases
2.1x 10" kJ mol’. Over 200 different isotopes of 35 different elements have been found in the fission products of 25U.
 The f elements 15

Element Mass of most Half-life Decay

stable isotope(s) | mechanism*
90 Th 232 1.40% 10" y Oo
onPa 231 3.25 x10" o
92 U 234 2.45 x 10° y a
235 7,037 * 10° y o
238 AAT XID ¥ o
93 Np 236 135% 10 y cipal 2h@
237 214 210" ¥ Oo
94 Pu 239 241 x 10:79 o
240 6.563 x 10° y Oo
242 3.76 x 10° y Oo
244 8.26 x 10’ y 7
95 Am 241 432.7 y a
243 7.38 10 y o
96 Cm 244 18.11 y o
245 8.5 x 10° y o
246 AT3.9¢- 10" y a
247 156.5:107 ¥ o
248 3:4 X10" y 7
250 ca.1 X10 ¥ SF
97 Bk 247 138 < 10" y a
249 320 d B
98 Cf 249 351 y a
250 Py a
251 898 y o
props 2.64 y a
OO Tis poe 472d o
253 20.47 d o
254 276 d Oo
255 39.8 d B
100 Fm 257 100.5 d o
101 Md 258 56d or
102 No 299 Lh a, EC
103 Lr 262 3.6 h (es
“ Radioactive decay mechanisms. &: ejection of ae from the nucleus. (i:
ejection of an electron from the nucleus. EC (electron capture): capture of an
inner core electron by the nucleus. SF: spontaneous fission.

Table 2.3 Half-lifes and radioactive decay mechanisms of the most stable isotopes of the
actinides.

For heavy nuclei such as the actinides, nuclear fission is an exothermic

process and under certain conditions a self-sustaining chain reaction can occur.
One such condition is that there is sufficient mass of fissile material to
prevent excessive loss of neutrons through the surface. This minimum mass
is referred to as the critical mass and for quantities in excess of the critical
mass (a supercritical mass) very few neutrons escape the surface. The chain
reaction rapidly increases the number of fissions, and may lead to a nuclear
explosion.
16 Properties of the atoms and ions

In 1939 Albert Einstein wrote to This was the principle behind the first nuclear weapons developed at Los
President Roosevelt of the USA Alamos in the USA and used to such devastating effect at Hiroshima and
emphasising the potential military
Nagasaki in Japan at the end of the Second World War. For example, the
applications of nuclear fission and
raising the possibility that the Nazis ‘Little Boy’ bomb that was dropped on Hiroshima consisted of two subcritical
could develop an atomic bomb. masses of *°U (i.e. two amounts each less than the critical mass) which were
Roosevelt decided that it was vital that slammed together using conventional chemical explosives to create a single
the USA investigate the viability of supercritical mass. The uncontrolled chain reaction rapidly led to a nuclear
such weapons, and in late 1941 the
explosion equivalent to the detonation of 20 000 tons of TNT.
go-ahead was given for the ‘Manhattan
Project’, the goal of which was to build
a bomb based on the fission process. Nuclear power generation
The fissile properties of *°U are put to altogether more peaceful use as the
power source in nuclear power stations. **U is, for all practical purposes, the
only naturally occuring fissile nucieus. However, while natural abundance
uranium is capable of sustaining a fission chain reaction, the low
concentration of **°U (0.72%) is such that by the time the effects of the fuel
cladding and the other reactor materials are taken into account, it is
advantageous to increase the proportion of **°U in the nuclear fuel. Hence
nuclear fuel is typically composed of UO, enriched to 2-3% **U.
Figure 2.8 is a schematic diagram of a typical nuclear reactor. Aside from
the fuel rods (zirconium or stainless steel tubes containing UO, pellets), the
other principal features are the circulating cooling fluid, the control rods, and
the moderator. The kinetic energy of the fission products is dissipated by
collisions with surrounding atoms, releasing huge amounts of heat (ca. 10°
times that produced by burning an equivalent mass of coal). In a reactor, this
Control
rods
heat is absorbed by the cooling fluid (usually water or heavy water, D,O)
which is subsequently used to drive steam turbines, thereby generating
electricity. The control rods are good neutron absorbers - usually boron steel
or boron nitride - and are used to regulate the flux of neutrons and thus
Fuel prevent the reactor from overheating. Although the **°U concentration is not
rods large enough for a nuclear reactor to explode like an atomic bomb,
overheating can cause sufficient damage for radioactive materials to be released
into the environment, as happened at Chernobyl in the former Soviet Union
in 1986,
The neutrons produced by **U fission are highly energetic (ca. 2 x 10° kJ
mol’), which is problematic in that they are not very good at causing fission
in other ~*°U nuclei. Fission is much more efficient with less energetic
Circulating neutrons, ideally ‘thermal’ neutrons with energies of ca. 2 kJ mol’.
cooling fluid/
Moderators are therefore used in nuclear reactors to slow down the *°U
moderator
neutrons. The best moderators are light nuclei such as '*C or 7H (D), and thus
in heavy water reactors the cooling fluid may also act as the moderator.
Fig. 2.8 Schematic diagram of a
nuclear reactor,
Separation of the isotopes of uranium
One of the greatest difficulties faced by the scientists working on the
Manhattan Project was to obtain enough **U to build a nuclear device. The
first attempts to separate *°U from **U were made by Ernest Lawrence at the
University of California in Berkeley. Lawrence’s approach was to take UCI,
and pass it through a cyclotron-like device in which ionized *°UCI, followed
a slightly different path from the heavier *“U species. Unfortunately the
 The f elements 17

process did not work at all well, and only about | g of **°U was produced by
Lawrence.
The technique that was developed as an alternative relies on gaseous
diffusion. This approach, which has been extensively used in the production
of **°U enriched fuel for nuclear reactors, passes gaseous UF, through porous
metallic membranes (usually composed of nickel or aluminium) at 70-80 °C.
Graham’s law states that the rate of diffusion is inversely proportional to the
square root of the relative molecular mass, and thus repeated passes through
the membranes results in the mixture of “*UF, and **°UF, becoming richer in
the lighter molecules. Up to 3000 passes are made in practice, leading to a
90% concentration of *°UF,. Unfortunately gaseous diffusion plants are very
large and expensive to run. All of the materials that come into contact with
the UF, must be fluorine resistant, and the metal membranes must be
manufactured to high tolerances. Furthermore, the pumping of the UF,
through the plant is energetically very demanding.
Another method of isotope separation involving UF, is the gas centrifuge.
If UF, is spun in a gas centrifuge the heavier ***UF, will concentrate toward
the walls and the UF, in the axial position. The higher the rotation speed
and the lower the temperature the better the separation.
Lasers have also been used to separate *°U from ***U. One approach is
based on the fact that **°U has a slightly different ionization energy from **U,
and hence the irradiation of uranium vapour with a laser whose wavelength ts
tuned to the ionization energy of **°U will produce exclusively *°U*, which
can then be collected at a negatively charged electrode.

Fast breeder reactors

The scarcity of *°U has lead to extensive efforts to develop nuclear power
based on two other fissile nuclei, **U and *’Pu. Although these do not occur
naturally, they may be produced artificially in nuclear reactors from **’Th and
>38U) respectively (nuclei that are far more abundant than **°U).

aethtpies
ool
“ppt grat U},5 = 22 min) (2.1)
233Pa—733U +B (t,,. = 27days)
SOS errs B
“(t,, = 24 min)
3Np
233U-373+B (2.2)
233Np—754Put+B (t,,, = 2.36days)
The vast majority of the uranium in conventional reactor fuel is **U.
Hence some *°’Pu is produced in all operating reactors (via the processes
shown in Egn 2.2) although the neutron yield from the 2-3% *°U is
insufficient to produce significant quantities of plutonium. However, if more
neutrons were available the production of **’Pu from ***U would increase to
the point at which it would become greater than the consumption of *°U.
This may be achieved in practice by removing the moderators in conventional
reactors and enriching the fuel to a greater extent than usual, to produce the
18 Properties of the atoms and ions

so-called ‘fast breeder’ reactors. The advantage of such reactors is that both the
>33UJ and **8U are used to produce heat, allowing up to 60 times the energy to
be extracted from naturally occuring uranium. However, not only does the
design and construction of fast breeder reactors pose many difficult technical
problems, but there are serious political objections to the use of a technology
which generates fissile >*°Pu, as it may be used to make nuclear weapons (the
‘Fat Man’ bomb dropped on Nagasaki was a **’Pu device). Thus fast breeder
reactors remain at the prototype stage. 4

Nuclear fuel reprocessing and nuclear waste storage

As fission proceeds in the fuel rods of a nuclear reactor, the fission products
increase in concentration. Many of these fission products are themselves good
As illustrated for °5U in Fig. 2.7, a
fissile actinide nucleus will usually
neutron absorbers and bring the chain reaction to a halt before all of the fuel
split into two nuclei of unequal mass, is used up. In some countries (é.g. France and the UK) the fuel rods are
one in the range 90-100 mass units removed periodically and the useful uranium and plutonium separated from the
and the other in the range 130-145. fission products and transplutonium elements (fuel reprocessing). Figure 2.9
Many of these fission products are is a flow scheme for the separation of uranium and plutonium from the fuel
themselves radioactive.
rods. It must be emphasised that the spent fuel rods are highly radioactive and
must be handled by remote control throughout. Care must also be taken to
avoid the accumulation of critical mass quantities of actinide (as low as 500-
1000 g of plutonium in saturated aqueous solutions). Reprocessing is also
complicated by the fact that the separations are often not complete and must
be carried out several times (e.g. ruthenium, a major fission product, binds
quite strongly to tributylphosphate and thus has appreciable solubility in the
organic phase as a nitrosyl complex).
In many countries, notably the USA, reprocessing is not considered to be
economically viable, and the spent fuel rods are currently kept at reactor sites
until a long term storage location is found. Indeed, the long term storage of
highly radioactive nuclear waste is one of the greatest environmental
problems faced by the World today. The most promising solution is to take
dried nuclear waste and to calcine and heat it with ground glass ‘frit’ to
produce a borosilicate glass. This glass will then be buried deep underground.
In 1982, the United States Congress passed the Nuclear Waste Policy Act that
requires the location of a suitable storage site and the implementation of an
underground disposal facility. Yucca Mountain in southern Nevada has been
targeted as a possible location and the Yucca Mountain Site Characterization
Project has been set up to determine if the site is suitable for a spent nuclear
fuel and high-level radioactive waste repository. The project involves
extensive study of Yucca Mountain’s geology, hydrology, biology, and
climate. A suitable repository must be located in a stable geological
environment (i.e. a region of low seismic activity) as the radioactivity of the
waste will not decay to safe levels for many thousands of years. Furthermore,
it is important that the movement of water in Yucca Mountain and its
surrounding region be thoroughly understood so as to minimise the
possibility that the repository will be breached by water, as the transport of
radioactive nuclei in aqueous solution is potentially one of the most
hazardous aspects of long term waste storage.
 The f elements 19

Irradiated fuel
elements

‘Cooling off’ period (up to 100 days). Rods are immersed in ponds of water to allow the intensely
radioactive but short-lived decay products such | as '3'| (t,.=8.04 days) to lose most of their activity

Rods dissolved in 7 M HNO3 |

Solution of mainly UOz(NO3)9,

TBP = ("BuO)3PO
Pu(NOs3),4 and other metal nitrates

Aqueous solution of
Solution treated with TBP in kerosene fission products and
transplutonium elements

[UO,(NOs)a(TBP)2] and
[Pu(NO,) ,(TBP),] in kerosene

Fe(NHzSO3)o, hydrazine, or hydroxlamine nitrates | reduce Pu(IV) to Pu(Ill) but do not affect U(VI)

[UO »(NO3)o(TBP)>] Aqueous solution

in kerosene of Pu(Ill)

0.2M mae) |ae with

UO2(NO3)z (aq) gb. (aq)

evaporate |oxalic acid

UOx(NO3)2.6H20 Precipitate of
Pu(Cp O4)2.6H2O

400 cal
300 °C

[ra]
UO;

H,/700 °C

UO,

Fig. 2.9 Flow scheme showing the principal steps in nuclear fuel reprocessing.
20 Properties of the atoms and ions

Other applications of actinide radioactivity

Although the use of uranium and plutonium in nuclear power generation and
nuclear weapons is by far the most important application of the actinides,
there are other, less contentious uses of actinide radioactivity. For example,
the heat generated as a by-product of the intense @ particle radioactivity of
*38Pu is coupled with PbTe thermoelectric elements to produce batteries which
are used to power heart pacemakers (the regular electrical pulses are used to
stimulate the heart muscle). The lifetime of the nuclear powered batteries is
about five times longer than conventional devices. Similarly, ~*PuO, is
employed on the kilogram scale as the power source for spacecraft, including
many Earth satellites and probes to other planets in the Solar System.
**1Am is used as the ionization source in smoke detectors and thickness
guages. In smoke detectors, the @ particles emitted by the **'Am collide with
the oxygen and nitrogen in the air in the detector’s ionization chamber to
produce ions. A low-level electric voltage applied across the chamber is used
to collect these ions, causing a steady small electric current to flow between
two electrodes. When smoke enters the space between the electrodes, the a
radiation is absorbed by smoke particles. This causes tne rate of ionization of
the air and therefore the electric current to fall, which sets off an alarm.

2.5 Exercises
1. | Radioactive decay follows first order kinetics, i.e. ifN is the number of
atoms of a particular isotope at time rt, N, the number at time r=0, and
ty is the half-life, then
0.693
N = No exp(- yt
fy
How long does it take for a sample of **°Pu to decay to 10% of its
activity? (Note that the activity of a radioactive sample - the rate at
which it decays - depends only on the size of the sample).

tO How many (a) radial and (b) angular nodes do the following atomic
orbitals possess?
St 75 Gd 6p 7p

3. Rationalise the trends in the following first ionization energies (kJ

mol!')
Ca 590 Zn 906
Sr 549 Cd 867
Ba 503 Hg 1007

4. Nuclear fission is not the only nuclear process on which the generation
of electricity may be based. Research continues into the feasibility of
using nuclear fusion in electricity production. What are the principal
features of nuclear fusion, and for which elements is it an exothermic
process?
3 Relativity, electronic
spectroscopy, and magnetism

3.1 The chemical consequences of relativity

Relativity is most commonly associated with the world of the very large. For
example, the motions of the stars and planets are governed by the laws of
gravity which are contained in the theories of relativity developed by Albert
Einstein in the early part of the 20" century. It is less well known, however,
that relativity also has a significant impact at the atomic level. Indeed it is
ironic that Paul Dirac, who in the late 1920s unified the theories of special
relativity and quantum mechanics, believed relativistic effects to be ‘of no
importance in the consideration of atomic and molecular structure and ordinary
chemical reactions’. This statement has proved to be some way off the mark,
and there is now no doubt that the chemical consequences of relativity are
very significant, in particular for heavy elements such as the actinides. These
consequences may be divided into two main areas - the modification of atomic
orbital energies and the effects of spin-orbit coupling - and each is described
below.

3.2 The modification of radial atomic wavefunctions and

associated energies
Einstein’s Special Theory of Relativity tells us that it 1s impossible to
accelerate a particle to the velocity of light or beyond. The mass m of a
particle of rest mass mj moving with velocity v is given by Eqn 3.1, where c
is the velocity of light. Clearly, as v — c, m — ©, and this increase in mass
is known as the relativistic mass increase.
Mo
m = ———— (Gall)

ral
In atomic units, the average radial velocity, <v,,y>, of the electrons in the
ls shell of an atom is approximately Z, where Z is the atomic number. For
uranium, for which Z = 92, <v,,,>/c is given by Eqn 3.2, where the velocity
of light is also expressed in atomic units.
a ere ee fe
ai — = —— = 0.67 (3:2)
e 137
The average relativistic mass increase of the Is electron in uranium is
therefore given by Eqn 3.3 where m, is the rest mass of the electron. As the
22 Relativity, electronic spectroscopy, and magnetism

expression for the Bohr radius has a 1/m dependence, this mass increase
produces a marked contraction of the Is electron, and a concomitant energetic
stabilisation.

me = 135m, (3.3)
a A= 0:67"

The dependence of relativistic mass corrections on atomic number means

that for light elements the effects of relativity are generally small. This is
clearly not the case for heavy elements, however, and the consequences of
relativity for the valence electronic structure of such atoms are often profound
(in percentage terms, the modification of the valence electron radial
distributions and energies in heavy elements upon the inclusion of relativity
is often greater than that of the core orbitals). Although there are many
competing factors that determine the precise valence electronic structure of a
heavy element, several general trends may be identified. The s functions in
higher primary quantum shells must be orthogonal to the Is orbital, and
hence the relativistic stabilisation of the Is orbital results in a stabilisation of
all of the other s functions in the atom. This is known as the direct
relativistic orbital contraction. Similar, though smaller, effects are
experienced by p electrons. By contrast, relativistic valence d and f electrons
are expanded and destabilised with respect to their non-relativistic analogues.
This arises from the increased shielding of the nucleus by the outer core s and
p electrons of similar radial distribution to the d and f functions, and is known
as the indirect relativistic orbital expansion.
In very heavy elements such as the actinides, the relativistic expansion and
destabilisation of the valence f orbitals is sufficient to alter their chemistry
markedly in comparison with (hypothetical) non-relativistic analogues. For
example, non-relativistic calculation of the 5f atomic orbitals of uranium
yields a binding energy of 1665 kJ mol'', whereas an analogous relativistic
approach produces 869 kJ mol'!. Furthermore, because the actinide 5f orbitals
are more destabilised by relativistic effects than are the lanthanide 4f orbitals,
they are more weakly bound and hence more chemically active. The larger
range of oxidation states observed for actinides (see Chapter Two, Section
Two) and their greater tendency to form covalent bonds in comparison with
the lanthanides (see, for example, Chapter Six, Section Six) may therefore be
traced (at least in part) to the greater effects of relativity in the Sf series.

3.3 Spin-orbit coupling

The time-dependent Schrodinger In order to make quantum theory compatible with relativity, Dirac suggested
equation is incompatible with relativity
an alternative form of the Schrdédinger equation. One of the triumphs of
because while it contains second
order partial derivatives with respect
Dirac’s equation is that it leads in a natural way to electron spin, which has to
to the three spatial dimensions, it has be treated rather clumsily as an ‘extra’ in non-relativistic quantum mechanics.
a first order partial derivative with When an electron is part of an atom, the angular momentum associated with
respect to time. The Special Theory of its spin couples with that generated by its orbital motion. The factors which
Relativity demands an even-handed
govern the magnitude of this spin-orbit coupling in many electron atoms are
treatment of space and time.
 The f elements 23

complicated, but in general it increases significantly with increasing nuclear

charge and for a given primary quantum shell decreases in the order p > d > f.
We define and describe the states produced by spin-orbit coupling (and their
associated energies) according to their spin and orbital angular momenta.
There are two limiting approaches, both of which are based upon addition of
the vectors describing the orbital and spin angular momenta. One approach -
LS or Russell-Saunders coupling - is most appropriate when spin-orbit
coupling is weak in comparison with interelectronic repulsion. The other
limiting approach is more appropriate when the spin-orbit interaction is
strong in comparison with the electrostatic interactions between the electrons,
and is known as j-j coupling. Both of these approaches are now discussed.

