Atypical compounds of gases, which have been called ‘noble’{
Wojciech Grochala
ab
Received 12th April 2007
First published as an Advance Article on the web 20th June 2007
DOI: 10.1039/b702109g
In this critical review I describe fascinating experimental and theoretical advances in ‘noble gas’
chemistry during the last twenty years, and have taken a somewhat unexpected course since 2000.
I also highlight perspectives for further development in this field, including the prospective
synthesis of compounds containing as yet unknown Xe–element and element–Xe–element
bridging bonds, peroxide species containing Xe, adducts of XeF
2
with various metal fluorides,
Xe–element alloys, and novel pressure-stabilized covalently bound and host–guest compounds of
Xe. A substantial part of the essay is devoted to the—as yet experimentally unexplored—
behaviour of the compounds of Xe under high pressure. The blend of science, history, and
theoretical predictions, will be valued by inorganic and organic chemists, materials scientists, and
the community of theoretical and experimental high-pressure physicists and chemists
(151 references).
1 Lead-in
They say noblesse oblige.
1
Nothing more inaccurate and
misleading than that has been said, however, in the context of
the heavier gaseous elements of Group 18 (Rn, Xe, Kr), which
for a century have been colloquially called ‘noble’. Since the
discovery of argon (‘the lazy one’) in 1894 by Rayleigh and
Ramsay,
2
and the realisation of the monoatomic and unreactive
nature of its homologues, it was long believed that group 18
elements cannot form chemical compounds. And the very
concept of the stable ‘octet valence configuration’, so vital to
chemistry, was born.
3
‘Nothing can force a noble gas (Ng) atom
into a chemical bonding’ said this new law of nature. But human
nature loves challenges. Led by intuition and qualitative
considerations, von Antropoff
4a,b
and Pauling
4c
have forseen
that heavier Ng’s might, in fact, be chemically awaken in
powerfully oxidizing conditions while entering the oxide or
fluoride environment. Yet nearly forty years were to pass since
the apposite predictions saw confirmation in the breakthrough-
making 1962 discovery of the first Xe compound.
5,6
In 1989 Xe
has completely lost traces of its nobility and inertness when it
was turned into metal by use of ultra-high pressure.
7
About half a thousand compounds of Ng’s have been
synthesized since 1962.
8
Due to the above mentioned historical
reasons and for the purpose of this short essay, we would like
draw here the line between ‘usual’ and ‘unusual’ compounds. By
‘usual’ or ‘classical’, I refer to those compounds predicted by
Pauling, etc. so long ago. They contain fluoride, oxide and related
ligands (such as SbF
6
2
, TeOF
5
2
, IO
2
F
4
2
, WOF
4
, etc.) and
contain Xe at oxidation states +2, +4, +6, +8 (Fig. 1), or Kr
II
in
fluoride environment; chemistry of ‘classical’ Ng compounds—
also undergoing violent progress
9
—has been reviewed else-
where
10–12
and only rarely will be mentioned here.
13
Whatever
goes beyond that (and any chemical bonding involving Ar) will be
treated as ‘atypical’, and described from a chemical perspective to
a greater detail, with particular emphasis on chemistry of Xe.
2 The known Ng–element bonds
2.1 The first ‘unusual’ Ng–nonmetal bonds
Encouraged by fast enriching spectrum of the Xe–F and Xe–O
compounds, several groups have tried to break another
a
Laboratory of Technology of Novel Functional Materials,
Interdisciplinary Center for Mathematical and Computational Modeling,
University of Warsaw, Pawin
´
skiego 5a, 02106 Warsaw, Poland
b
Laboratory of Intermolecular Interactions, Department of Chemistry,
University of Warsaw, Pasteur 1, 02093 Warsaw, Poland
{ This essay is dedicated to Roald Hoffmann, on the occasion of his 70
birthday, who has taught me to feel the pulse of science.