Russell-Saunders coupling
Russell-Saunders coupling considers that the individual orbital angular
momenta of all the electrons combine into a total atomic orbital angular
momentum with quantum number L. Similarly, the individual electronic spin
angular momenta combine to yield a total spin angular momentum for the
whole atom with quantum number S. The total atomic angular momentum is
given by the coupling of L and S and is described by the quantum number J.
Atomic or ionic energy levels are characterised by a term symbol, of
general form
Cen
J

To obtain a term symbol, the value of L is determined, and assigned the

appropriate letter from the series below

i O12 3.4.5 6.7

symbol SP DPGHIK

The value of S is then used to obtain the spin multiplicity, (2S+1), of the
term. Finally the permitted values of J are determined according to Eqn 3.4,
and are known as the levels of the term.
PES, LRA (Leo! (3.4)
The most important term of an atom or ion is the ground term (the most
stable term). Fortunately there is a simple way to determine atomic/ionic
ground term symbols, using the three rules due to Hund.

Rule 1: The ground term always has the largest value of S. This rule is
known as the rule of maximum multiplicity.
Rule 2: If two terms have the same multiplicity, the one with the highest
value of L lies lowest in energy.
Rule 3: For electronic subshells that are less than half full, the level with the
lowest value of J lies lowest in energy. For greater than half-filled
subshells, the level with the highest value of J lies lowest in energy.

The electronic structures of lanthanide atoms, and in particular Ln** ions,

are well described by the Russell-Saunders coupling scheme. As an example,
24 Relativity, electronic spectroscopy, and magnetism

let us determine the ground term arising from the [Xe]4f? configuration of
Dy**. The best way to do this is to use Hund’s rules and the
‘electrons-in-boxes’ approach, in which the individual f orbitals are
represented by boxes and the electrons by arrows (up-and down arrows are used
to represent up spin (m, = +1/2) and down spin (m, = -1/2) electrons
respectively). The arrangement of the nine 4f electrons of Dy** which satisfies
Hund’s first two rules is shown in Fig. 3.1.

Fig. 3.1 The most stable arrangement of the nine 4f electrons of Dy**.

This arrangement has the maximum number of unpaired electrons, five, and
hence an S value of 5/2 (the sum of all of the individual electron m, values)
Note that we need consider only the
and (2S+1) = 6. The value of L in the above arrangement is five (the sum of
partly filled 4f subshell to obtain the
ground level. This is because all filled the individual electron m, values), and hence the ground term is °H. J can take
electron shells and subshells have no integer values from 5+5/2, 5+5/2-1......, |5-5/2| and because Dy* has a
net angular momentum and hence do greater than half filled 4f shell, the level with the highest J value (15/2) lies
not contribute to the term symbol.
lowest in energy. Thus the full symbol for the ground level of Dy** is °H,<,.
The ground levels of all of the Ln** ions are given in Table 3.1.

L.,/Bohr Magnetons
Ln* Electronic Ground Colour Calculated Observed
Configuration _ level (Eqn 3.5)
Ce* — [Xe]4f' "Bos Colourless 2.54 ya
oesNe:
Pr* [Xe]4f 3H, Green 3.58 3.4-3.6
Nd* — [Xe]4f? 4, Lilac 3.62 Dea
Pm* [Xe]4f* y, Pink 2.68 :
Sm* — [Xe]4f? 23 Yellow 0.85 (Aa7
Eu* [Xe]4f° usr Pale pink 0 33255
Gd* — [Xe]4f’ a Colourless 7.94 7.9-8.0
Th [Xe]4f* USP Pale pink 9.72 9.5-9.8
Dy* [Xel4f? Hg, Yellow 10.65 10.4-10.6
Ho* =‘ [Xe]4f"° *t. Yellow 10.60 10.4-10.7
Et [Xe]4f" “Hisp Rose-pink 9.58 9.4-9.6
Tm* = [Xe]4f”” 7H Pale green 7.56 That 5
Yb* =‘ [Xe]4f "sys Colourless 4.54 4,3-4.9
Lu*__ [Xej4f* Sa Colourless _0 0

Table 3.1 Spectroscopic and magnetic properties of Ln* ions in hydrated salts.

J-j coupiing
In this coupling scheme, the individual electronic orbital and spin angular
momenta combine to give a total angular momentum for each electron,
 The f elements 25

denoted j. The j values then couple to produce the total atomic angular
momentum J.
Spin-orbit coupling in actinide atoms and ions and their compounds is
much greater than for the lanthanides, to the point that the Russell-Saunders
coupling scheme is much less valid. It would be both elegant and convenient
if we could treat actinide spin-orbit coupling using the j—j scheme, but
unfortunately a purely j—j based approach does not work either. This is partly
because even in actinide systems spin-orbit coupling does not dominate over
interelectronic repulsions, and partly because the 5f orbitals are much more
sensitive to atomic/ionic environment than are the lanthanide 4f orbitals
(particularly the 5f orbitals of the early actinides). Experimental studies of
actinide spin-orbit coupling, for example electronic spectroscopy and
magnetic measurements, are consequently more difficult to interpret than
analogous lanthanide data. The spectroscopic and magnetic properties of the
lanthanides and actinides are discussed in more detail in the following
Sections.

3.4 Electronic absorption spectroscopy

The excitation of an atom, ion, or molecule from its ground electronic level
to higher lying levels may be brought about by the absorption of
electromagnetic radiation. Electronic absorption spectroscopy (sometimes
known as optical or ultraviolet/visible spectroscopy) is the study of these
photon/matter interactions, and has been extensively used to investigate
f element compounds.

Electronic spectroscopy of Ln**; comparisons with transition

metal spectra
The majority of the electronic transitions in Ln* ions involve only a
redistribution of electrons within the 4f subshell i.e. they occur between the
ground and excited levels arising from the ground electronic configuration
({(Xe]4f). As with the formally d — dé transitions of transition metal
compounds, electric dipole selection rules forbid these f — f promotions. In Both the Laporte (g — u) and the
transition metal compounds, these rules are relaxed primarily by vibronic Al=+1 selection rules forbid f > f
promotions. The former rule disallows
coupling. In this process a molecular vibration temporarily lowers the
transitions between levels with the
symmetry around the metal atom and, in the new (transient) geometry, one or same inversion symmetry, while the
more of the metal’s d orbitals shares the symmetry of a p orbital. The latter prohibits transitions between
transition acquires some d — p (or p — d) character and therefore gains levels with the same orbital angular
intensity. This coupling of the vibrational and electronic parts of the total momentum quantum number, /.
wavefunction of the system requires a significant interaction between the
transition metal d orbitals and the surrounding ligands. In Ln** ions, however,
the 4f orbitals are radially much more contracted than the d orbitals of
transition metals, to the extent that the filled Ss and Sp orbitals largely shield
the 4f electrons from the ligands. The result is that vibronic coupling is much
weaker in Ln** systems than in transition metal compounds, and hence the
intensities of electronic transitions are much lower. As many of these
electronic transitions lie in the visible region of the electromagnetic
26 Relativity, electronic spectroscopy, and magnetism

spectrum, the colours of Ln** compounds are typically less intense than those
of the transition metals.
The colours of the Ln** ions in hydrated salts are given in Table 3.1. The
lack of 4f orbital/ligand interaction means that the f — f transition energies
for a given Ln* change little between compounds, and hence the colours of
Ln* are often characteristic. A further consequence of the small interaction of
the Ln** 4f atomic orbitals with the surrounding ligands is that f — f
transition energies in Ln** compounds are well defined, leading to much
sharper bands in their electronic absorption spectra than are observed for
transition metal compounds. A typical spectrum, that of aqueous Pris
shown in Fig. 3.2, and the energies of the levels of free Pr** relative to the
3H, ground level are given in Fig. 3.3.
25000

3 P5

3 P,

20000

24 20 16 10 5 0

emi x 107

Fig. 3.2 Electronic absorption spectrum of aqueous Pr* [modified from Fig. 11.11 (b) of
15000
‘Physical Inorganic Chemistry’ by S.F.A. Kettle].

There is a complementarity between transition metal and _ lanthanide

electronic structure that makes the study of their electronic absorption spectra
€(Ss) particularly fascinating. In the transition metals, the effects of the surrounding
ligands upon the d orbitals - the crystal field - is very much greater than the
10000
spin-orbit coupling of the d electrons. The opposite is true for lanthanide
compounds, where 4f spin-orbit coupling dominates over crystal field effects
(typically 2000 cm’! vs 100 cm’'). Nevertheless the effects of the crystal field
3 F,
cannot be totally discounted in Ln** spectra, as evidenced by the so-called
‘hypersensitive’ bands in the electronic absorption spectra of some Ln*,
3
notably Nd**, Ho**, and Er**. The intensities and energies of these bands show
5000 Fo
a marked dependence upon the coordination environment of the Ln’,
implying that the crystal field can affect the energies of the spin-orbit coupled
4f levels in certain cases. It should be noted, however, that the mechanism of
SH.
this hypersensitivity 1s not well understood.
In addition to the f + f transitions discussed thus far, Ln** (notably Ce*
Sy
and Tb**) display electronic absorption bands which are more intense than the
0
f — f bands and which are usually found in the ultraviolet. These bands are
not due to f > ftransitions, but rather to [Xe]4f" + [Xe]4f"'Sd' promotions,
Fig. 3.3 Electronic energy level
diagram for Pr** ([Xe]4f). which are formally allowed by the electric dipole selection rules. The energy
of the [Xe]4f! > [Xe]4f°5d' transition in Ce** is particularly noteworthy
because of its strong dependence upon the environment of the metal ion. The
lowest excited [Xe]4f°5d! level of gaseous Ce** lies 49 737 cm’! above the
 The f elements 27

°F; ground level, but the f > d transition has been reported at only 22 000
cm! in Ce**-doped Y,A1,O,,. Even more remarkably, the peak at 17 650 cm’!
in the spectrum of [Ce{n°-C,H;(SiMe;),},] has also been assigned to the
f — d transition. The large differences in the energy of this transition for
different Ce** environments reflects the greater radial extension of the Sd
atomic orbitals with respect to the 4f, and their greater interaction with the
surrounding ligands.
Note that Ln** ions are often highly coloured. This arises because the 4f
orbitals in Ln’* are destabilised with respect to those in Ln**, and hence lie
closer in energy to the 5d orbitals. This change in orbital energy separation
causes the f — d transitions to shift from the ultraviolet into the visible
region of the spectrum.

Actinide electronic absorption spectra

As indicated at the end of Section 3.3, the electronic absorption spectra of
actinide compounds are more difficult to interpret than those of the
lanthanides. The increase in spin-orbit coupling and the greater interaction of
the Sf atomic orbitals with the surrounding ligands (especially for the early
actinides) result in J no longer being a good quantum number, and thorough
treatments of the electronic structure of actinide compounds must account for
mixing of the J levels obtained from the Russell-Saunders approach. In
practice this means that the interpretation of actinide spectra is best done on a
compound by compound basis, and alterations of the ligand set for a given
actinide element in a given oxidation state can significantly alter the
absorption spectrum.
It is, however, still possible to make some general observations about
actinide electronic absorption spectra. Although vibronic coupling is greater
in actinide compounds than in those of the lanthanides, f + f transitions are
still weak, albeit less so than for the lanthanides. The increased 5f
orbital/ligand interactions also result in actinide f — f bands being broader
than their lanthanide counterparts. f — d transitions also occur in actinide
compounds, and the smaller energy difference between the Sf and 6d atomic
orbitals of the actinides in comparison with the lanthanide 4f/Sd separation
results in these transitions lying at lower energy than in the lanthanides,
although they still generally fall in the ultraviolet. Another class of transition
that occurs in actinide compounds is ligand to metal charge transfer. The band
maxima of these transitions are generally in the ultraviolet, although the
bands can tail into the visible. As with charge transfer transitions in general,
these promotions are fully allowed, which accounts for the intense colours of
certain actinide compounds, in particular those with metals in high oxidation
states and easily oxidizable ligands.
Before leaving electronic absorption spectra, it is worth returning to one of
the central themes of this book; the similarities between the later actinides
and the lanthanides. The spectra of both BkCl, and CfCl, feature sharp, low
intensity peaks which resemble the spectra of Ln** much more than those of
the lighter actinides.
28 Relativity, electronic spectroscopy, and magnetism

3.5 Fluorescence of Ln**; the use of lanthanide ions in

colour television sets
The compounds of many Ln* ions fluoresce following stimulation by
ultraviolet irradiation or electrical discharge, the fluorescence arising from
f — f transitions within the Ln** ion. The mechanism for this is represented
schematically in Fig. 3.4, and is as follows. The ultraviolet light or electrical
discharge promotes a ligand-based electron from the singlet ground level into
one of the vibrational levels of an excited singlet. The compound then rapidly
relaxes to the ground vibrational level of the excited singlet, after which a
non-radiative intersystem crossing (ISC) occurs to one of the vibrational
levels of an excited triplet, which lies at slightly lower energy than the
excited singlet. Following relaxation to the ground vibrational level of the
triplet, a second intersystem crossing occurs to a nearby excited level of the
Ln* ion, which fluoresces back to the ground level via an f — f transition.

Excited
singlet
Excited
triplet

Excited
Ln3* level

Lns+

fluorescence

Singlet ground level

Fig. 3.4 Schematic representation of the mechanism of Ln** fluorescence.

This process is particularly favoured for Tb** and Eu**, which have excited
levels at slightly lower energy than the excited triplets of typical ligands. The
main emissions for Tb** are between the °D, and ’F, (n = 6-0) levels, and
generate green light, while for Eu** the °D, — ’F, (n = 4-0) transitions emit
red light. This fluorescence is employed in colour television sets, the screens
of which are made up of a large number of tiny clusters, each containing three
phosphor dots. The three dots in each cluster emit red, green, or blue light
respectively. The red phosphors are typically Eu** in Y,0,S or Eu**:Y,0,,
while one of the choices for green emission is Tb**:La,O,S. The best blue
emitter is Ag,Al:ZnS, which has no Ln** component. The television set has
three separate cathodes, one for each colour. A metallic mask behind the
screen has tiny holes in it which allow only electrons from the green cathode
 The f elements 29

to hit the green phosphor dots in each cluster, and similarly for the other two
primary colours.
There is currently a great deal of research into replacements for the cathode
ray tube in television sets. One promising prospect for the generation of
colour in flat panel displays are tris(pyrazolyl)borate compounds of cerium,
europium, and terbium (Fig. 3.5), which have the potential for bright, highly
efficient, durable, low power light emission with a narrow bandwidth over the
full spectral range.

Fig. 3.5 Tris(pyrazolyl)borate complexes of cerium, europium, and terbium are promising
phosphors for flat panel displays.

3.6 Neodymium lasers

Solid matrices containing Nd** ions can be made to exhibit laser action. The
matrix material is sometimes glass, but the most popular matrix is yttrium
aluminium garnet (YAG), Y3,Al,;O,,. The Nd* ions replace yttrium in the
lattice, up to a maximum doping level of about 1.5%. Nd* has the [Xe]4f°
electronic configuration, and the energies of the lowest lying terms and levels
that originate from this configuration are given in Fig. 3.6.
A typical Nd:YAG laser consists of a rod a few centimetres long with a
mirror at each end, one of which is partly transmitting. The rod is then
irradiated with a pulse of light from a simple lamp (e.g. a tungsten halogen or
high-pressure mercury discharge lamp) to promote the Nd™ into excited
levels, e.g. the “F,, or *F,,. These then undergo a rapid non-radiative decay to
the long-lived “F;,. level. The pulse of light is continued for several
milliseconds, in order to achieve a situation where the majority of the Nd** are
in the *F,. level (known as population inversion). If a photon of the laser
wavelength, generated by an excited Nd* ion relaxing to the “I,,, level,
encounters another *F,,, Nd**, the second ion will be stimulated to release a
photon of exactly the same wavelength and phase as the first, as it relaxes to
the “I,,, level. This stimulated emission rapidly depopulates the *F;,, level,
with the photons being reflected back and forth within the rod until an intense
beam of coherent, monochromatic (1.06 lum wavelength - near infrared)
radiation emerges from the end of the rod. The “I,,, level then rapidly relaxes
to the “I, ground level and the whole process can begin again. As the
depopulation of the upper lasing level is much more rapid than the time
required for a population inversion to be created, the laser output consists of a
series of infrared pulses.
30 Relativity, electronic spectroscopy, and magnetism

etc

“Fi 1/2 1.81

4F a0 1.66
4p a aaa es 1.54
“Fay 1.42

[Xe]4f8

4145/2 0.74

I Alaa 0.52

“lit 0.28

“loro 0

Electronic Interelectronic Spin-orbit Energy above

configuration repulsion coupling ground level (eV)

Fig. 3.6 Electronic energy levels of Nd*.

3.7 Magnetism

Magnetic behaviour of Ln**

The paramagnetism of Ln* ions arises from their unpaired 4f electrons which,
as has been stated, interact little with the surrounding ligands in Ln*
compounds. The magnetic properties of these compounds are therefore similar
to those of the free Ln** ions. We know that in the Russell-Saunders
appreach, spin-orbit coupling splits an atomic or ionic term into a series of
levels characterised by their J values. For most Ln** the magnitude of the f
orbital spin-orbit splitting is sufficiently large so that the excited levels are
thermally inaccessible, and hence the magnetic behaviour is determined
entirely by the ground level. The effective magnetic moment [U,, of this level
is given by Eqns 3.5 and 3.6.

Meg = 8, VHT +1) (3.5)

where 2, 3,
3
55+) a
Eiichi)
2 FFs) eae
 The f elements 31

Notice the similarity between Eqn 3.5 and the spin-only formula which
For first row transition metal
works so well for first row transition metal compounds. In these compounds
compounds the effective magnetic
the orbital contribution to {,, is quenched by the interaction of the d orbitals moment is well approximated using
with the surrounding ligands, so that only S$ is required (and hence only the the spin-only formula
number of unpaired electrons). The replacement of § by J in Eqn 3.5 is Hey = ajn(n + 2)
necessary because the surrounding ligands fail to quench the orbital where n is the number of unpaired
contribution to {l,, in Ln** compounds, owing to the 4f orbitals being so electrons.
contracted.
The calculated (Eqn 3.5) and observed u,, values for all of the Ln’* are
given in Table 3.1. It may be seen that there is good agreement between the
calculated and experimental values in all cases except for Sm** and Eu**. The
discrepancies for the latter arise because both ions have excited levels (°H,,
for Sm** and ’F, and ’F, for Eu**) which are sufficiently close to the ground
level to be thermally accessible. If allowance is made for this (by assuming a
Boltzmann population distribution over the energy levels) then calculated and
experimental [,, values once again agree.
The high J of the ground levels of the later Ln** result in very high [,
values (Table 3.1). As a consequence, placing salts of these ions in strong
magnetic fields leads to a slight warming because the stabilisation energy
given out as the salts are attracted into the field manifests itself as heat. This
effect is exploited for obtaining very low temperatures. Gadolinium and
dysprosium salts are typically used, and are cooled with liquid helium in the
presence of a strong magnetic field. When the system is at liquid helium
temperature the magnetic field is removed, which causes further cooling as the
salts lose their magnetic orientation, a process known as _ adiabatic
demagnetisation.
The strongest known permanent magnetic material is an alloy of
neodymium, iron, and boron, of chemical formula Nd,Fe,,B. The unit cell of
this material contains 68 atoms, with six distinct iron sites, two different
neodymium sites, and one boron site. The iron and neodymium sites each
have their own magnetic moments, and these align in the same direction to
produce a bulk magnetism more than 50 times that of steel. Compounds with
the general formula Ln,Fe,,B have been identified for all of the lanthanides
except promethium and europium, although none has the same permanent
magnetic strength as the neodymium alloy.

Magnetic properties of actinide compounds

The factors that account for the increased complexity of actinide electronic
absorption spectra with respect to their lanthanide counterparts are also
responsible for the correspondingly complicated magnetic behaviour of
actinide compounds. Eqn 3.5 is less applicable than for the lanthanides, and
there is a much greater temperature dependence of actinide magnetic
susceptibilities. This is illustrated by UF, and certain UO,’*-containing
species, all of which feature U and should therefore be diamagnetic (as are
compounds of Th** and Pa®*, which also have the [Rn]5f? metal configuration
and a 'S, ground level). However, these compounds exhibit temperature
independent paramagnetism (TIP) on account of the mixing of paramagnetic
excited levels with the ground level.
32 Relativity, electronic spectroscopy, and magnetism

One of the most widely studied actinide ions is U* ({(Rn]5f’). The magnetic
data obtained for U** compounds are usually interpreted by considering only
In spite of the reduced applicability of
the Russell-Saunders coupling
the 7H, ground level but unlike the Ln** ions, the effects of the surrounding
scheme to actinide electronic ligands must also be taken into account. The *H, ground level is spilt by the
structure, term symbols based upon surrounding ligands into several new energy levels, and the magnitude of the
this approach are often used as the splitting is comparable to thermal excitation energies. The interaction of the
starting point for discussions of the
5f atomic orbitals with the surrounding ligands therefore creates a range of
spectroscopic and magnetic
properties of actinide compounds.
thermally accessible excited levels from the free ion ground level.
There have been several magnetic studies of [NEt,],[U(NCS)g], which
above 30 K contains U** in a site of cubic symmetry surrounded by eight
nitrogen atoms. Magnetic data have been used to show that below 30 K, the
geometry distorts to Dy, around the U**. [U(n?-C,H;);R] (R = BHy, BF,, OR,
F, Cl, Br, I) have also received a good deal of attention, and have been divided
into two categories on the basis of their magnetic behaviour: (a) molecules
with small dipole moments and a small range of TIP and (b) molecules with
larger dipole moments and a more extended range of TIP. The differences
between the two types of behaviour are attributed to an increasing trigonal
distortion for molecules in category (b).
The magnetic properties of the actinidocenes - [An(n*-C,H,).] - are
discussed in Chapter Six, Section Six.

3.8 Exercises
1. | Use Hund’s rules to determine the Russell-Saunders ground levels of

U2 Sm?+ Tm?* Bk** Ce**

2. The table below shows the extinction coefficient € of the strongest

absorption band in the electronic spectra of four species, labelled A—D.
A-D are [NiCl,]”, Pr**(aq), [Mn(H,0),]?*, and [MnO,], but not in that
order. Match A-D to the correct compound.

Compound _|_edm? mol! cm’!

A > 1000
B 100
Cc ital
D 0.03

3. Why are aqueous solutions of Ce** colourless but those of Ti** purple?

4. Why are there four f — ftransitions in the electronic absorption spectra

of U** in octahedral environments?

5. Define the terms diamagnetism, paramagnetism, ferromagnetism,

antiferromagnetism, and antiferrimagnetism.
4 Solid state compounds

In this Chapter we will focus on two classes of binary compounds - halides

and oxides - which between them illustrate many of the important trends in f
element solid state chemistry, and which may be used to compare the
lanthanides with the actinides. The Chapter concludes with an introduction to
the factors which determine the electrical properties of solids, and applies the
ideas developed to elemental lanthanides and divalent lanthanide compounds.

4.1 Halides

Trihalides of the lanthanides

Anhydrous trihalides are known for all of the lanthanides. They are ionic,
crystalline substances with high melting points whose principal use is as
starting materials for the synthesis of pure lanthanide metals and other
lanthanide compounds. Apart from the trifluorides, they are all highly
deliquescent.
Addition of HF to aqueous Ln(NO,), results in the precipitation of
LnF,.1/2H,O. Heating of these hydrates leads to anhydrous lanthanide
trifluorides, which may also be prepared by the reaction of HF with Ln,O;:

300 °C, in vacuo

LnF3.1/2H,O0
or 600 °C/HF(g)

LnF3

HF(g)
Ln2O03
700 °C

LnF,; (Ln = La—Pm) adopt the ‘tysonite’ (LaF,) structure (Fig. 4.1) in
which the Ln** is coordinated by nine F in a tricapped trigonal prismatic
arrangement, with a further two F at a slightly greater distance. Beyond Fig. 4.1 The LaF, (‘tysonite’)
promethium, all of the LnF, have the YF; structure which features eight close structure. The shaded circles
represent the two more distant fluorine
Ln3*-F contacts at ca. 2.3 A and one longer Ln**-F' distance (ca. 2.6 A),
atoms.
with the F in an approximately trigonal prismatic arrangement around the
Ln*. This reduction in the primary coordination number from nine to eight is
a result of the decreasing size of the Ln**.
LnF, have several technological uses, including thin film coatings of
optical elements, high pressure anti-wear lubricants, and host lattices for
phosphors and scintillators.
34 Solid state compounds

Heating of the hydrated lanthanide trichlorides, tribromides, or triiodides

generally results in the formation of oxyhalides, and hence other routes to
LnX, (X = Cl, Br, I) must be followed. A completely general method is the
direct combination of the elements, but there are also many other approaches.
For the trichlorides these include:

Cel

HCI(g)
LnCl3.xH,O
105-350°C
LnCl,

xs NH,Cl 300 °C, in vacuo ~~

LnCl3.xH2O —— (NH4)3LnCle

LnCl3.xH,O
reflux

LnCl, (Ln = La—Gd) adopt the nine coordinate UCI, structure, a tricapped
trigonal prismatic arrangement which is like the structure of LaF, but with
the two’ more distant F removed. This structure is also adopted by
{Ln(H,0),]** of the early lanthanides. TbCl, has the eight coordinate PuBr,
structure (which may be regarded as the UCI, structure with one of the
capping Cl removed) and all of the other lanthanide trichlorides the six
coordinate AICI, structure. Thus the trend toward lower coordination number
with decreasing Ln** radius, noted above for LnF;, is also apparent in the
trichlorides. The size of the anion is also important in determining the
coordination number of a given Ln**, with a trend toward decreasing
coordination number with increasing anion radius.
Routes to LnBr, and Lnl, include:

100 °C, in vacuo NH,

LnBr3.6H,Q ———> Ln Br (Gd-Lu) Lnl3.xH,0 ———_——_____ Ln,
heat

HBr(g) HI/H>
n> nis Lnclj. —————> Ll,
400-600 °C heat

There are three different structural types among the LnBr,. LaBr,, CeBr,,
and PrBr; ali have the UCI, structure, while the tribromides of neodymium-—
europium adopt the PuBr, structure. The remaining LnBr, have the six
 The f elements 35

coordinate FeCl, structure. LnI, have either the PuBr, (Ln = La—Pm) or FeCl,
structure.

Dihalides of the lanthanides

By contrast to the trivalent compounds, only a few lanthanides form dihalides
with all of the halogens. The known LnX, are shown in Table 4.1. There is a
variety of preparative routes to LnX,, including reduction of the appropriate
trihalide with elemental lanthanide, H,, or an alkali metal and, in some cases,
thermal decomposition of the trihalide (e.g. NdCl,, SmI, Eul,, and YbI,). As
shown in Table 4.1, LnX, adopt a wide range of structural types, and exhibit
the same trend toward lower coordination number with decreasing cation
radius and increasing anion radius noted for LnX,. The diiodides of lanthanum,
cerium, praseodymium, and gadolinium are distinct from the other LnX, in
that they exhibit metallic lustre and conductivity. The origin of these
properties is discussed in more detail in Section 4.3.

F Gil Br I
ic ~ - - MoSi,/8
PG - - _ MoS1,/8
Nd ~ PbC1,/7+2 PbCL,/7+2 SrBr,/7,8
Pm CaF,/8 ~ - -
Sm CaF,/8 PbCI,/7+2 PbClI,/7+2, Eul,/7
SrBr,/7,8
Eu - PbCI,/7+2 SrBr,/ 7,8 Eul,/7
Gd - = - MoS1,/8
Tb - - - —
Dy - SrBr,/7,8 SrI,/7 CdCl,/6
Ho ~ - - —
Er - - - -
Tm - orLJ7 SrL/7 CdL/6
Ao} CaF,/8 SrL/7 Sri, CdlL,/6
CaCl,/6
Ga - - - -

Table 4.1 Dihalides of the lanthanides. The structural type and metal coordination number
are indicated.

Other lanthanide halides

Tetravalent lanthanide halides are confined to the fluorides of Ce(I1V), Pr(1V),
and Tb(IV), of which only CeF, is thermally stable. It may be made by direct
combination of the elements, by fluorination of CeF, or CeCl,, or by the
addition of F to aqueous solutions of Ce(IV) which results in the
precipitation of CeF,.H,O. TbF, is made by the reaction of F, with TbF;.
The procedure for making PrF, is more complicated, requiring initial
fluorination of a mixture of NaF and PrF, with F, to form Na,PrF,, followed
by treatment with HF.
Many lanthanide halides have been made in which the metal is in a formal
oxidation state less than two. These include Ln,Cl, (Ln = Y, Gd, Tb, Er, Lu),
36 Solid state compounds

which consist of trans edge sharing Ln, octahedra arranged in single chains
with the Cl ions capping the faces of the octahedra. Further reduction can lead
to compounds of the general formula LnXH, (X = Cl, Br; Ln = Sc, Y, Gd,
Lu), which again feature edge sharing Ln, octahedra (now arranged in double
metal layers) but which also have interstitial hydrogen atoms.

Halides of the actinides

The actinide elements form a wide range of+ binary compounds with the
halogens, and several trends may be identified. The actinide ‘group valence’
(corresponding to the removal of all of the electrons outside the [Rn] core) is
achieved only in ThX,, PaX, (X = F, Cl, Br, I), and UX, (X =F, Cl), but not
by elements heavier than uranium. Indeed, the halides of the later actinides
resemble those of the lanthanides more than those of the early actinides, with
the +3 oxidation state being by far the most common. Thus, beyond
plutonium the highest metal oxidation state observed in an actinide halide is
+4 (in AnF,, An = Am-—Cf), while all four trihalides are known for all of the
elements uranium-einsteinium. The early actinides display a greater range of
oxidation states in their halides than the later actinides, reflecting the general
trend in actinide oxidation states discussed in Chapter Two, Section Two.
There are numerous preparative routes to actinide halides, and the method
chosen depends upon the desired product (particularly the oxidation state of the
metal). Broadly speaking, direct combination of the elements leads to halides
with the actinide in high oxidation states, while thermal decomposition or
reaction with HX produces lower oxidation states. Other routes include
reduction with H,, reaction with halogenating species such as CIF;, CCl,, or
BCI,, and halogen exchange reactions. Some of these methods are illustrated
below for uranium:

iz Fo C.c
4. > uF (6 a ee es
250-400 °C Bt 4

BCl, lp
UF; ——. UC U Uly
-107 °C 500 °C/20 kPa

HBr 900 °C
UF; ————— UF, Urn See URE,

CCl, Br,
al tee
160 °C/20 atm 500 °C

The coordination number of the actinide increases with decreasing oxidation

state (and hence increasing radius) and decreasing halogen radius. For example,
the early AnF,; (An = U-Cm) have the 9+2 coordinate LaF, structure
 The f elements 37

(Fig. 4.1), but BkF, and CfF, have the eight coordinate YF, structure. Note
that this change in structural type is also seen for LnF, but occurs three
elements further along the actinide series on account of the greater size of the
actinide ions. A similar reduction in coordination number occurs in AnBr,,
with AcBr; to NpBr, adopting the nine coordinate UCI, structure and PuBr, to
BkBr; (plus another NpBr, phase) having the eight coordinate PuBr, structure.
Actinide hexahalides are confined to UF,, UCI,, NpF,, and PuF,. All four
are strongly oxidising and highly moisture sensitive. The pentahalides are
more numerous; all four PaX,; are known as well as UX, (X = F, Cl, Br) and
NpF,. All AnX, are very moisture sensitive. All four tetrahalides are known
for thorium—neptunium, although only the flouride is known for plutonium—
californium. Eight coordination is common, often in dodecahedral geometry
(e.g. ThCl, ThBr,, and UCI,). The UX, family provides another illustration of
decreasing coordination number with increasing anion radius; UF, and UCI,
feature eight coordinate uranium, UBr, seven coordination (pentagonal
bipyramid) and UI, six coordination via edge sharing UI, octahedra. AnX, (X
= Cl, Br, I) are hygroscopic, and are readily soluble in polar solvents such as
(CH;),CO and CH;CN.
AnF, (An = U-Es) are high melting and insoluble in water, by contrast to
the other trihalides of these elements which are hygroscopic and water
soluble. Many may be crystallised out of water solutions as the eight
coordinate hexahydrates [AnX,(H,O),]*. The only trihalides of thorium and
protactinium are the triiodides; ThI, is poorly characterised, but Pal, is well
established. It may be made by heating Pal, at 360 °C in vacuo, and has the
eight coordinate PuBr; structure.
AnX, (An = Am, Cf, Es; X = Cl, Br, I) and ThI, are known. The latter has
metallic properties and has two forms, @ (black) and B (gold). By analogy
with the [Xe]4f’ configuration of Eu’, it might be anticipated that americium
would have a divalent chemistry, and the existence of black AmX, supports
this expectation.

4.2 Oxides

Oxides of the lanthanides

The most common stoichiometry of the oxides of the lanthanides is the
sesquioxide, Ln,O;, which are all well characterised. With the exception of
cerium, praseodymium, and terbium, Ln,O; is the end product of the
combustion of metallic lanthanide or of the hydroxides, carbonates, nitrates
etc. Cerium, praseodymium, and terbium form the tetravalent LnO, under
these conditions, but the dioxides may be reduced to Ln,O,; with H,. Ln,O,;
may be divided into three groups on the basis of structural type.
38 Solid state compounds

A-type: {LnO,} units arranged in an approximately capped octahedral

geometry; favoured by early lanthanides.
B-type: also {LnO,} units, but of three different types. Two are face capped
trigonal prisms and the third a distorted capped octahedron; favoured
by middle lanthanides.
C-type: related to the fluorite structure with one quarter of the O* removed to
reduce the coordination number around the metal from eight to six;
distorted octahedral coordination; favoured by middle and late
lanthanides.

All of the Ln,O, are strongly basic. They are insoluble in water but
dissolve readily in aqueous acids to produce solutions that contain
{Ln(H,0),]* (provided the pH is kept below 5). Ln,O, absorb both CO, and
H,O from the atmosphere.
LnO (Ln = Nd, Sm, Eu, Yb) may be prepared (Eqn 4.1) by reduction of
Ln,O, with elemental lanthanide at high temperature and pressure (high
pressure is not required for europium)
Ln + Ln,O; — 3LnO (4.1)
All four LnO have the NaCl structure. However, while EuO and YbO are
insulating or semiconducting, NdO and SmO exhibit metallic conductivity.
The mechanism for this process is believed to be the same as that discussed
for LnS in Section 4.3.
LnO, (Ln = Ce, Pr, Tb) adopt the fluorite structure in the limiting dioxide
stoichiometry, although a range of non-stoichiometric phases exists between
Ln,O, and LnO,. CeO, (‘ceria’) is white when pure but is usually pale yellow
on account of the non-stoichiometric phases. The non-stoichiometry of ceria
is exploited in the catalytic converters used to remove pollutants from motor
vehicle exhaust emissions. These converters contain several catalytically
active components, including platinum and other metais from groups nine and
ten, as well as ceria. Ceria has several functions in the catalytic converter:
e Promotion of the water-gas shift reaction
CO + H,O > CO, + H, (4.2)

e Enhancement of the NO, reduction capability of rhodium

e Oxygen storage
In the third role, ceria provides elemental oxygen in fuel-rich but air-poor
conditions, to ensure oxidation of unburnt hydrocarbons and the removal of
CO. It accomplishes this by going non-stoichiometric to CeO,,. In leaner
(fuel-deficient, air-rich) conditions it reoxidises to CeO,, i.e. it stores oxygen
during the air-rich periods.
Another use of CeO, is as a coating for the walls of ‘self-cleaning’
domestic ovens in which it can prevent the formation of tarry deposits.
 The f elements 39

Oxide superconductors
Prior to the mid 1980s the highest recorded value of 7, (the temperature at
which a material becomes superconducting) was ca. 23 K for Nb,Ge. Then in Superconductivity was discovered in
1911 by the Dutch physicist H.
late 1986 Bednorz and Miiller reported their discovery of a new ceramic
Kamerlingh Onnes. He cooled
compound, La, ,.Ba,CuO,, which has a T. of 30 K, and in March of the mercury to below 4.2 K and observed
following year Wu and Chu reported a T, of 92 K for YBa,Cu,O,; (0 < 8 < that it lost all resistance to the flow of
1). This latter result is especially significant as a material that is an electrical current.
superconducting at liquid N, temperatures (between 63 and 77 K) is much
more likely to find a practical use than one with very low T..

@ =Cu

O-0
= O (removed to create
O-deficient phases up
to YBa,Cu,0¢)

Fig. 4.2 The structure of YBa,Cu,O..

YBa,Cu,O, is an oxygen-deficient perovskite, the structure of which is

shown in Fig. 4.2. It may be seen that there are two types of coordination
about the copper ions - square planar and square pyramidal - and electron spin
resonance data indicate that there is a mixture of Cu** and Cu** ions
distributed across the coordination sites. Removal of oxygen from the
partially shaded sites leads to the semiconducting YBa,Cu,O,. Many other
high 7, superconductors have been synthesised, including Nd) ,Ce,CuO, and
HgBa,Ca,Cu,O, (the current high 7. record holder at 133 K, rising to 150 K
40 Solid state compounds

under 23.5 GPa of hydrostatic pressure) and all feature puckered planes of
CuO, (as does YBa,Cu,0,) which are believed to provide a path for
superconduction. The exact mechanism of superconductivity remains
uncertain, although it is widely believed that in a-superconducting material
pairs of electrons (known as Cooper pairs) move through the solid, the first
electron distorting the lattice in such a way that the second can follow it very
easily.
Unfortunately the commercial exploitation of superconductivity in, for
example, magnetic devices, power transmission, and communications, is
hampered by the ceramic nature of the high 7. compounds. The innate
brittleness of ceramics prevents their extrusion into wires, and it now seems
that the most promising way forward is the deposition of thin films of
superconductor onto metal oxide or silver surfaces to form flexible tapes.