Wojciech Grochala was born
in 1972 in Warsaw, Poland,
where he received his MSc,
PhD and DSc from the
University of Warsaw. He
has studi ed chemistry with
Roald Hoffmann in Ithaca
(Cornell, USA) and with
Peter P. Edwards (then at
Birmingham, UK). Inorganic
solid state chemistry and
m o u n t a i n e e r i n g a r e
Wojciech’s real passions. He
entered the field of noble gas
chemistry through the back
door, when studying the most
potent oxidizers available to chemistry (higher fluorides of
Ag).
144
His scientific interests encompass superconductivity,
145
quantum modelling of solids and molecular materials, hydro-
gen
37
and energy storage, hydrogen transfer cata lysis,
146
applications of high pressures in chemistry,
120
molecular devices,
unusual oxidation states of the chemical elements,
147
and more.
He now heads the Laboratory of Technology of Novel
Functional Materials, a joint enterprise of the ICM and of the
Department of Chemistry at Warsaw University.
Wojciech is the happy husband of Liliana and father of three
kids: Ilona Maria, Oliver Lucian and Leonard Karol.
Wojciech Grochala
CRITICAL REVIEW www.rsc.org/csr | Chemical Society Reviews
1632 |
Chem. Soc. Rev.
, 2007, 36, 1632–1655 This journal is
ß
The Royal Society of Chemistry 2007
paradigm, and to extend bonding to Xe beyond the two most
electronegative elements.
The first genuine Xe–N bond saw the daylight in 1974, with
the discovery of (F–Xe
II
)[
2
N(SO
2
F)
2
]. This new white
compound is isolable in gram quantities and is stable to
decomposition up to +70 uC (Table 1)! This synthesis paved
the road for further developments but these went uphill;
during the next 25 years about twenty other compounds have
been made in chemists’ flasks (including related
Xe
II
[N(SO
2
F)
2
2
]
2
, (RCN–XeF
+
)(AsF
6
2
) where R = H, Me,
derivatives of perfluorinated pyridine, perfluoroalkyls, tri-
fluorotriazine etc.),
14
Xe[N(SO
2
R)
2
2
](2,6-F
2
C
6
H
3
2
) where
R = F, CF
3
,
15
and [F
3
SN–XeF][AsF
6
].
16
Compounds of
[
2
(NH)–TeF
5
] and related anions
17
have also been made,
while some others were suggested as short-lived reaction
intermediates.
18
All of them were less thermally stable than the
native F–Xe
II
–N(SO
2
F)
2
.
As time passed by and the knowledge of researchers
accumulated, the first examples of the unstable Xe
VI
–N and
Xe
VIII
–N bonds were delivered.
19
Finally, Kr has also been
linked to N (and to O), but only below 260 uC (for O: 290 uC)
in the BrF
5
(for O: SO
2
ClF) solvent.
20,21
It was also quite difficult to get compounds which have real
Xe–C bonds. The precedents, [(F
5
C
6
)Xe
+
][B(C
6
H
5
)
3
F
2
] and
[(F
5
C
6
)Xe
+
][B(C
6
H
5
)F
3
2
] were synthesized as late as 1988 as
colorless solids, and their solutions in acetonitrile proved to be
stable at y0 uC.
22
Soon the preparation method was extended
and modified, and new synthetic paths have been invented
yielding more related aryl derivatives, the compounds of
(partially or entirely fluorinated) unsaturated hydrocarbon
groups (i.e. alkenyl
23
and alkynyl),
24
and even the first
compound with the simultaneous Xe–C and Xe–N bonds.
25
[(F
5
C
6
)Xe
+
](AsF
6
2
), which is so thermally stable that it
decomposes rapidly only well above its melting temperature
(T
m
= 102 uC), has become an important reagent in the
emerging organoxenon chemistry.
26
It took one decade to
synthesize the important siblings (F
5
C
6
)XeF, Xe(C
6
F
5
)
2
and
{[(F
5
C
6
)Xe]
2
F
+
}(AsF
6
2
).