Oxides of the actinides

All of the actinides up to and including einsteintum form more than one
oxide, with the exception of thorium which forms only ThQ,.
Polymorphism, non-stoichiometry, and intermediate phases are very
common, and hence the stoichiometries of the actinide oxides are best regarded
as limiting cases. Figure 4.3 shows the most stable oxide of each actinide.
The greater range of oxidation states exhibited by the early actinides is
reflected in their most stable oxides, and the increasingly lanthanide-like
nature of the later actinides 1s demonstrated by the increasing stability of the
sesquioxide as the series is crossed.

fed) +6
s
D
=
@ U;0;
2 +5 Pa,Os ®
o
x AmO>;

o +4 e NpO, @ @ ®@ @ 8kO,
E

+3 P @ @€£5,0,
Cm, O3 Cfo Oz

Th Pa U Np Pu Am Cm Bk Cf ES

Fig. 4.3 The most stable oxides of the actinide elements.

ThO, (thoria) may be prepared by the reaction of thorium meta! with O, at

250 °C. It has the fluorite structure and has the highest melting point of any
oxide (3390 °C). It gives off a bluish light when heated, and was used for
many years in gas mantles. Pa,O, is formed by igniting Pa(V) hydroxide in
air or by direct reaction of protactinium metal with O, at 300-500 °C. PaO,
is also known.
Many uranium oxide phases have been reported, although not all are well
characterised. UO,, the importance of which as a nuclear fuel was discussed in
 The f elements 41

Chapter Two, Section Four) has the fluorite structure (as do all of the actinide
dioxides) and this is preserved on the addition of extra oxygen until U,O, is
formed, at which point the structure is related to fluorite but with interstitial
oxygen. The next well characterised phase is U,O,, which is the end product
of heating any uranium oxide in air at temperatures above 650 °C, and which
contains pentagonal bipyramidal {UO,} units. UO,, the only anhydrous
actinide trioxide, has several crystalline forms, most of which contain uranyl]
(UO,”) groups linked by bridging oxygen. One form (y-UO,) has a structure
based on edge and corner sharing UO, octahedra. A variety of preparative
routes exists to UO;, which may be reduced by H, to UO).

400 °C
UO,.2H,O

air, 500 °C Hp
0, U0, —— U0,
300-600 °C

Oz
UO>(NO3)5.6H,O
400-600 °C

NpO, is the usual end product when neptunium compounds are burnt in air
at elevated temperatures. Reaction of Np(I[V) hydroxide with O, gives
NpO,.H,O, which when heated to 300 °C under vacuum produces the highest
oxide of neptunium, the non-stoichiometric Np,O;. Beyond neptunium the
highest oxide of each element is the dioxide, which is formed by heating of
the oxalate or hydroxide in air or O,. The sesquioxides become increasingly
stable beyond plutonium, and have the same structures as the analogous
lanthanide compounds. The C-type is the most stable form of each An,O,,
but some elements also form the A- and B-type structures (e.g. Am,O,
exhibits all three structural types).

4.3 The electronic structure of lanthanide metals and

divalent lanthanide compounds
As has already been noted, many divalent lanthanide compounds exhibit very
high electrical conductivities while others are insulators or semiconductors.
This Section provides an explanation for this behaviour in terms of the band
theory of electrical conductivity and the Hubbard approach to electron-electron
repulsion in solids. These models are briefly described, and subsequently
applied to the elemental lanthanides and divalent lanthanide compounds.
Band theory is one of the most successful approaches to the electronic
properties of solids. It assumes that the energetic separation between the
electronic levels that arise from overlap of orbitals on adjacent atoms is so
small that the levels merge into a continuous energy band. If a band is only
partially filled, then electrons with energies close to that of the least stable
42 Solid state compounds

electron (known as the Fermi level) can be easily promoted into low-lying
empty energy levels. As a result they are relatively free to move through the
solid, giving rise to high electrical conductivity. By contrast, electrons in a
solid in which all the bands are full cannot undergo this process, and the solid
will be an electrical insulator.

The Hubbard model

Many compounds of the lanthanides that might be expected to be metallic (on
the grounds that they appear to have partially filled d or f bands) are in fact
insulating. This is due to the effects of electron-electron repulsion, which
localises electrons on individual atoms and prevents the high electrical
conductivity described above. This situation arises when the overlap of
orbitals on neighbouring atoms is small, and hence the width of the resulting
energy band is also small. Full treatment of electron-electron repulsion in
solids is extremely complicated, but we can make progress via the Hubbard
approach, which makes the assumption that the only important
electron-electron repulsion occurs between electrons on the same atom.
Although this would appear to be a fairly drastic approximation, intra-atomic
electron-electron repulsion does seem to be the principal cause of failures of
band theory, and the Hubbard model is a surprisingly useful approach to
electron localisation.
Consider a model solid consisting of a line of atoms each with a single
valence s orbital and one s electron per atom. The s band will be half full and
the solid will be metallic. If, however, the overlap of the s orbitals on
adjacent atoms is very small, it will be energetically preferable for the solid to
localise its valence electrons as shown in Fig. 4.4, in order to avoid the
electron-electron repulsion that arises when two electrons are forced to pair on
the same atom, as must occur were an electron to move through the solid via
the s band. The solid will therefore be insulating.

YDODOOOD OOO
Fig. 4.4 Ina line of half-filled, weakly interacting s atomic orbitals, electron-electron repulsion localises the electrons one per atom
(above), in order to avoid having to pair two electrons in the same atom (below).

DOODOBOOO
We may quantify these ideas as follows. The energy required to remove an
electron from an atom is the ionization energy, J. If we now place the electron
in an atom which already has a half-filled s orbital we get back the electron
affinity of that atom, A,. The energy required to move the electron is therefore
given by Eqn 4.3, where U is called the Mott-Hubbard splitting, or Hubbard
U, and may be interpreted as the repulsion energy between two electrons in
the same atom.
 The f elements 43

U=I-A (4.3)
In order for metallic conductivity to occur, the overlap of the s orbitals must
be sufficiently large that the band width, W, is greater than the value of U. If
U > W, electron-electron repulsion prevails and the solid is insulating.

Elemental lanthanides
We have previously seen that the 4f orbitals of lanthanide atoms and ions are
radially contracted and interact to a small extent with the surrounding ligands
in lanthanide compounds. A further consequence of this contraction is that the
overlap of the 4f atomic orbitals on neighbouring atoms in the elemental
lanthanides is very small, and hence 4f band widths are also small (typically <
0.1 eV). The metallic nature of the elemental lanthanides comes from the 5d
and 6s atomic orbitals, which are much more diffuse and overlap to a much
greater extent. That the 4f orbitals cannot be responsible for the metallic
properties of the lanthanide elements is confirmed by Fig. 4.5, which plots
the value of the Hubbard U for the lanthanide 4f orbitals. These data have
been experimentally determined using solid state X-ray photoelectron
spectroscopy, and clearly show that U is very much greater than 0.1 eV for
the 4f orbitals of all of the lanthanides.

Ce Pk Nd Pm sm Eu Gd Tb Dy Ho Er Tm

Fig. 4.5 Hubbard U values for the 4f atomic orbitals of the lanthanides.

The U values for europium and gadolinium are particularly striking, as

they are so much greater than any of the others. It is useful to consider the
elemental lanthanides as being composed of Ln* ions, with the [Xe]4f
configuration, embedded in a delocalised valence electron ‘sea’ made up of
electrons in the 5d and 6s bands. Each lanthanide atom contributes three
electrons to the band structure. In gadolinium, this results in the [Xe]4f’
configuration which has the particular stability associated with the half-filled f
44 Solid state compounds

subshell. This is lost when an electron is added to make the [Xe]4f*

configuration, and hence U is very high. The [Xe]4f’ configuration is so
favourable that it is also adopted in metallic europium, which is best regarded
as containing Eu’* (as mentioned in Chapter Two, Section Three). Although
this gives a high value of U for the europium 4f orbitals, it also means that
only two electrons per atom enter the band structure and are available for
metallic bonding. This manifests itself in a lower sublimation energy and a
larger atomic volume for europium than the other lanthanide elements except
for ytterbium, for which the stability of the [Xe]4f'* configuration also results
in only two metallic bonding electrons per atom.
It is also worth noting that the lanthanide 4f U values predicted from
gas-phase ion data are in the region of 25 eV. This is clearly significantly
greater than the values in Fig. 4.5, and is an example of a general effect in
which the U values measured for solids are much less than expected on the
basis of atomic/ionic energies. This is due primarily to polarisation effects, in
which the electrons surrounding a hole created by the removal of an electron
relax in toward the hole and away from the removed electron, thereby
lowering the energy required to move an electron from one atom to another.

Divalent lanthanide compounds

Although the +3 oxidation state 1s the most common for the lanthanides,
several series of divalent lanthanide compounds exist. These include Lnl,,
LnE (E = S, Se, Te), and LnB,, the latter containing octahedral B,’ ions
linked in a continuous three-dimensional framework.
The electronic configuration of most gas-phase Ln** is [Xe]4f"*', and hence
it mightbe expected that the divalent solids would be electrically insulating
on the grounds that the only electrons outside the [Xe] core are in the 4f
subshell and are thus highly unlikely to produce bands of a width great
enough to overcome the electron-electron repulsion. However, many of the
divalent lanthanide compounds listed above are metallic, and it appears that
the [Xe]4f"5d' configuration is more stable than the [Xe]4f"*! in some
lanthanide compounds, giving rise to a single electron per lanthanide atom in
a broad 5d band. The partially filled 5d band is believed to be responsible for
the high conductivities of Lal,, Cel,, PrI,, Gdl,, NdO, and SmO noted in
Sections 4.1 and 4.2, as well as that of most LnB,. However, for certain Ln’*
the [Xe]4f"*' configuration remains more stable than the [Xe]4f"5d! even in
compounds. Thus the monotellurides of europium, samarium, thulium, and
ytterbium are non-metallic, as are EuS, SmS, and YbS, the diiodides of
neodymium, samarium, europium, dysprosium, and ytterbium, and LnB, (Ln
= Fu, Yb).
 The f elements 45

4.4 Exercises
hie LnO (Ln = Nd, Sm, Eu, Yb) have the NaCl structure, while LnO, (Ln
= Ce, Pr, Tb) and all of the actinide dioxides have the CaF, structure.
Both of these structural types may be described by the filling of holes
in face centred cubic (fec) lattices. Which type of hole is filled in each
case, and which ions form the fee lattice? What is the coordination
number of the anions and cations in each structural type, and what is
the relationship between the ratio of the anion:cation coordination
number and the stoichiometry of the oxide?

to From the values of AH,° for LnCl,(s) given below, estimate values for
EuCl,(s) and YbCI,(s). Compare your estimates to the experimental
values of -936 and -960 kJ mol’ respectively. Rationalise the
differences between the estimated and actual values. Why is there a trend
toward decreasing AH,° for LnCl,(s) with increasing Ln atomic
number?

Compound | AH&/kI mol!

SmClL,(s) -1026
GdCl,(s) -1008
TmCl,(s) -991
LuCl,(s) -986

What is the Meissner effect?

Discuss the following elemental melting points (K)

1002 La 1194 Nd 1294 Eu 1095 Yo. 1097

W 3680
O73 Ac 1320 U_ 1406
5 Coordination chemistry

The coordination chemistry of the lanthanides and actinides has as much in

common with that of groups one and two as it does with that of the d block.
At the same time, the unique characteristics of f block complexes have
required the development of new techniques for investigation of their
chemistry, and have subsequently led to a growing number of technological
and scientific applications in, for example, medicine and catalysis. Of
growing concern to the public at large is what happens to the products of the
nuclear power and fuel reprocessing industries after their release into the
environment. Coordination chemistry is at the heart of unravelling this issue.

Fig. 5.1 Molecular structure of [Nd(NO,).(H,O),]*. The f element ions are large, hard acids
and form high coordination number complexes with hard bases.

5.1 Keys to understanding

The following summarises the properties of the lanthanides and actinides
which are of direct relevance to understanding their coordination chemistry.
(a) As hard Lewis acids, the lanthanide and actinide ions prefer to coordinate
hard bases such as F and H,O (Fig. 5.1). The ‘classical’ coordination
chemistry in an aqueous environment is thus very different from that in
say anhydrous hydrocarbons, and this has led to the development of
what is sometimes termed the neo-classical coordination chemistry of
these elements and the advent of a unique organometallic chemistry
(Chapter Six).
(b) To an incoming ligand, lanthanide ions have the appearance of a noble
gas atom, except with a positive charge (most commonly +3). This is
because the 4f orbitals which contain the valence electrons do not
extend out far enough to interact to any great degree with ligand
 The f elements 47

orbitals. The complexes thus formed are held together largely by

electrostatic interactions (ionic bonding). The Sf orbitals of the actinides
are rather more accessible, however, leading to some overlap and
covalent character, but this property decreases with increasing atomic
number, and the later actinide ions behave like the lanthanides.
(c) The f elements are large, which means that any charge on the ion is
distributed over a large area. However, since the most common
oxidation states of these elements are +3 and above, they still have a
high charge density. As we cross the 4f and Sf series, ionic radii
decrease uniformly (lanthanide and actinide contractions, see Chapter
Two, Section Three) leading to higher charge densities and stronger
ionic bonds.

5.2 Complexes with water

Most of the early coordination chemistry of the lanthanides and actinides was
performed in water or in protic solvents. The highly charged ions formed by
the f elements are well suited to this highly polar environment.

Comparative features of lanthanides and actinides

Few monodentate ligands can compete with water for coordination sites of
lanthanide ions in aqueous solution. By contrast, the actinides are far better at
forming complexes with other bases under these conditions. For example,
while dissolution of LnCl, in water leads to the formation of ({Ln(H,O),]*
(x= 8, 9) ions, ThCl, dissolves to give hydrated [ThCl,]**. It is also
generally the case that isolation of lanthanide complexes from aqueous
solution is much more difficult than it is for the actinides. In all cases, the
use of chelate or macrocyclic ligands leads to more predictable outcomes, and
it is here where much modern research ts focused.
The greatest difference between the aqueous chemistry of the lanthanides
and that of the early actinides arises from the tendency of some of the latter
(U-Am) to form pentavalent and hexavalent actinyl ions AnO,* and AnO,”*
(discussed in Section 5.3). The aqueous chemistry of thorium and
protactinium resembles more closely that of the group four and five transition
metals respectively, while the later actinides (Cm—Lr) behave rather like the
lanthanides.

Oxidation states
As we have seen in Chapter Two, Section Two, the early actinides have many
more stable oxidation states than do the lanthanides.

Lanthanides
The divalent lanthanides tend to reduce water to hydrogen (Eqn 5.1), although
Eu* is more stable in aqueous solution than are Sm** and Yb**. Also, since
all these ions are oxidised by molecular oxygen, the solutions must be
handled under an inert atmosphere. The tetravalent ions of neodymium,
dysprosium, praseodymium, terbium, and cerium oxidise water to oxygen
(Eqn 5.2), and of these only Ce** is sufficiently kinetically stable to form
48 Coordination chemistry

aqueous coordination compounds. Hence, the trivalent ions Ln* are by far the
most readily studied.
2Ln**(aq) + 2H*(aq) = 2Ln**(aq) + H,(g) (520)

4Ln*(aq) + 2H,O(1) = 4Ln**(aq) + 4H*(aq) + O,(g) (5.2)

Actinides
The oxidation states adopted by the actinides and the-relative stabilities of the
di-, tri-, and tetravalent states are represented in Figs. 2.3 and 2.4. In this
section we will illustrate the general trend of increasing stability of the lower
oxidation states across the series. It is worth noting that complexes of the
actinide ions in the same oxidation state usually have the same structures, for
example, the hexavalent ions all exist as AnO,”* in aqueous solution.
We have noted that the stability of the trivalent state for the actinides
increases as we move across the series. Th** is rapidly oxidised by water, and
the aqueous chemistry of this ion is exclusively of the tetravalent state. The
U* ion is, however, readily obtained by reduction of higher valent species and
is stable for days in the absence of air. Eventually it reduces water to
hydrogen, and the reaction in Eqn 5.3 is accelerated by the presence of acid.
The tetravalent ion U* is quite stable in the absence of oxygen.
U*(aq) + H,O(l) > U*(aq) + 1/2H,(g) + OH’ (aq) (5.3)
Moving further along the 5f series, aqueous solutions of Np** are stable in
air, and this ion is readily prepared by electrolytic reduction of acidic solutions
of Np**.. By the time we reach americium, the trivalent ion is the preferred
state. All the actinide elements after protactinium can exist as An** in aqueous
solution.
A corresponding trend is observed for the penta- and hexavalent actinyl
ions. Their stability falls across the series from uranium, for which UO,”*
predominates in the presence of air, to curium for which the actinyls are
unknown. The physical properties of these ions are dealt with separately in
Section 5.3.
The redox chemistry of plutonium in aqueous solution is fascinating in its
complexity. The element is capable of attaining five oxidation states in water,
from Pu(II) to Pu(VIJ), although the latter only exists in strong alkali.

py 0-986 ie NEN puo,t O818Y Bg 2

ee ————— a
(IIL) (IV) (V) (V1)

The redox potentials (see above) separating the most common states are
roughly equal. Interconversion between oxidation states occurs through
disproportionation and conproportionation reactions, and the matter is further
complicated by the reducing effect of the a decay of the plutonium atoms.
Given these factors, it is not surprising that all four oxidation states Pu(IID) to
Pu(VI) commonly occur together in aqueous solution, and it is difficult to
prepare pure solutions of any one ion. The redox couples Pu(IV/III) and
Pu(VI/V) are reversible, and rapid reactions occur with one-electron oxidising
 The f elements 49

and reducing agents. Those couples which would require the making or
breaking of bonds, e.g. the oxidation of Pu** to PuO,", are irreversible.

Coordination numbers and geometries

As a consequence of the large size of the lanthanide ions, high coordination
numbers (up to 12) are found. In aqueous solution, the Ln** ions are thought
to be surrounded by eight or nine oxygen-bound water molecules for the later
and earlier metals respectively, but owing to the extremely rapid exchange of
bound and unbound water these coordination numbers are not known with
certainty. By contrast, the much smaller group three element scandium exists
as [Sc(H,O),]* in water. The situation is more complicated for the actinides
because of their redox chemistry, but, for example, aqueous U** is probably
nine-coordinate.
The lack of covalent (directional) bonding means that geometries of
lanthanide complexes are known with even less precision than are the
coordination numbers in solution. However, many hydrated lanthanide ions
have been examined by X-ray crystallography in the solid state. They
generally have tricapped trigonal prismatic structures (Fig. 5.2). In complexes Fig. 5.2 Tricapped trigonal prismatic
where geometrical or optical isomers might be expected, these are extremely structure adopted by [Ln(H,O),]** in
difficult to detect or isolate because of rapid equilibration. the solid state. Compare with the
tysonite structure in Fig. 4.1.
Hydrolysis
When a metal cation dissolves in water, it will undergo hydrolysis to some
extent (Eqn 5.4). Highly charged cations such as those formed by the f
elements will polarise the O—H bonds in water strongly, and the aquo cations
[M(H,0),]"* tend to act as Brgnsted acids.

[M(H,0),}"*= [M(H,0),.,(OH)]""" + H* (5.4)

The acidity of aqueous solutions of the group three element trications
decreases rapidly as we move down the group, and that of the trivalent
lanthanides increases smoothly across the series as might be expected from
the reduction in ionic radius (lanthanide contraction - Fig. 2.5) and
concomitant increase in charge density of the ions. In the actinide series, the
acidity of the aqueous cationic species decreases in the order
(most acidic) An** > AnO,’* > An** > AnO,* (least acidic)
The +4 cations are the most acidic since they have the highest charge density.
For the actinyl ions AnO,"" (Section 5.3) the question arises as to whether the
extent of hydrolysis depends on the net charge on the ion (i.e. +1 or +2) or on
the formal charge on the metal at the centre (i.e. +5 or +6). In the case of
AnO,”* the O* ions do not fully ‘quench’ the charge on the metal, and the
effective charge as experienced by an approaching water molecule has been
calculated to be ca. +3.3. This explains the position of the AnO,”* ions
between An* and An* in the above order of decreasing acidity (decreasing
hydrolysis).
50 Coordination chemistry

Rates and mechanisms of exchange of ligands

The characteristic lifetimes for exchange of water ligands (Eqn 5.5) in aquo
complexes of the trivalent transition metals range from ca. 10° s for ions with
large ligand field stabilisation energies such as Cr**, to 10° s for those which
do not, such as Fe**. For the f elements, these lifetimes are of the order 10° s
which means essentially that rate of exchange of water ligands is controlled
Complexes which survive for long by the rate of diffusion of the molecules in and out of the inner coordination
periods (usually taken as about one
sphere. All f element complexes with monodentate ligands are regarded as
minute) are termed inert. Those
which do not are called labile. being very labile. This makes the study of substitution reaction mechanisms
difficult.
[M(H,0),]"* + H,O* = [M(H,0), ,(H,O*)]"* + H,O (5.5)
Let us assume that the intimate mechanism of substitution (Eqn 5.5) is
The intimate mechanism describes
dissociative, i.e. that in the rate determining step a bond to an inner sphere
the formation of the transition state water ligand begins to break before the new bond to an incoming ligand is
in the rate determining step. formed. Since small, more charge-dense ions should form stronger bonds with
water we would expect the rates of exchange to fall with size of the ion. This
is found to be the case by experiment. For an associative rate determining
step, where a bond with an incoming ligand begins to form before other
bonds are broken, we expect the same behaviour since the smaller ion is less
susceptible to nucleophilic attack by an incoming group on steric grounds.
The rates of substitution of water by other ligands, although also very fast,
have been measured in a few cases. The variation in rate constant for the
formation of 1:1 Ln**/oxalate complexes (Eqn 5.6) is shown in Fig. 5.3. If
oxalate is replaced with a similar bidentate ligand, the same sort of graph is
obtained, and the rate of substitution is independent of the incoming group.
Ln* (aq) + ‘O,C—CO, (aq) > [Ln(O,C—CO,)]* (aq) (5.6)

kx10’
dm*
mol
s"

La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Ef Tm

Fig. 5.3 Rate constants for the formation of 1:1 Ln*/oxalate complexes at 25 °C.
 The f elements 51

The rates are essentially constant at ca. 8 x 10’ mol dm® s"' in the region
La* to Eu* but then fall rapidly via Gd** and Tb** to another region Dy** to
Tm* where they are again constant at ca. | x 10’ mol dm® s’'. It is worth
noting that the ion midway between the two regions of constancy is Gd**, and
it is widely held that it is at this point along the lanthanide series that the
coordination number in aquo complexes changes from nine (early lanthanides)
to eight (late lanthanides). This reduction in the number of water molecules in
the primary coordination sphere would lead to a sudden increase in the
strength of the remaining Ln—OH, bonds. Hence, if the breaking of this bond
is rate determining, there will be a dramatic reduction in the rate of
substitution.
The stoichiometric mechanism in Fig. 5.4, consistent with these data, has The stoichiometric mechanism
describes the sequence of elementary
been proposed. It has the following steps.
steps by which the reaction takes
(i) The incoming ligand becomes associated with the aquo complex, place.
probably through hydrogen bonding with the highly. polarised water
ligands.
(ii) One of the water ligands begins to dissociate, perhaps under the
influence of the incoming group.
(1) The loss of water ligand (breaking the strong Ln—O bond) is rate
determining.
Since oxalate is a bidentate ligand, it might be expected to replace two
water ligands, and not just one as shown for simplicity in Fig. 5.4. A second
substitution reaction converting an 7'-oxalate to n° will follow step (iii).

(i)
[Ln(H»O)9]**(aq) + (oxalate)* (aq) ee ([Ln(H2O)g--7*- (oxalate)]*(aq)

(iii)
[Ln(H2O)g, (oxalate)]*(aq) + HO | (LO) (oxalate): (H2O)}*(aq)

Fig. 5.4 A stoichiometric mechanism for substitution of H,O ligands for a simple bidentate
ligand such as oxalate (C,0,*).

5.3 Actinyl ions

Most complexes of the actinide elements in oxidation states higher than +4
(i.e. those containing uranium, neptunium, plutonium, or americium) contain
the actinyl ions AnO,"* (n = 1, 2). The physicochemical properties of these
ions have been studied in great detail, not least because of their technological
and environmental importance in nuclear fuel reprocessing and in waste
management.

Structure of uranyl ions and complexes

The actinyl ions contain essentially linear O-An—O units. By contrast, the
transition metal dioxo complexes are almost invariably bent. Compare, for
52 Coordination chemistry

example, the structures of [MO,(Ph,PO),Cl,] (M = U, Mo) in Fig. 5.5. In the

uranium compound the oxo ligands adopt a trans configuration while in the
molybdenum compound the structure is cis.

Fig. 5.5 Molecular structures of (a) [UO,(Ph,PO),Cl,] and (b) [{MoO,(Ph,PO),Cl,].

Perhaps the most closely related species to the actinyl ions in the d block
is the d? osmyl(VI) ion OsO,** which is also linear with Os—O double bonds
of ca. 1.75 A. Since the ionic radius of U(VI) is nearly 0.2 A greater than that
of Os(VJ) it is quite surprising that in the uranyl ion UO,"* the U—O distances
are as short as 1.7 to 1.8 A. This is taken as good evidence for the presence of
a significant degree of multiple (covalent) bonding in the actinyl ions (see
later).
A further key feature of the structure of coordination complexes of the
actinyl ions is that the auxiliary ligands are almost always accommodated in
the equatorial plane. Usually between four and six ligands are supported in
this way giving (a) octahedral, (b) pentagonal bipyramidal, and (c) hexagonal
bipyramidal structures (Fig. 5.6).

2 O 3 O
20
os H Oo en a ae -O== N
=Cl. ro

ca e FSaes |ee
ee i2 | es
-=

Ay gl eee een
Ci-* ~ ios
& = SET
yeas Wis. 70?

O O
| i O
|
(a) (b) (c)

Fig. 5.6 Coordination in the equatorial plane of uranyl (VI) ions. (a) The octahedral dianion of the salt Cs,[UO,Cl,], (b) pentagonal
bipyramidal trianion in K,[UO,F.}, and (c) hexagonal bipyramidal uranyl nitrate dihydrate [UO,(NO,),(H,O),).
 The f elements 53

The latter class of structure is only formed when three bidentate ligands such
as carbonate, nitrate, or sulphate are present. In the cases with six equatorial
monodentate ligands, the equatorial plane is puckered.
Actinides in the environment, whether they be naturally-occurring or
arising from accidental contamination, are usually found in the form of
actinyl ions, except of course in the case of the later actinides where the
trivalent ions are more stable. A highly topical area of research is in the
coordination chemistry of actinyls with ligands such as carbonate and
bicarbonate which are found in relatively high concentrations in natural
groundwaters. Such systems can be quite complicated, and several different
species may exist in rapid equilibrium with one another. However, using a
wide range of techniques, it has been shown that the dominant species in the
uranyl(VI) carbonate system at pH = 6 is the trimer [(UO,),(CO,),]°, Fig.
ay i

Fig. 5.7 Molecular structure of [(UO,),(CO.),]” determined by X-ray crystallography.

Bonding in actinyl ions

If we first consider the two oxygen atoms, there is a total of six possible
linear combinations of their six 2p orbitals (0,, O,, 2 X M,, 2 X 7,). All of
these have symmetry-allowed bonding combinations with either 5f or 6d
orbitals on the actinide. Two such combinations (Fig. 5.8) are (a) O(p,) -
An(dy,) which has m, symmetry and (b) O(p,) - An(f,,2) which has 1, The f,,2 orbital is member of the
symmetry. It should be noted that the latter interaction is not possible using d general set of f orbitals, see Chapter
orbitals on the metal. We will come across more examples of the Two, Section One.

complementarity of f and d orbital symmetries in Chapter Six.

54 Coordination chemistry

O----C-----0

Fig. 5.8 Examples of possible An—O bonding combinations in actinyl ions. (a) 7,: out of
phase oxygen p/An d,, and (b) 7: in phase oxygen p/An f,,2.

The principal features of the ordering of the six bonding molecular orbitals
shown in Fig. 5.9 have been established for the uranyl(VI) ion using a variety
of spectroscopic and theoretical techniques. It can be seen that the = bonding
molecular orbitals are lower in energy than the o bonding molecular orbitals,
implying that o bonding is less important than 7 bonding. This is quite
unusual since we normally expect 6 overlap to be more efficient than 1
overlap where p orbitals are involved. One explanation for this may be that at
short U—O distances there is antibonding overlap between the o-oriented
O----C-----0
oxygen 2p, orbitals and the toroidal lobes of the uranium d and f orbitals.
Fig. 5.10 The o,, bonding interaction This is illustrated in Fig. 5.10.
in the uranyl(VI) ion between uranium
f,3 and oxygen p showing the Ps ae anti-bonding (o and m)
possibility of destructive overlap.

O(Gd) P= hs <s8 ene {—_] 6d non-bonding (8)

oe mea | anti-bonding (o and 7)

U(5f) [k= ------- {____] 5f non-bonding (5 and 9)

eee O(2p) combinations

Oye ae Sg (6 in total)
6 bonding orbitals } %o ie
(filled in UOQ**)

Fig. 5.9 A schematic molecular orbital energy level diagram for the uranyl(VI) ion UO,**.
 The f elements 55

Of course, not all the f and d orbitals on the metal are required, and some
remain non-bonding, for example the 5f based molecular orbitals which are of
5 and @ symmetry. In the case of UO,**, we have zero electrons from U(VI)
and six p electrons from each O* making a total of 12.' This completely fills
the bonding molecular orbitals shown in Fig. 5.9 and accounts for the Oo 2+

unusual stability of the uranyl ion.

As we have seen, both d and f orbitals play a role in bonding between
uranium and oxygen, and all 12 valence electrons participate. Hence each U-O
bond is of order three (i.e. a triple bond, Fig. 5.11). Some further interesting O
points arise in relation to the molecular orbital treatment above.
(a) MO,2* (M = Np, Pu, Am), which have 13, 14, and 15 electrons Fig. 5.11 Valence bond
respectively in this model, will have one, two, and three electrons in representation of the U—O triple bonds
in the uranyl(VI) ion.
the non-bonding Sf based 5 and @ orbitals. These ions are also linear
and quite robust, but the stability with respect to lower oxidation state
species decreases in the following order.
(most stable) U > Np > Pu > Am (least stable)

This is a good example of the increasingly lanthanide-like behaviour of

the later actinides.
(b) The sensitivity of f orbital energies and thus An—O overlap to overall
charge on the ion causes the uranyl(V) ion UO,* to be unexpectedly
unstable. It disproportionates readily (Eqn 5.7).
2U0,* + 4H* = U* + UO,” + 2H,O (3.7)
(c) On the basis of steric effects alone we might expect the linear O-M—O
structure to be quite commonplace for ions MO,"*, but this is not the
case. As pointed out earlier, some of the six linear combinations of
oxygen p orbitals for the linear uranyl(VI) ion have symmetry matches
with metal f orbitals only. The bent structure of MoO,”* [Fig. 5.5(b)]
thus probably arises from the inability of transition metals to utilise f
orbitals. In order to be able to accommodate all six valence d electrons
in bonding molecular orbitals, the ions must adopt a distorted
geometry.
(d) The molecule ThO, has been isolated in a cold matrix and detected in
the gas phase. Despite being isoelectronic with the linear UO,” ion, it
has a bent structure. This is probably because the thorium 5f orbitals
are much higher in energy than they are in uranium, thus reducing the
magnitude of the p/f interaction (Fig. 5.9). Thorium, like the transition
metal ions, thus has to rely more on p/d overlap.

Actinyl ions in nuclear fuel reprocessing

The process of separation and recycling of the components of irradiated fuel
elements is outlined in Fig. 2.9. The key stage of separation of uranium and
plutonium is dependent upon the unusual stability of the uranyl(VI) ion to

t+ A ‘neutral ligand’ counting approach here would be as follows: six electrons

from uranium, four electrons from each oxygen, less two for dipositive charge
= 12 electrons.
56 Coordination chemistry

reducing agents. Treatment of a mixture of [UO,(NO;),(TBP),] and

[Pu(NO,),(TBP),] in kerosene with Fe(II) leads to reduction of Pu(IV) to
Pu(IID which, like the trivalent lanthanide and transition metal ions, does not
form a kerosene-soluble TBP adduct and instead migrates to the aqueous
phase. The extremely stable uranyl nitrate is unaffected and stays in the
organic phase.

5.4 Chelate and macrocyclic complexes

The extremely high rates of exchange of ligands at f element centres makes
the isolation of their coordination complexes difficult, particularly from
aqueous solutions. This kinetic instability was discussed in Section 5.2. A
successful method of overcoming this is to increase the thermodynamic
stability of the complexes by exploitation of the chelate and macrocyclic
effects.

Formation constants and the chelate effect

If we consider the formation constants K for Eqns 5.8—5.10 we can see that as
the denticity of the ligand (its number of ‘teeth’) increases, the formation
constant increases dramatically. Moving from the monodentate chloride ligand
to the chelate (or ‘claw’) ligand acetylacetonate (ACAC), K increases by five
orders of magnitude. With the ethylenediaminetetraacetic acid (EDTA*) ion,
six water molecules are eliminated on complexation and the resultant large
increase in entropy leads to a more thermodynamically stable complex.

[Nd(H,0),}** 16 (@Q) nes [Nd(H,0),Cl]* + H,O (5.8)

2 molecules 2 molecules ell

[Nd(H,0),]** + ACAC = [Nd(H,0),(ACAC)}** + 2H,0 (5.9)

2 molecules 3 molecules K=10°

[Nd(H,O),]** + EDTA* = [Nd(H,O),(EDTA)] + 6H,O 10)

2 molecules 7 molecules K = 10"

The chelate effect may also be explained in the following way. When one end
of a bidentate ligand attaches itself to a metal centre, the effective
Some of the dramatic increase in K for
Eqn 5.10 can be attributed to charge concentration of the other end of the ligand is artificially high, hence affecting
neutralisation of Nd** by EDTA* the equilibrium position. For macrocyclic ligands such as
which would loosen the solvation tetraazacyclododecanetetracetic acid (DOTA*) this phenomenon is amplified
sphere around the ions and thus since the ligand is unlikely to be able to distort in such a way so as to
increase the entropy of the system.
remove a ligating atom from the coordination sphere.

Trends in stability of chelate complexes

The reaction (Eqn 5.11) of EDTA* with lanthanide metal trications in
aqueous solution has an overall equilibrium constant K which varies across
the lanthanide series as shown in Fig. 5.12.
 The f elements 57

Ln**(aq) + (EDTA)*(aq) = [Ln(EDTA)] (aq) (nl)

This trend is characteristic of a number of similar polydentate ligands. There

is a general increase in stability with increasing atomic number or decreasing
ionic radius. This is readily explained in terms of the increasing Coulombic
attraction of the ligand to the metal as the charge density of the latter
increases. Note the slight irregularity in the graph at Gd**. This is a
characteristic of nearly every ligand that has been studied in this way and is
referred to as the ‘gadolinium break’. This has been explained in terms of the
ligand field stabilisation of the [Xe]4f’ configuration of the Gd* ion, but as
we have discussed, these effects are small for the f elements and in particular
the lanthanides. It seems more likely, as we have discussed in Section 5.2,
that Gd** represents the border point between two stable coordination numbers
in aqueous solution.

La Ce Pr Nd Pm sm Eu Gd Tb Dy

Fig. 5.12 Equilibrium constants K for the formation of 1:1 Ln*/EDTA* complexes in aqueous solution at 25 °C.

Ion exchange chromatography

The steady variation in chelate complex stability across the series is exploited
in ion exchange chromatography which is used in the industrial scale
separation of the lanthanides and in the purification and identification of
transuranium elements. In this technique, metal cations are partitioned
between a stationary solid phase and a mobile aqueous phase. The stationary
phase is most usually a resin with binding sites for cations, for example
sodium polystyrene sulphonate. The f element ions have a high affinity for
the sulphonate groups, and when they are introduced as an aqueous solution to
the top of the column they readily exchange with the sodium ions. The
mobile aqueous phase contains a chelate anionic ligand such as citrate or
2-hydroxyisobutyrate. An equilibrium is established between the complexed f
58 Coordination chemistry

element ions in the aqueous phase and those attached to the resin [Eqn 5.12,
(r) denotes a resin-bound species].
Ln**(r) + 3citrate(aq) + 3Na*(aq) = [Ln(citrate),](aq) + 3Na*(r) (5.12)

The order in which the ions are eluted (Fig. 5.13) reflects the balance
between the affinity of the ions for the stationary phase and the stability
constant K of the chelate complex. The ions with the most stable chelate
complexes (i.e. the smallest ions) are eluted first.

————

eluant
in
concentration

concentration
eluant
in
——3>

total volume of eluant ———s

Fig. 5.13 The elution of lanthanide (upper) and actinide (lower) ions from an ion exchange
column. The heavier elements appear first.