27
The electrochemistry of
organoxenon(II) derivatives has been explored,
28
and even
the first organoxenon(IV) compounds have been prepared
([(F
5
C
6
)Xe
IV
F
2
+
](BF
4
2
) and Xe
IV
F
4
(CN
2
)).
29
The formation of the Xe
II
–Cl bond (an analogue of the well-
known Xe
II
–F bond) at temperatures close to ambient, was
first seen in 1997.
30
This new bond was brought to life in
two novel crystalline species, (F
5
C
6
)XeCl and
{[(F
5
C
6
)Xe]
2
Cl
+
}(AsF
6
2
), which showed reasonable kinetic
stability at ambient temperature. Since then only one more
Fig. 1 Classical chemistry of Xe involving connections to F and O
ligands at various oxidation states of a ‘‘noble’’ gas.
Table 1 Chemical formula and ranking of the thermal decomposition temperature, T
dec
, of selected compounds which exhibit the Xe–E bonds
(E = N, C, Cl)
Chemical formula T
dec
/uC Comments
F–Xe
II
–N(SO
2
F)
2
70 Bulk solid
(MeCN–XeF
+
)(AsF
6
2
) .210 Solution in anhydrous HF
Xe[N(SO
2
F)
2
]
2
.240 Solution in SO
2
ClF
[(F
5
C
6
)Xe[N(SO
2
F)
2
] 155 Bulk solid
[(F
5
C
6
)Xe[N(SO
2
CF
3
)
2
] 120 Bulk solid
[(F
5
C
6
)Xe(NC
5
H
3
F
2
-2,6)
+
](AsF
6
2
) ,20 Solution in MeCN
[(F
5
C
6
)Xe(NC–CH
3
)
+
](F
2
B(C
6
F
5
)
2
) 14 Bulk solid
[(F
5
C
6
)Xe
+
](AsF
6
2
) .130 Bulk solid; fast decomp. . 180 uC
[2,6-(F
2
H
3
C
6
)Xe
+
](BF
4
2
) 130 Bulk solid
[2,6-(F
2
H
3
C
6
)Xe
+
][C(SO
2
F)
3
) 113 Bulk solid
[(F
5
C
6
)Xe
+
](
2
OOC–C
6
F
5
) 85 Bulk solid
[(F
5
C
6
)Xe
+
][B(C
6
H
5
)F
3
2
] ,35 Solution in MeCN
[(F
5
C
6
)Xe
+
](BF
4
2
) .25 Solution in anhydrous HF
[(F
5
C
6
)Xe
+
](BF
4
2
) .25 Solution in anhydrous HF
(F
5
C
6
)
2
Xe 20 Bulk solid, explosive decomposition
[(CF
3
CMC)Xe
+
](BF
4
2
) 20 Stable for at least 2–3 h
[(F
5
C
6
)Xe
+
](HF
2
2
) ,20 Solution in MeCN
[(F
5
C
6
)Xe
+
][B(C
6
H
5
)
3
F
2
] .0 Solution in MeCN
[(4-ClH
4
C
6
)Xe
+
](BF
4
2
) .24 Bulk solid
[(F
5
C
6
)Xe
IV
F
2
+
](BF
4
2
) .220 Bulk solid
(F
5
C
6
)XeF ,278 Solution in CH
2
Cl
2
(F
5
C
6
)Xe(CN) ,278 Solution in CH
2
Cl
2
{[(F
5
C
6
)Xe]
2
Cl}(AsF
6
2
) 100 Bulk solid
(F
5
C
6
)XeCl 36 Bulk solid, decomposition with melting
(XeCl
+
)(Sb
2
F
11
2
) 220 Bulk solid, decomposition with melting
This journal is
ß
The Royal Society of Chemistry 2007
Chem. Soc. Rev.
, 2007, 36, 1632–1655 | 1633
compound, (XeCl
+
)(Sb
2
F
11
2
),
31
has been added to the list of
compounds isolable ‘in the flask’. The chemistry of the Xe–Cl
bond thus remains largely unexplored.