Ion exchange 1s rapid and selective, and since the positions of elements in
the elution series can be predicted with some accuracy, the technique was of
key importance to the discovery of new (transcurium) elements. The first
synthesis and identification of berkelium, element 97, is described thus by
Seaborg.
...we were somewhat surprised to see the rather large gap between
97 and curium; we shouldn’t have been surprised because there is a
notably large gap between the elution peaks of the homologous
lanthanide elements terbium and gadolinium.
 The f elements 59

Lanthanide complexes in NMR spectroscopy

Lanthanide shift reagents

Since most complexes of trivalent lanthanides are paramagnetic, they do not
usually give very useful NMR _ spectra. The resonances are at unusual
chemical shifts and are often very broad. Nevertheless, certain complexes
called lanthanide shift reagents are used to induce chemical shift changes
selectively in other molecules via through-space interactions. These
The ions Eu** and Pr** are often used
complexes comprise a lanthanide ion surrounded by three B-diketonate ligands because their electronic relaxation
such as in 5.1. Ligands are chosen such that the complexes are soluble in times are very short. This reduces
non-polar solvents but still have reasonably accessible coordination sites so line-broadening for the nucleus under
that the molecule being analysed can form a bond via a heteroatom to the study, most commonly 'H.
paramagnetic centre.

5.1 5.2

The first use of lanthanide shift reagents was in the 'H NMR analysis of
molecules which under normal conditions display overlapping peaks in the
alkyl region of the spectrum, and which contain a ligating group (e.g.
aliphatic alcohols and amines). Under the influence of the unpaired electrons,
the resonances are spread out over a much greater chemical shift range, thus
simplifying the spectrum. Also, since the groups closest to the paramagnetic
centre are most strongly affected, it is possible to use the lanthanide induced
shift to estimate structural data. With the advent of high field NMR
spectrometers, these reagents are used less often. However, chiral lanthanide
shift reagents such as 5.2 are widely applied to the NMR analysis of
mixtures of enantiomers. Since one enantiomer is bound more strongly than
the other, it experiences on average a larger contact shift of its resonances.
Hence, signals for the two enantiomers are separated. Integration of the
spectrum obtained allows the determination of enantiomeric excess.

Contrast agents for magnetic resonance imaging

Whole body magnetic resonance imaging (MRI) scanners detect the 'H NMR By contrast to Eu** and Pr*
mentioned above, Gd** has a rather
signals of water in the tissues and fluids of the patient. The presence of
long electronic relaxation time,
highly paramagnetic, stable chelate complexes of gadolinium such as making it an efficient relaxer of 'H
[Gd(DTPA)(H,O)] reduces the relaxation times of nearby 'H nuclei, and thus spins at levels where any lanthanide
causes enhancement of the signal. induced shift is negligible.
60 Coordination chemistry

The complexes are distributed in extracellular fluids, but are preferentially

absorbed in brain tumours since the blood/brain barrier is ruptured in such
cases allowing access of the complex. Hence, the contrast in the magnetic
resonance image between tumour and healthy tissue is increased. Doses are
approximately 0.1 mmol kg! (ca. 7 g for an adult human), and since
gadolinium ions and also the free ligands are toxic at this level the complex
must be stable enough to be excreted unchanged by the body. Hence the use
of the strongly chelating DTPA®* ligand. Several gadolinium complexes are
used in clinical practice today and 40% of all MRI scans are taken with the aid
of these contrast agents.

5.5 Alkoxides and dialkylamides

Coordination chemistry does not have to take place in water, and indeed given
the high affinity of the f elements for this ligand it is clear that their
chemistry will be dramatically modified by the use of anhydrous media and
techniques for the exclusion of air. In recent years, a new coordination
chemistry of the lanthanides and actinides has been developed using this
approach. This Section describes the key issues and outcomes of these
studies. These neo-classical coordination complexes have structures and
properties more akin to the relatively inert organometallic compounds
described in Chapter Six than they do the labile hydrated complexes described
in Section 5.2.

Bonding
In principle, the alkoxide ligand RO» can act as a one, two, or five electron
donor to a metal centre depending on the hybridisation of the oxygen atom as
(a) sp’, (b) sp’, or (c) sp.

R—O R——O R—oO==m

SX

(a) (b) (c)

Since the lanthanide and actinide elements tend to form ionic complexes and
have large numbers of orbitals with suitable symmetry for overlap with
ligand orbitals, it is most usually the case that linear or near-linear R-O-M
fragments are observed, and we may describe the alkoxide radical as a five
electron donor. Similarly, the fragments R,N—M are almost invariably planar
and the dialkylamide radical is a three electron donor. In general, however, the
 The f elements 61

number of electrons donated has little effect on how many ligands the metal
centre can accommodate. This is largely determined by steric effects (i.e. the
size of the ligands).

Nuclearity, coordination number, and bulky ligands

The oxygen and nitrogen atoms of alkoxide and dialkylamide ligands may, by
virtue of their lone pairs, act as bridging ligands between two metal centres,
giving oligomeric structures in the solid state (as determined by X-ray
crystallography) and in solution (as determined by freezing point depression
and other colligative properties). For example, the tetravalent
tetrakis(diethylamide) complex [U(NEt,),] is actually a dimer 5.3(a) with six
terminal amide and two u’-bridging amide groups. Each uranium atom is five
coordinate. It is important to note that all the U-N bridging bonds are
equivalent, and the valence bond picture 5.3(b) is an aid to assignment of
formal oxidation state rather than prediction of structure.

5.3 (a)

Nuclearities of complexes are governed or can be controlled by:

(a) Size of groups R This is the key to the design and synthesis of
monomeric homoleptic complexes (i.e. complexes where all the ligands
are the same).* Unlike 5.3, the tetrakis(diphenylamide) complex
[U(NPh,),] is a tetrahedral monomer 5.4. Use of the even bulkier re €\ =
bis(trimethylsilylamide) ligand means that only three amide groups can LS \ NE
be accommodated at an f element centre. Synthetically useful complexes UN N \ y
such as [UCI{N(SiMe,),},] and [Nd{N(SiMe;),},] can be made by the /
methods described later in this Section. N
(b) Oxidation state Since both alkoxide and amide are monovalent ligands,
the formal oxidation state of the metal in a neutral monomeric
compound corresponds to its coordination number. Hence, higher
oxidation state complexes will tend to have lower degrees of
polymerisation. The deep red hexavalent alkoxide complex [U(OMe),] 5.4

* The corresponding parameter, ionic radius of metal, is also important in

determining nuclearity, but few data are available.
62 Coordination chemistry

5.5, for example, has been shown by X-ray crystallography to be

monomeric in the solid state with an octahedral UO, core. It retains this
structure in solution. As a consequence of this low nuclearity and the
encapsulation of the metal atom inside a sphere of low-polarity alkyl
groups, the complex sublimes at 87 °C in a vacuum. By contrast, the
analogous pentavalent complex [U(OMe).] is probably a trimer and
correspondingly does not sublime below 140 °C at the same pressure.
The volatility of [U(OMe),] led to its
being investigated for use in gas-
phase methods for the isotopic Me
enrichment of uranium.
\
O

Me——04,,,, | w'tO— Me
"1 eet

om |
So Me
fo 9

5.5

For a given alkyl group R, the dialkylamide ligand R,N- is far more
sterically demanding than the alkoxide ligand RO-, simply because there are
two alkyl groups in the former. As a consequence, it is more difficult to
design monomeric alkoxide complexes of the f elements. For example, the
trivalent cerium complex [{Ce(OCHBu',),},] 5.6 is a dimer despite having
two ‘Bu groups on each alkoxide ligand 5.7. For this reason, several classes
of exceptionally bulky aryloxide ligands have been used. One of the most
successful 5.8 is derived from an inexpensive substituted phenol which is
used as an antioxidant in polymers and foods. The complex [Ce(OAr),] 5.9 is
monomeric and sublimes readily.

Me a

a Me

-O Me

Me Me

5.7
 The f elements 63

Me Me

5.8

Synthesis and reactivity

Alkoxide and dialkylamide complexes of the f elements have been synthesised
by many routes, but the following two are the most general.

Salt elimination (metathesis)

The deprotonated (or base) form of the desired ligand is allowed to react with a
metal salt incorporating anions which are good leaving groups (Eqns 5.13—
14. The anhydrous lanthanide and actinide chlorides LnCl, and AnCl, (An =
U, Th) are by far the most popular starting materials here because they are
commercially available or are readily synthesised (Chapter Four, Section
One). Where lower valent complexes are desired, the iodides or bromides must
usually be used. For example the divalent lanthanide sources LnI, (Ln = Sm,
Eu, Yb) are readily available and the uranium triiodide [UI,(THF),] provides a
useful synthon for the trivalent coordination chemistry of this element in
anhydrous solvents.
LnCl, + 3LiOAr* — [Ln(OAr*),] + 3LiClN (5.13)

AnCl, + 3 K[N(SiMe,),] > [AnCI{N(SiMe,),},]+3KCN (5.14)

Control of stoichiometry is important here, particularly with less bulky
ligands. If an excess of ligand salt is used, the strongly Lewis acidic f element
centre can accommodate anionic groups in excess of the number required by
its formal oxidation state to give what are known as ate complexes. Although
this phenomenon usually occurs unintentionally and causes problems in
isolation of the products, ate complexes are sometimes synthetically useful as
shown in Fig. 5.14. The salt 5.10 is probably zwitterionic, with lithium
ions incorporated into the structure by coordination to methoxide lone pairs.

OMe 2-
- 4LICN MeOr,,, | ss OMe Pb‘+
UCI, + 6 LiOMe————> (Lit), 2S ————_ [U(OMe)¢]
Meo |™ ome
5.5
OMe

5.10
Fig. 5.14 Synthesis of a hexavalent uranium alkoxide via a tetravalent ate complex.
64 Coordination chemistry

The pre-formed octahedral array of methoxide ligands required for steric

saturation of the hexavalent product 5.5 facilitates the smooth oxidation of
tetravalent 5.10 by lead(IV)acetate.

Protonolysis
The protonated or acid form of the desired ligand is allowed to react with a
pre-formed complex such as an alkyl, dialkylamide, or alkoxide. This
strategy, which circumvents the problems mentioned earlier of ate complex
formation, works for the f elements because the bonds are so polar, and thus
is not so useful in the more covalent transition metal systems. The driving
force of the reaction (Fig. 5.15) can be provided by (i) the relative strengths of
the metal-element bonds or the pK, of the ligands (ii) the production of a
volatile ‘acid’ or (iii) the chelate effect (Section 5.4).

OH
ul But But But

Me3Si. H
—..--——_
ile ge
O Bu

Me3Si

Bu!

DAR
N(CH,CH NHR), 4 PhaNH
<< 1/2 [{U(NEt,),})] ————_ [U(NPho)4] + 4 EtpNHT
(it) 5.3 (i) 5.4

Fig. 5.15 Synthesis of f element alkoxides and dialkylamides using acidic reagents.

5.6 Lanthanides in organic chemistry and catalysis

The unique properties of the lanthanides (e.g. their redox chemistry and
affinity for hard ligands) are exploited in organic functional group
transformations and in various areas of catalysis.

Oxidising and reducing agents

Although complexes of Ce*t and Sm** are readily accessible, the most stable
oxidation state of these elements under normal conditions is +3. This means
that Ce** and Sm** will act as one electron oxidising and reducing agents
respectively.

Cerium(IV) promoted oxidations

Ceric ammonium nitrate [Ce(NH,),(NO;),] (known as CAN) and _ the
analogous sulphate CAS have been used extensively for many years in a
variety of selective and efficient oxidative transformations. For example,
 The f elements 65

while oxidation of polynuclear aromatic systems such as fluoranthene usually

gives a complex mixture of products, use of CAN leads to the o-quinone
product in high yield [Fig. 5.16(a)].

oe Ce(IV) 2

: silent ke ee mt oe
is OAc

Fig. 5.16 Selective oxidation using tetravalent cerium.

Oxidative coupling reactions are also mediated effectively by tetravalent

cerium. For example, 1|,4-dicarbonyl compounds are readily prepared by the
CAN promoted reactions of ketones with vinyl acetates [Fig. 5.16(b)]. This
reaction proceeds in higher yield and is more regioselective than traditional
oxidative coupling procedures using Mn(III) salts.

Samarium (II) promoted reductions

Samarium(II) iodide (SmI,) is a mild, soluble one electron reducing agent
widely used in organic synthesis. It is usually prepared in situ by the reaction
of samarium metal with CH,I, to give a deep blue solution.

Sml>

Fig. 5.17 Sml, promoted reductive cleavage of C—X bonds in the o.-position of ketones.
66 Coordination chemistry

A wide range of a-substituted ketones is reduced by Sml, under mild

conditions (Fig. 5.17). The reaction is very selective and, for example, will
not affect isolated ketones. Even hydroxyl groups are reductively cleaved in
this way, thus providing a useful entry to unsubstituted ketones in natural
product syntheses.
Reductive cyclisations can also be performed, and the reactions of SmI,
with ®-bromoacetoxycarbonyl substrates gives valerolactones__ with
exceptionally high degrees of diastereoselectivity. This is proposed to arise
from the formation of a rigid cyclic transition state structure enforced by
chelation, and is a good example of exploitation of the combined oxophilic
and reducing properties of samarium(II) (Fig. 5.18).

H O-=> Smi(IIl)

Fig. 5.18 Chelate control in a reductive cyclisation reaction mediated by Sm(Il).

Lewis acid reagents and catalysts

Lanthanide based Lewis acid catalysts have many advantages over
conventional main group and transition metal compounds:
(a) The ions coordinate preferentially to hard bases such as oxygen and
nitrogen, thus leading to selectivity.
(b) Lanthanide complexes are labile, and thus substrate and product
molecules may exchange rapidly at the metal centre. This leads to fast
acid catalysed reactions.
(c) The lanthanide ions are tolerant to air, water, and most importantly a
wide range of functional groups in the organic molecule.
ae -- Ce(ll
(It) Ketones such as 2-indanone [Fig. 5.19(a)] are susceptible to enolisation
(i.e. deprotonation at the CH, groups) by basic nucleophiles such as alkyl
xae CH»Ph lithium reagents. Pre-treatment of lithium reagents with anhydrous CeCl,
leads to the formation in situ of less basic cerium alkyl species, ‘RCeCl,’,
the actual structures of which are unknown. Nevertheless, the ability of the
Fig. 5.20 Directing affect of Ce(IIl) in Ce** ion to coordinate to the carbonyl group and thus to direct the
carbonyl alkylation.
nucleophilic attack of the carbanion (Fig. 5.20) leads to the formation of
alcohols with high selectivity.
 The f elements 67

PhCHp>Li/CeCl, OH
(a) OF
CH>Ph

e PhCH,Cl oe
(b) ee oe
Cat. TmCl,

Fig. 5.19 Exploiting the Lewis acidity of lanthanides in (a) selective nucleophilic addition and
(b) Friedel-Crafts alkylation.

The well known Friedel-Crafts alkylation of aromatic molecules catalysed

by AICI, suffers from several problems. It is difficult to prevent multiple
alkylation, the catalyst is sensitive to water, and it cannot easily be recovered
in an active form. The later (smaller) lanthanide trichlorides are also effective
catalysts for this reaction. Only small amounts of dialkylated products are
formed and the lanthanide catalyst can be recycled effectively after aqueous
work-up.

Heterogeneous catalysts
One major use of lanthanides in heterogeneous catalysis is the stabilisation of
zeolite structures used in petroleum cracking applications. Although the
precise function is somewhat unclear, the addition of small amounts of
lanthanide halides to the zeolite allows it to maintain the high degree of
acidity required, under the harsh conditions present in an industrial reactor.
Another important use is in the exhaust catalysts in cars. This was
discussed in Chapter Four, Section Two.

5.7 Exercises
1. Why is it that carbonate CO,” in natural groundwater can compete
effectively for coordination sites at actinyl centres even though the
excellent ligand H,O is in vast excess?

2. Using Fig. 5.12, predict the equilibrium constant for the formation of
the 1:1 Y**/EDTA* complex in aqueous solution.

ak The hexavalent alkoxide [U(OMe),] (5.5) has C—O-U angles ranging

from 125°-175°. Why do you think they are not all linear i.e. 180°?

4. Suggest a synthetic route to the complex [U{ N(SiMe;), },(OPh)].

6 Organometallics

Some of the most important advances in understanding the behaviour of the f

elements have arisen from the study of their organometallic complexes. In
this Chapter we describe the methods that are used to synthesise these
compounds, the structures that they adopt, and the physical and chemical
properties that they display. The aim is to paint a clear picture of the often
unique behaviour of the organometal!lics of the f elements and to explain it
using the concepts described in earlier Chapters.

6.1 Comparisons with the d block organometallics

The discovery of ferrocene [Fe(n°-C,H;),] 6.1 in 1952 was followed in 1956
by the synthesis of [U(n°-C5H,);Cl], the first well-defined organometallic
complex of an f element. It was, however, the synthesis of the
cyclooctatetraene sandwich complex uranocene 6.2 in 1968 which indicated
that, while the f elements would have an organometallic chemistry related to
that of the transition metals, there would be some contrasts between the two
series. The differences in stability, reactivity, and structure between d and f
block organometallics can be traced to the large atomic radii of the latter and
also to their lower propensity to form covalent bonds (Chapter Five, Section
One). For example while carbon monoxide readily forms complexes with
many of the transition metals, this behaviour is extremely rare for the
lanthanides and actinides and such carbonyl compounds have only recently
been isolated (Section 6.4).
For the transition metals, the effective atomic number (or 18 electron) rule
can be used very effectively to forecast the stability of a proposed
organometallic molecule. For the f elements, no such simple rules exist.
Formal electron counts and coordination numbers are determined almost
exclusively by steric effects, i.e. the size of the metal atom and its ligands. In
addition, molecular geometries of transition metal complexes can usually be
predicted or rationalised with reference to the crystal field or molecular orbital
theories. For the lanthanides and actinides, the large numbers of electrons
involved and the effects of relativity (Chapter Three, Section Two) greatly
increase the complexity of the theoretical models required to describe their
electronic and geometric structures, and accurate calculations have only
recently become possible with the advent of powerful computers.
The sensitivity of most f element organometallics to oxygen and water is a
property which they share with the early the d block compounds, but this
problem is exacerbated by the large atomic radii of the lanthanides and
actinides. It is with the advent of modern techniques for the rigorous
exclusion of air and the development of sterically demanding ligands that the
field has expanded to the point where, like the d block organometallics,
applications in catalysis and other areas are beginning to emerge.
 The f elements 69

6.2 Alkyls: achieving steric saturation and stability

Synthesis
In the context of the problems outlined in Section 6.1, the synthesis of f
element alkyls MR, presented a great challenge, but nevertheless this has
been achieved by two general methods.
First, by analogy with the aryloxides and dialkylamides described in
Chapter Five, Section Five, it is possible to synthesise homoleptic
complexes of the f elements such as [U{CH(SiMe,),},] 6.3 using the
extremely bulky bis(trimethylsilyl)methyl ligand.

i ne Me 3-
Vi
Me3Si. H U tte
Saree
SiMe, Me aa
‘lk = Me |
" En
Me

M : H
“SiMe, sn Me
a~\~
| Me
€3Si M
SiMe, Me

6.3 6.4 6.5

Second, in some instances where smaller alkyl groups are used, the metal
centre may accommodate more alkyl ligands than it requires in order to satisfy
its valency. For example, the trivalent lutetium centre in 6.4 has four
2,6-dimethylphenyl groups in a tetrahedral array and thus has an overall
negative charge. Such anionic ate complexes are relatively common in f
element chemistry (see Chapter Five, Section Five). A quite spectacular
example of this behaviour is given by complexes such as_ the
hexamethylerbiate trianion 6.5. Similar compounds are known for most of
the lanthanides and thorium.