Summarizing this section, we note that close to a hundred
organoxenon, Xe–N and Xe–Cl compounds are known up to
date.
32
Xenon chemistry finally ‘went out from the cold’,
33
and
started flourishing.
2.2 Where are the bonds
Despite the 1974–2006 successes of xenon chemistry, only
y15% of all known solids containing short Xe–C, Xe–N and
Xe–Cl contacts have been structurally characterized. Selected
structures are shown in Fig. 2.
Xe
II
is isoelectronic to I
I
. Not surprisingly, XeF
2
is
isostructural with the linear symmetric IF
2
2
anion (present
for example in its Cs
+
salt), and all Xe
II
compounds show two-
coordinated Xe in an E
1
…
Xe
…
E
2
unit (E
1
, E
2
= Lewis base)
with its bond angle close to 180u (¡10u). Covalent Xe–C bond
lengths vary from 2.08 to 2.39 A
˚
, the Xe–N contacts are at
2.02–2.20 A
˚
, while the Xe–Cl bonds are 2.31–2.85 A
˚
long.
Even in the very limited set of experimental data, a clear
trend can be perceived, calling for an obvious generalization
(Table 2 and Fig. 3).
34
It is clear from Table 2, that for a given
E
1
…
Ng
…
E
2
set of elements (for example: Ng = Xe, E
1
= N,
E
2
= F), a relationship always holds: when R
1
becomes longer,
R
2
becomes shorter, and vice versa. Some will attribute it to the
varying basicity of one of molecule’s counterions, using the
following reasoning: since the (Sb
3
F
16
2
) ion is much less
basic than F
2
, the Xe
II
…
F contact must be longer in the
(Sb
3
F
16
2
)
…
Xe
II
–N(SO
2
F)
2
derivative than in the (F
2
)
…
Xe
II
–
N(SO
2
F)
2
one. In turn, the Xe
II
center must be more powerful
Lewis acid in the former compound, and therefore it is ligated
by the
2
N(SO
2
F)
2
base at shorter separation than that found
for the latter species.
Fig. 2 Structures of: (A) F–Xe
II
–N(SO
2
F)
2
; (B) Xe
II
–
N(SO
2
F)
2
(Sb
3
F
16
); (C) Xe(C
6
F
5
)
2
; (D) {[(F
5
C
6
)Xe]
2
Cl}(AsF
6
2
); (E)
(XeCl
+
)(Sb
2
F
11
2
). Xe – blue, F – light blue, Cl – green, C – black,
N – dark blue, S – yellow, O – red. Molecular units, extracted from
full crystal structures, are shown here, except for (D) where only one
C atom from each benzene ring is shown.
Table 2 Chemical formula and the lengths, R
1
and R
2
, of two hypervalent Xe–E bonds (E = N, C, Cl, F) and of two hypervalent Kr–F bonds
observed in several compounds of these Ng’s.