B-elimination and the agostic interaction

Many alkyl complexes of the f and d elements undergo decomposition via
B-elimination of an alkene (Fig. 6.1).

oak B-elimination Py es

Ren R
fi eee
Ap 6.7
6.6

Fig. 6.1 Elimination of alkenes from metal alkyl complexes.

The f elements do not form strong bonds with 7 ligands such as alkenes
(Section 6.4) and so the intermediate complex 6.6, if formed at all, will
rapidly lose alkene leaving a hydride species 6.7 which may decompose
further.
70 Organometallics

In order for B-elimination to occur:

(i) |The B carbon of the alkyl must bear a hydrogen substituent.
(ii) The M—C-C-H fragment must be able to take up a planar conformation
such that the hydrogen atom makes a close approach to the metal.
(iii) There must be an empty orbital on the metal which is of appropriate
energy and orientation to bond to the 8 hydrogen atom.
Considering requirement (iii) and the fact that the f elements have a large
number of orbitals in various orientations, it is not surprising that their alkyl
complexes are particularly susceptible to B-elimination. It is thus a common
tactic in the design of low coordination number f element (and indeed
transition metal) alkyl complexes to use alkyl ligands for which
B-elimination is not possible. For example, the alkyl ligands in 6.3 have no
hydrogen atoms on the B position (in this case, silicon) with respect to the
metal. Other important alkyl ligands are neopentyl -CH,Bu' and benzyl
-CH,Ph.
In instances where B-elimination is prevented by the above strategy,
another phenomenon may occur. A C-H bond may approach the metal centre
SiMe, closely as if it were moving toward the transition state in an elimination
Me™ Varn /simeg process. The C-H—M bridge thus formed (Fig. 6.2) is known as an agostic
‘SiMe, interaction. For example, the formal coordination number in 6.3 is three, but
H H it has been determined that an agostic interaction is formed between one
Me3Si SiMe; methyl group from each ligand toward the uranium centre. The C—H bonds
involved are in the y position with respect to the metal, and so this is known
Fig. 6.2 One of three y-agostic
as a Y-agostic interaction. It is probably the case that many f element
interactions present in the uranium
alkyl 6.3. A straight half-arrow is used
organometallics which apparently have very low coordination numbers
to represent the agostic bond. actually contain agostic interactions.

6.3 Cyclopentadienyls: spectator ligands

As is the case for the transition metals, much of the organometallic chemistry
of the f elements is based around complexes containing the formally
monoanionic cyclopentadienyl ligands, C5R; (R = e.g. H, Me). Such
complexes have useful properties including high reactivity, and stability with
respect to decomposition. The cyclopentadienyl ligands themselves do not

A
often take part in chemical transformations, and thus they are an example of
what are known as ‘spectator ligands’. Cyclopentadienyls are capable of
forming both covalent and ionic complexes with between one and five carbon
Yo— OBu! atoms bonded to the metal centre. In the n° bonding mode they can be

a 6.8
considered to occupy three coordination sites. For example, we can describe
the trivalent ytterbium complex 6.8 as being seven coordinate. One
important advantage of these ligands is that their steric demand (or how
effective they are at protecting the metal centre) can be altered subtly and
methodically by introduction of alkyl group substituents on the ring. Early
chemistry was based around the simple C;H; ligand known as Cp, but today
the use of very bulky ligands is more common.
 The f elements 71

Simple cyclopentadienyls
For the trivalent lanthanides, three general types of cyclopentadienyl
complexes are available, [LnCp,X;,,] where n = 1-3. These compounds are
made by reaction of the anhydrous lanthanide halides with sodium
cyclopentadienyl (Eqn 6.1). Ether solvents such as THF work best, but since
they are themselves powerful ligands to lanthanide and actinide centres, it is
often difficult to remove them subsequently.
LnCl, + n NaCp > [LnCp,Cl,.,] + n NaCl (6.1)
The compounds [LnCp,] react with FeCl, to give ferrocene, as would an
ionic cyclopentadieny! such as NaCp, and this is taken as an indication of the
polar nature of the Cp—Ln bond. This polarity also manifests itself in the fact
that quite rapid exchange of Cp ligands occurs between lanthanide centres.
Nevertheless, compounds containing the Cp,LnCl unit can be isolated as
dimers 6.9 or as Lewis base adducts such as 6.10.

The tris(cyclopentadienyl) lanthanide compounds [LnCp,] sublime when

heated in a vacuum, and this is a convenient way of removing donor solvents
(e.g. THF, diethyl ether). The solid state structures of the ‘base-free’
compounds subsequently obtained vary across the 4f series in a way which
reflects the lanthanide contraction (Chapter Two, Section Three). The larger
lanthanides such as lanthanum and praseodymium have ten coordinate
structures 6.11 where each metal centre has one 1°-Cp and three 1°-Cp
ligands. The n?-Cp ligands are further bonded 7° to another metal centre, thus
forming a coordination polymer in the solid state. The slightly smaller
thulium and ytterbium are nine coordinate 6.12. Lutetium, the smallest
lanthanide, forms an eight coordinate structure 6.13. Interestingly, the group
three elements fit into this scheme according to their atomic radii with
yttrium adopting structure 6.12 and scandium that of 6.13.

6.11 6.12
The trivalent actinide complexes behave similarly to their lanthanide
analogues, and although the compounds [AnCp,] are well known, complexes
72 Organometallics

such as [AnCpCl,] are often poorly characterised. Many of these problems are
solved by the use of bulky substituents on the cyclopentadienyl ligands, as
we shall see later.
Most of the cyclopentadienyl chemistry of the actinides, however, involves
the metals in a tetravalent state, and complexes [AnCp,Cl,.,,] (n = 3, 4) can be
prepared as in Eqn 6.2. A combination of the high affinity of the actinide ions
for cyclopentadienyl rings and the relatively modest steric demand of the
unsubstituted Cp ligand makes it impossible to stop reaction in Eqn 6.2 at
the [AnCpCl,] and [AnCp,Cl,] (i.e. n = 1, 2) stages. The vast majority of
tetravalent cyclopentadienyls of the actinides are, as expected, of thorium and
uranium, although some protactinium and neptunium complexes are known.
AnCl, + n NaCp > [AnCp,Cl,.,] + n NaCl (6.2)
The fully cyclopentadienylated compound [U(n’-Cp),] 6.14 has been
shown to have the Cp ligands arranged at the vertices of a tetrahedron. This is
the only type of organometallic compound which has four Cp groups all
bonded in the n° fashion. In comparison, the analogous zirconium compound
has the structure [Zr(n°-Cp),(n'-Cp)].
Unlike the lanthanide cyclopentadienyls, tetravalent [UCp,Cl] does not
react with FeCl, to form ferrocene, and this has been taken as an indication of
a significant degree of covalency in the Cp—An(IV) bond. The chloride ligand
in [UCp,Cl] and in the analogous [ThCp,Cl] can be readily substituted
making these compounds among the most important synthons in_ the
organometallic chemistry of the actinides (Fig. 6.3). The three
cyclopentadienyl groups protect the metal centre quite effectively, but at the
same time allow chemistry to take place at the remaining coordination site.

AnCl, [AnCp,(OEt)]
W (i)

Si

(ii & rl (iv)

[AnCp3(CH3)] [AnCp3(BH,)]

Fig. 6.3 Synthesis and some reactions of [AnCp,Cl]. Reagents (i) 3 TICp, (ii) NaOEt,
(iii) LiMe, and (iv) NaBH,,.

Substituted cyclopentadienyls
Replacement of C5H, (Cp) with bulkier analogues such as C,Me, (Cp*) and
C;H;(SiMe;), (Cp") leads to complexes with more convenient properties.
They are soluble, form crystals more readily, and do not suffer so much from
the aforementioned ligand redistribution reactions. In the same way that the
CH, ligand facilitated the exploration of the organometallic chemistry of the
d block, the use of bulky cyclopentadieny! ligands such as Cp* paved the way
for the f elements.
 The f elements 73

The tris(pentamethylcyclopentadienyls) [MCp*,] are unknown for the

actinides and for most of the lanthanides, most probably because Cp* is too
bulky. This relative inaccessibility of complexes with three Cp* ligands is,
however, quite convenient since it makes it rather easier to synthesise
complexes of the lanthanides [LnCp*,Cl], and actinides [AnCp*,Cl,] (n = 1,
2). These compounds are very versatile starting materials for the exploration
of the chemistry of the MCp*, fragment. The complexes [AnCp*,Cl,] (An =
U, Th) are structurally analogous to [ZrCp,Cl,] and, like the zirconocenes,
have an extensive chemistry. The thorium chemistry is particularly well
developed, not least because complexes of thorium(IV) are diamagnetic and
thus readily studied by NMR spectroscopy. The dichloride 6.15 (Fig. 6.4)
can be readily converted to the dimethyl 6.16 and subsequently the dihydride
6.17 which is an active catalyst for the polymerisation of ethene and the
hydrogenation of a-alkenes. The hydride also undergoes insertion reactions
with polar double bonds such as carbonyls. CO itself is reduced at low
temperatures to the n*-formyl compound 6.18.

6.18 6.17

Fig. 6.4 Some reactions of the compounds Cp*,ThX;.

The trivalent lanthanide complexes [LnCp*,X] are extremely reactive. In

particular the lutetium compounds (X = H, Me) are one of the few systems
capable of mediating the exchange of C—H bonds, a process known as 0 bond
metathesis (Fig. 6.5).
74 Organometallics

CK CH LE CHs -CH 2

SX
Be SS L 0)ww H ee Lu— CH,

Wee Se
Fig. 6.5 Hydrocarbon activation via o bond metathesis mediated by a lutetium alkyl 6.19. The methyl group in 6.20 can be thought of
as acting as a strong base which deprotonates the alkane.

Low oxidation state cyclopentadienyl complexes

The favourable spectator ligand properties of cyclopentadienyls make them of
great use in the synthesis of complexes of the f elements in low oxidation
states. Many divalent complexes have been made with samarium, europium,
and ytterbium since these lanthanides have the most readily accessible +2
oxidation state. Other more unusual oxidation states have been also observed,
e.g. Th(iI).

Samarium(II)
Lewis base free complexes such as [SmCp,] are expected to be strongly Lewis
acidic, as well as being powerful one electron reducing agents. Unfortunately,
the bis(cyclopentadienyl) complex 6.21 synthesised by the route in Fig. 6.6
incorporates at least two THF ligands into its coordination sphere, and these
are difficult to remove. Use of the bulkier Cp* ligand, however, allows the
removal of coordinated THF from the complex 6.22. The molecular structure
of 6.23 (Fig. 6.7) reveals that it has a bent structure, as do the analogous
europium and ytterbium compounds. The reasons for this are not fully
understood, and similar transition metal complexes (such as ferrocene 6.1)
adopt a parallel structure. As expected, the complex 6.23 is highly reactive,
Fig. 6.7 Molecular structure of as we shall see in Section 6.4.
[SmCp*,] 6.23.
2 NaCp
[SmCp2(THF)o] —«————_—_——-._. Sm
6.21 THF
2NaCp*| THF

C
wae ce LE eee
1 Sm"

4 Ss _e
6.23 6.22
Fig. 6.6 Synthesis of samarium(ll) cyclopentadienyls.
 The f elements 75

Thorium(III)
The aqueous chemistry of thorium is confined to the +4 oxidation state since
the trivalent state reduces water rapidly. Even in organometallic chemistry,
only a few compounds containing trivalent thorium have been detected, and
these are exceptionally sensitive to oxygen and water. It was only by the use
of the sterically demanding Cp" ligand that the structure of such a compound
6.24 could be characterised by X-ray crystallography (Fig. 6.8).

SiMe;

wee SiM €3
Na/K alloy pies ak) ”
3 iiiCp Cl) —_{___ 2 +

toluene Me,Si ae
IM@3

6.24
SiMe,

Fig. 6.8 Synthesis of a Th(IIl) organometallic.

In the same way that the Ti(II) centre in [TiCp;] has the electronic
configuration [Ar]3d' we might expect Th(III) in 6.24 to be [RnJ5f'.
Relativistic calculations on the model complex {[ThCp,], however, show that
the 6d level in this molecule is surprisingly stable and that the electronic
configuration is [Rn]6d'. This is consistent with experimental data in that the
electronic spectrum of the complex consists of a series of strong (i.e. Al = +1
allowed) d — f bands, and the molecule is an intense blue colour.

6.4 Carbonyls and related complexes: 2 bonding?

The carbonyl ligand CO is extremely important in transition metal
organometallic chemistry. The stability of the M—CO bond arises from the
synergic combination of ligand—metal o donation and metal—ligand mt back-
donation as shown in Fig. 6.9.

o donor m™ acceptor

P2205
chopo0 GSEs
Fig. 6.9 The main bonding interactions in metal carbonyl complexes.

For the f elements, this type of covalent bonding interaction is difficult to

achieve; the filled metal orbitals are generally too deeply hidden within the
core of the atom to overlap effectively with ligand 7 orbitals. Nevertheless,
co-condensation of lanthanide vapours with CO diluted with argon at low
temperatures (< 40 K) results in the formation of mixtures of the carbonyls
M(CO), (n = 1-6) as identified by their infrared spectra. On increasing the
temperature, these transient complexes decompose immediately.
76 Organometallics

It was pointed out in Chapter Two, Section Two that the greater range of
oxidation states exhibited by the early actinides indicates that the valence
electrons are less tightly bound than they are for the lanthanides and later
actinides, and that these electrons may take part in covalent bonding to
ligands. This phenomenon is illustrated by the synthesis under fairly standard
laboratory conditions of CO complexes of trivalent uranium surrounded by
bulky cyclopentadienyl spectator ligands, [UCp';(CO)] 6.25 and
[U(n>-C;Me,H),(CO)] 6.26. Infrared spectroscopy shows that the
wavenumbers of the carbonyl stretch v(CO) for these two compounds are
1976 cm and 1900 cm’ respectively. Since these values are rather lower than
that in free CO (2146 cm'') there must be a significant amount of U-CO n
backbonding. This conclusion is supported by molecular orbital calculations
on the model compound [UCp,(CO)}.

oS
Ss
Me3Si

> SiMe,

6.25 6.26

No such molecular CO complexes of the lanthanides exist, and this is

consistent with the emerging picture that these elements are less capable of
forming covalent bonds than are the (early) actinides. This difference in
behaviour is also borne out by studies of complexes of the isocyanide ligands
CN-R which are isoelectronic with CO. The lanthanide complexes
[LnCp,(CN-Et)] have v(CN) values that are actually higher than that in free
EtNC. This behaviour is also observed in the isocyanide complexes of high
oxidation state transition metal ions and indicates that there is little or no 7
backbonding to the ligand from the metal. By contrast, the early actinide
compounds have v(CN) values which are slightly lower than the
uncomplexed isocyanide.
Dinitrogen N, is isoelectronic with CO but is far less efficient at forming
complexes, even with the transition metals. For the f elements, dinitrogen
complexes are exceptionally rare. The complexes 6.27 and 6.28 are formed
by the addition of N, to low valent metal centres Sm(IT) (Section 6.3) and
UID respectively. We will return in the following section to the importance
of low formal oxidation states in the formation of m complexes with the f
elements.
Alkene and alkyne complexes of the lanthanides and actinides are very
difficult to make. The complex 6.29 is a rare example, and is stable only
because the Pt(0) centre is a good 7m donor, making the bridging 1’-ethene
ligand in turn a sufficiently strong Lewis base to coordinate to the divalent
[YbCp*,] fragment.
 The f elements 77

b- - f- - =Pt(PPh3)3

6.27 6.28 (R = SiMe,Bu') 6.29

6.5 Arene complexes: an unexpected (covalent) bond

By contrast to CO which is a 6 donor/m acceptor ligand, n°-arenes are best Throughout Sections 6.5 and 6.6, o, 7,
5, and @ refer to the symmetry of the
regarded as m™ donor/d acceptors. Metal—arene backbonding interactions are
bonding interaction with respect to the
comparatively weak since overlap between 6 orbitals is difficult to achieve metal/igand axis. They indicate the
both on steric grounds and because arene orbitals with 6 symmetry are usually presence of zero, one, two, and three
high in energy relative to metal d orbitals. The consequence of this is that nodal planes respectively.
compared with CO, n®-arenes form highly electron-rich (or high energy)
complexes, and are usually formed only where the metal is in a low oxidation
state. The f elements, with their deeply buried valence orbitals and tendency to
form tri- or higher-valent complexes, are thus not the ideal candidates to form
such compounds. Despite this, some notable successes have been achieved in
the synthesis of complexes which cannot be adequately described by an ionic
bonding model.

SmCl, 3 + Al + AICI, 3 6 Meg ———>

+ CeMe af ol Cc i Cl

i Cl
a LY:
ra
Cl Cl

Bu!

-196°C | Bu! Ln = Pr, Nd, Gd, Tb, Dy,

LNatoms + CeHgBu'3 ——————- ar [ Ho, Er, also Sc, Y

< > Bu
Bul 6.31

Fig. 6.10 Synthesis of arene complexes of the lanthanides.

The classical (Fischer-Hafner) synthesis of arene complexes such as

zerovalent [Cr(n°-C,H,),] involves reduction of the metal halide in the
78 Organometallics

presence of the arene proligand. Under similar conditions (Fig. 6.10),

samarium forms a trivalent complex 6.30. Similar U(II) and U(IV)
complexes are also known, but in all cases the M-(arene) bond is, as expected
from the above arguments, rather weak.
Co-condensation of the vapours of certain lanthanides with an excess of the
extremely bulky arene C,H,Bu', leads to formation of the zerovalent
complexes [Ln(n°-C,H;Bu';),] 6.31, some of which are sufficiently stable to
survive sublimation at 100 °C in a vacuum. The extremely bulky ligand
helps to stabilise the complex kinetically, just like the Cp* ligand in
cyclopentadienyl chemistry.
The bonding situation in these molecules has been explained by analogy
with the bis(arene) complexes of the early transition metals whose electronic
structures have been studied in depth. Molecular orbital diagrams for these
complexes, e.g. [Cr(n°-C,H,),] are described in most text books on
organometallic chemistry. A summary of the properties of the five 3d based
molecular orbitals (€),, a), €;,) of such a system is presented in Fig. 6.11. In
18 valence electron [Cr(n°-C,H,),], all the orbitals up to and including the
weakly 0 bonding a,, would be filled. The 15 valence electron compound
[Sc(n°-C,H,Bu';),] would have an empty a,, orbital and a three quarters-filled
e,, 6 bonding orbital as shown. This latter complex is extremely air sensitive,
but is otherwise quite stable, and so it seems that the presence of these three d
electrons in the 6 bonding orbital is enough, in combination with the o and 7
bonding orbitals of lower energy, to form a satisfactory arene-metal bond.
Note that the ground state Let us replace the scandium atom in this model with a lanthanide atom of
configuration of Sc is [Ar]3d'4s? but in
typical electronic configuration [Xe]4f"5d°6s* (see Chapter Two, Section
its zerovalent complexes it is
regarded as being [Ar]3d°. One). Can these f electrons interact effectively with the arene 6 orbitals?
Probably not: lanthanide forbitals are quite core-like and are unlikely to take
part in ML backbonding, as we have seen in Section 6.4.

C49,0"

Filled HOMO
in 18 electron = —-_ 419 9
complexes

= B29, 8

12
jelectrons o and

Fig. 6.11 A section of the molecular orbital energy level diagram for [Sc(y ®-C,H,Bu',).]
showing the orbitals around the HOMO.
 The f elements 79

So what is it that is holding the bis(arene) lanthanide molecules together?

Consider the variation in the energy required to promote an electron from a 4f
orbital to the 5d, i.e. [Xe]4f5d°6s? > [Xe]4f"'Sd'6s? as we move across the
lanthanide series (Fig. 6.12). The metals which form a stable compound
[Ln(n°-C,H;Bu',),] are indicated with dark circles. It can be seen that these YOU might
liketo think oftheprocess
complexes are formed where the promotion energy is low (or in the case of eas cee Naa ee
gadolinium, negative). The Ln-arene) bond energies gained compensate for merorahy! GOUDe Real
the promotion energy which has to be put in.
Figure 6.12 also shows that there is an area of instability of these
complexes at the start of the lanthanide series which is not explained by the
promotion energies model. The lanthanum and cerium complexes decompose
at room temperature or below. The argument applied here is that as these are
the lanthanides with the largest radii, even C,H,Bu'; is not bulky enough to
protect the metal centre.

30

ro)
8

oO

Promotion
cm"
10°
energy
x

La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Ef Tm Yb

Fig. 6.12 Variation in promotion energy [Xe]4f'6s* — [Xe]4f*'5d'6s? across the lanthanide series. This graph mirrors closely the
variation in third ionization energies (Fig. 2.2).

6.6 Cyclooctatetraene complexes: d versus f orbital

covalency
Cyclooctatetraene (COT) requires two electrons in order to achieve its stable
ten electron aromatic form (C;H,)’ and so one such n*-COT ligand would
normally be expected to satisfy two metal valencies. It is also a quite
sterically demanding ligand, being similar in effective size to Cp* (Section
6.3). COT is therefore ideally suited to the formation of complexes with
lanthanides and actinides, perhaps more so than with the transition elements.
As was mentioned earlier, COT complexes were among the first f element
organometallics, and they are still being studied in great depth, particularly in
relation to the nature of the M-(COT) bond.
80 Organometallics

Uranocene [U(n°-C,Hs),]: synthesis, structure, and bonding

Uranocene, [U(n*-COT),] 6.2, and various ring-substituted derivatives can be
prepared by a number of methods, but the simplest is the salt elimination
route (Eqn 6.3), similar to that used in the synthesis of the cyclopentadienyls.
UCL +2 K,(COT) —S [U@"-COT),] + 4 KCI (6.3)
The structure of this fascinating compound shows planar and parallel COT
rings which are eclipsed in conformation giving the molecule Ds, symmetry.

(c) @oy/f26.2-y2)

Fig. 6.13 Symmetry matched combinations of COT ring and actinide orbitals of 5
symmetry: the major source of bonding in actinidocenes. You might like to think of the metal
f,2,2) in (c) as comprising two d,2.,2 orbitals (a) of opposite phases stacked on top of one-
another. The d,, and f,,, orbitals in (b) and (d) are similarly related.

A number of spectroscopic and theoretical studies have concluded that the

The f orbitals shown in Fig. 6.13 are covalent interaction between ring €,, orbitals and the uranium 6d orbitals of
part of the general set (Chapter Two,
matching symmetry is the major source of bonding in uranocene. These 6
Section One).
bonds are represented in Fig. 6.13(a) and (b). Note that each individual COT
ring orbital has two vertical nodal planes but that there is no nodal plane
 The f elements

between the two rings. Hence these are gerade (subscript g) combinations and
have symmetry matches with the d orbitals shown.

655; 0

Se e:

a1, O* ———————-- - ----- : me 7 i

1g iT VONOCOnoitals

- .
wos
beg os
=.
wee eee eee eee ee ee seven 5f orbitals

Gnu and pq (5) =...

€y, and on SS...

Boy ANd Ayg (CO) Te

eis. =p Heajg and ap,

0) 0 U

Fig. 6.14 A molecular orbital scheme for D,,, symmetry [U(n®-COT),]: the ligand/metal orbital combinations giving rise to the &
bonding orbitals in the box are shown in Fig. 6.13.

The second most important interaction is between the ligand e,, orbitals
and the uranium 5f. These bonds are depicted in Fig. 6.13(c) and (d). In
addition to the two nodal planes within each COT ring, there is a nodal plane
between the two rings. These ungerade combinations have symmetry matches
with the f orbitals shown. It is important to note that these e,, symmetry 6
 82 Organometallics

interactions, which play a crucial role in stabilising the sandwich structure,

are not possible using d orbitals on the metal, and so this type of bond is
only available to compounds of the f elements.
Several factors combine to make 6 bonding more important in uranocene
than it is in the bis(arene) lanthanide complexes described above. First,
because the carbocyclic ring is larger, the metal must sit closer to the middle
of the ring in order to maintain the same U-C bond distance. This improves
the spatial overlap between metal and ligand orbitals of 6 symmetry. Second,
there is a much better energy match between ligand 6 orbitals and metal d and
f orbitals. It is for these reasons that the metal f orbitals can make a
significant contribution to the U-(COT) covalent bond.
A molecular orbital scheme for uranocene is shown in Fig. 6.14. The e,
interaction is considerably larger than the e,,, reflecting the fact that
ligand/metal d overlap is more efficient than ligand/metal f. Two of the
uranium f orbitals have e;, (or 0) symmetry with respect to the metal-ligand
axis, and although orbitals of matching symmetry are available on the
ligands, the overlap is very small. These degenerate f orbitals thus remain
essentially non-bonding, and in 22 electron uranocene, each contains a single
electron. Uranocene is therefore best regarded as containing a [Rn]5f (U*)
metal centre.

Other actinidocenes
Compounds analogous to uranocene are known for all the actinides which
|Pa_| 21VE |paramagnetic_| have a stable tetravalent state (i.e. thorium, protactinium, neptunium, and
plutonium). They all appear to have the same structure as uranocene, and their
magnetic properties (Fig. 6.15) suggest that they can be described by the
same bonding model. For example, thoracene would be expected to have the
electronic structure e,,"e,,° (Fig. 6.14) and therefore be diamagnetic, as would
Fig. 6.15 Valence electron (VE)
counts and magnetic properties of
plutonocene (e>,"e3,.).
the actinidocenes [An(n°-COT).,].
Lanthanide COT complexes
Of the lanthanides, only Ce forms a neutral sandwich of this type,
[Ce(n*-COT),]. Cerocene was upheld for many years as a rare example of a
tetravalent cerium organometallic, these species being generally unstable
because of the strong oxidising power of Ce(IV). Recently, however,
theoretical calculations and physical measurements have shown that this
compound is better described as having a trivalent metal Ce centre [Xe]4f!
complexed by two COT'®* ions (the e,, 5 bonding orbitals in Fig. 6.14
contain only three electrons). In ionic notation we might write
(Ce $COT.}"}:
Neutral cerocene is readily reduced, for example by potassium metal, to
give what is undoubtedly a trivalent cerium compound (Eqn 6.4). This type of
sandwich anion can be prepared for most of the lanthanides directly by the
route in Eqn 6.5.

[Ce(n’-COT),] + K — K[Ce(n*-COT),] (6.4)

LnCl, + 2K,COT —> K[Ln(n*-COT),] + 3 KCl (6.5)

 The f elements 83

Unlike the actinidocenes, lanthanide COT complexes are regarded as

containing essentially ionic Ln—(COT) bonds. Perhaps the most convincing
demonstration of this is that they react immediately with UCI, to give
uranocene.

6.7 Exercises
1. Why are agostic interactions difficult to detect in f element
organometallic complexes?

2. Why do you think it is that the mixed cyclopentadienyl complexes

[LnCpCp*Cl] have not been isolated in a pure state?

3. Predict the stability of the bis(arene) complex [Lu(n°-C,H;Bu',),].

4. | Would you expect the An—(COT) bonds in thoracene and plutonocene to

be more or less ionic than those in uranocene?

Further reading
Cotton, S.A. (1991). Lanthanides and actinides. MacMillan Education,
London, UK.
Cox, P.A. (1987). The electronic structure and chemistry of solids. Oxford
University Press, Oxford, UK.
Kaltsoyannis, N. (1997). Relativistic effects in inorganic and organometallic
chemistry, Journal of the Chemical Society, Dalton Transactions, 1—
ily
Katz, J.J., Seaborg, G.T. and Morss, L.R. (eds) (1986). The chemistry of the
actinide elements, (2nd edition). Chapman and Hall, London, UK.
Kettle, S.F.A. (1996). Physical inorganic chemistry: a_ coordination
chemistry approach. Spektrum, Oxford, UK.
Index

actinide contraction 12 electrical conductivity in

actinide halides 36-37 lanthanide solids 41-44
actinide oxides 40-41 electrode potentials 11
actinide radioactive decay electronic absorption spectroscopy
processes 15 25-27
actinidocenes 79-82 electronic configurations 6
actinyl ions 48, 49, 51-56 exchange energy 10
adiabatic demagnetisation 31 extraction of lanthanides and
agostic interactions 70 naturally-occurring actinides 3
alkene complexes 76
alkoxides 60-64 f orbitals 6
alkyl complexes 69-70 fast breeder reactor 17
arene complexes 77-79 Fermi level 42
fluorescence of Ln** 28
B-elimination 69-70
band width 43-44 gadolinium break 57
bastnaesite 3 group valence 11, 36

carbonyl complexes 75-76 heart pacemakers 20

carnotite 4 heterogeneous catalysis 67
catalytic converters 38 Hubbard model 42
ceria 2, 38 Hund’s rules 23
cerocene 82 hypersensitive bands 26
chain reaction 15
charge transfer transitions 27 intimate mechanism 50
chelate and macrocyclic complexes ion exchange chromatography
56-60 57-58
Cooper pairs 40 ionic radii 12-14
critical mass 15 ionization energies 8-10
cyclooctatetraene complexes
79-83 jJ-j coupling 24
cyclopentadienyl complexes
70-75 lanthanide contraction 12
lanthanide halides 33-36
dialkylamides 60-64 lanthanide oxides 37-40
dinitrogen complexes 76-77 lanthanide shift reagents 59
discovery of lanthanides and lanthanides in organic chemistry
naturally-occurring actinides | 64-67
ligand exchange reactions 50-51
effective magnetic moment 30 Ln* in colour television sets 28
effective nuclear charge 12
electric dipole selection rules 25
 Index 85

magnetic resonance imaging 59 spin multiplicity 23

Manhattan project 16 stoichiometric mechanism 5]
metallic bonding 13, 43-44 subcritical mass 16
metallic radii 12-13 supercritical mass 15
monazite 3 synthesis of transuranium
elements 4
neodymium:YAG laser 29
nuclear fission 14 temperature independent
nuclear fuel reprocessing 18, 55 paramagnetism 31—32
nuclear reactor 16 term symbol 23
nuclear waste storage 18 thoria 2, 40
transactinide elements 13
occurrence of lanthanides and tributylphosphate 3, 18-19, 56
naturally-occurring actinides 3 tris(pyrazolyl)borate compounds
oxidation states 8-11] 29
tysonite (LaF;) structure 33
PuBr, structure 34
UCI, structure 34
radial nodes 9, 11 uraninite 4
rare earth 3 uranium isotope separation 16
relativistic effects 11-13, 21-23 uranocene 68, 80-82
Russell-Saunders coupling 23
vibronic coupling 25
smoke detectors 20
spin-only formula 31 water complexes 47-51
spin-orbit coupling 22
 DTD
pam
 "80S-b0S ‘6S ‘UOIeONpF /eiweyD Jo /euINOP ‘s\uaWa|a ay} JO 8/qe} DIPOUEd ay) UI
‘||BMOd PUB “D'M ‘SNijawe+ :8es SWajsAs Buljjaqe| 184yJO PUe BSA} JO UOISSNOSIP e 104 “BOWY UI BSN UOWWOD U! SEM (g S}UBWA/e UO!NISUBI} VY
sjuawaja dnoi6 urew) jas jsej ay, “adoinz ul pasn Ajjuanbad pue suOIFepUaWWOIe! DYqM| JapjO UOdN paseg si (JyBu g ‘Ya! Y) BS) JO JSuly BY “Ayeujique Ajaj}ajdwoo ase
UOIS (Z861) HM
g pue y suoyeubisap ay) ~abesn UoWWON UT JIS ING ‘HuIsnjuOD ave AAU} BdUIS a|qesISEP SSE] aIe SWa}SAS OM] JAYJO AY ‘UOIUBAUOD OWI JUGUNO ay} S! (1 —|) Wa}sAs
Buijjaqe] ‘Ouawnu ‘do au, “sjaqe; dnos5 au) Buipunowns uoisnjuod ajqesapisuod si aJay] ‘SedR|d /EWIDaP aesy} 0} PapuNo aie SaNjeA pajonb ay} ‘seoejd jeEwWIDEp 8844}
UE} 18}}8q 0} UMOUY SI ]UBIaM DIWO}e au] B1ayM SBSED U| ‘pajonb sanjeA au} aie aSay} Pue SAODUeEPUNGe |eljS919} DNSUaJOeIeYO BAeY OP fF] Puke ‘ed ‘Y] S}UBWAI/a 9914}
ay] ‘pejonb si edojosi uMOUY Pari|-}SebuO} au) Jo JyHiem aU) Ue SapI|ONu ajqeis OU aAeY SaSauyjUaIed UIYIIM PaUle]UOD SI JYDIAM JIWO}e BY} YOIYM JO} S}UBWS|F -SO}ON
(Leege) |(Lo16s2) (01 es2)|(ge0 zs2)|(e80 252)(80°12) (0z0'2rz2)(0z0'z¥2)|(L90'eve)|(v90've) BrOLZEZ GzO8EZ GEO LEZ BEd zEZ |Bz0LZz|
m1 | on | pw | wa | sa | yo | xa | wo | wy | ng | dN | A | ed | ut | OV |
01 | zor | tor | oor | 66 se | ze | 96 | s6 | +6 | | 2 | i¢ | 06 | ce |
LOSPLL POELL vEGBIL 9eL91 |E691 |OGZII SzeeSI | SZLSL S9ELSL ESI ee ve'ryl B060rL Sit Ori 906°8e!
my GA WL 4g OH AQ | ep | ng | ws | wg | pN | dd | 80 | &F |
LZ OZ 69 e9 | z9 | 99 | so | v9 | €9 | 2 | 19 | 09 | 66 | Bf | Zs
(zzz) | (zzz) |(ize) |(992) |(S92)| (292) | (992) | (e9z) | (192) | | szo'9zz |(0z0'€z2)
= - | ~ | W | SH | ua | 5S | aad | 3 ey = ld
2 Se e a e i ML zit | ter | ovr | a Gor |ee iO | | vol
gor | zor sal| gor | sor jeor-ee|
| | ee88 || 2848
(g10-zzz) (296-602) (zeeg0z)| oge's0z | zzoz | ese'roz 69002 |296961 BOGE | 22261 EOE LOZ 9BL Pees BPEOsL ErsLt | L2eLE1 | S06 ZEl
uy Ww Od 1g ad IL 6H ony Id 41 | «SO eH MM Bb) ae) eg so
98 s8 88 ep | ce | ie | 08 62 | 82 i en ee SS
“ezter voee2s o9zz1 zszizt |ores |sie'vtt |tives |egezor |2r901 |906201 |LOLOL /(L0626) y6'G6 90626 veZ16 |906'8B 2978 BONE
2X | el qS us ul PO—soBy Pd te! ny 9L (OW ON | 4Z | A IS qu
ere rece eS) as er hy |e | ee |Se le | ee | ee oy | ee | ew | ze |
ogee |p06 9682 ze6rL | 19ZL EZLED | GE'SD OPED eEd'AS |ELERs LyBSG |BEES D6F1S ZPEOS | BBLr 9SEry BOOP |BEDE
| 44 | 48 38S sv 39 eS uZ nd IN 09 34 uN 49 A ae | 9S 70) >|
9€ Gemye eee ee ite) we | ee) ee | ze | ee | se | ee | ee | ee | le | 6|
Bree EGrSE 990ZE |PLBDE 98082 |28692 soeve|06622
jV ID Ss a i & IV Bw eN
Eon a 2Sedee vh EE. a i e
ogi0z s66'8l 66651 LOOv! 11021 LI'L Z106 | 169
a Oo 4 ON | S| 8 aq
ou) ee ee ee ee § eis
£00'r 800 800°!
2H =H | H
zZ i aa
a i — , | ae
VIIA VIIA VIA | VA VAL | OVIII dll dl | GIA AINA | GIA ailA alA | aA Al Vil vi
GINA allA IA | aA gl alll dll dl VIIA VIIA VIIA VIIA VIA VA VAI VII Vi
yr a Dy A
(Sanjea LE6L DVdNI wos paydepe) syyHiem Diwoje JUBWA}a Pue S}UBWI]a BU} JO 9/qe} DIPOad
 l[e|xX
=.
FORD
_l@vJH EMIS TRY
_—a
__|.JREM ERS

SERIES EDITORS

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, the Oxford Chemistry Primers offer just the right level of detail for
undergraduate 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 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 felements is a short accessible text which covers much of the fundamental
chemistry of the lanthanide and actinide elements, including coordination
chemistry, solid state compounds, organometallic chemistry, electronic
spectroscopy, and magnetism. Comparisons are made between the chemistry of
the lanthanides and actinides and that of the transition elements, which is generally
much more familiar to undergraduate chemistry students.

The chemistry of the f elements is used as a vehicle for the communication of

important chemical concepts that are not usually discussed in detail in
undergraduate courses, for example the chemical consequences of relativity and
the lanthanide and actinide contractions. Modern applications of f element
chemistry, e.g.. the use of actinides in nuclear power generation, and of the
lanthanides in magnetic resonance imaging and catalytic converters in motor vehicle
exhausts, are also discussed in depth.
Nikolas Kaltsoyannis is Lecturer in Chemistry, University College London, and
Peter Scott is Lecturer in Chemistry, University of Warwick.

ISBN 0-19-850467-5
IO

OXFORD UNIVERSITY PRESS glgorggisoae72!

———