Formula R
1
/A
˚
R
2
/A
˚
Comment
F–Kr
II
–F Kr–F: 1.894 Kr–F: 1.894 Symmetric
(Kr
2
F
3
+
)(SbF
6
)?KrF
2
Kr–F: 1.805 Kr–F: 2.041
(Kr
2
F
3
+
)(SbF
6
)?KrF
2
Kr–F: 1.799 Kr–F: 2.065
(F–Kr
II
)(BiF
6
) Kr–F: 1.775 Kr–F: 2.090
(F–Kr
II
)(AsF
6
) Kr–F: 1.765 Kr–F: 2.131
(F–Kr
II
)(SbF
6
) Kr–F: 1.765 Kr–F: 2.140
(F–Kr
II
)(AuF
6
) Kr–F: 1.751 Kr–F: 2.161
(Sb
3
F
16
)–Xe
II
–N(SO
2
F)
2
Xe–N: 2.020 Xe–F: 2.457
F–Xe
II
–N(SO
2
F)
2
Xe–N: 2.200 Xe–F: 1.967
[(F
5
C
6
)Xe
II
(NCMe)](F
2
B(C
6
F
5
)
2
) Xe–N: 2.681 Xe–C: 2.092
[(F
5
C
6
)Xe(NC
5
H
3
F
2
-2,6)
+
](AsF
6
2
) Xe–N: 2.694 Xe–C: 2.087
[(F
5
C
6
)Xe
II
](AsF
6
) Xe–C: 2.079 Xe–F: 2.714
[(F
5
C
6
)Xe
II
](AsF
6
) Xe–C: 2.082 Xe–F: 2.672
[(2,6-F
2
H
3
C
6
)Xe
II
](BF
4
2
) Xe–C: 2.090 Xe–F: 2.793
[(F
5
C
6
)Xe(NC
5
H
3
F
2
-2,6)
+
](AsF
6
2
) Xe–C: 2.087 Xe–N: 2.694
[(F
5
C
6
)Xe
II
(NCMe)](F
2
B(C
6
F
5
)
2
) Xe–C: 2.092 Xe–N: 2.681
{[(F
5
C
6
)Xe]
2
Cl}(AsF
6
2
) Xe–C: 2.111 Xe–Cl: 2.847
{[(F
5
C
6
)Xe]
2
Cl}(AsF
6
2
) Xe–C: 2.116 Xe–Cl: 2.784
(F
5
C
6
)
2
Xe
II
Xe–C: 2.350 Xe–C: 2.394 Close to symmetric
[(2,6-F
2
H
3
C
6
)Xe
II
](
2
OSO
2
CF
3
) Xe–C: 2.079 Xe–O: 2.687
Xe–C: 2.092 Xe–O: 2.829
(XeCl
+
)(Sb
2
F
11
2
) Xe–Cl: 2.309 Xe–F: 2.644
XeCl
2
Xe–Cl: 2.542 Xe–Cl: 2.542 Symmetric (theor. result for solid)
145
{[(F
5
C
6
)Xe]
2
Cl}(AsF
6
2
) Xe–Cl: 2.784 Xe–C: 2.116
{[(F
5
C
6
)Xe]
2
Cl}(AsF
6
2
) Xe–Cl: 2.847 Xe–C: 2.111
1634 |
Chem. Soc. Rev.
, 2007, 36, 1632–1655 This journal is
ß
The Royal Society of Chemistry 2007
Others will argue that such asymmetric (long–short)
bonding pattern is typical for any hypervalent (here: four-
electron–three-center) bonding;
35
it is seen even in a purely
fluoride environment for Kr
II
(compare Fig. 2) and of course
for Xe
II
, as well as for many other related species.
36
Notably,
the hypervalent E
…
Xe–C, E
…
Xe–N and E
…
Xe–Cl groupings
also obey this general relationship.
Coming back to structures: Xe
IV
resembles I
III
(or Br
III
) and
Te
II
(or Se
II
), and these species usually appear in a square-
planar coordination, as seen for Xe
IV
F
4
, Br
III
F
4
2
ion or
L
2
Se
II
Cl
2
(L = tetrahydrothiophene). Distorted square-planar
coordination of Xe
IV
is also likely to be the case for
[(F
5
C
6
)Xe
IV
F
2
+
](FBF
3
2
). A very rare planar pentacoordinated
geometry Xe
IV
is presumably adopted be Xe
IV
in the
F
4
Xe(CN
2
) anion. These are the only organoxenon(IV)
compounds synthesized so far.
2.3 Combatting against auto–redox reactions
The chemistry of nonmetal–Xe connections is fascinating but
quite difficult to do. There are many reactions possible which
may hinder formation and limit stability of Xe–N, Xe–C, Xe–
Cl and related bonds, and, indeed, unwanted redox processes
are a major headache of researchers of Xe chemistry. Some
such processes are illustrated in eqn (1)–(3) for three
prototypical Xe–N, Xe–Cl and Xe–C compounds:
2 F–Xe
II
–N(SO
2
F)
2
A Xe
0
+ XeF
2
+ [N(SO
2
F)
2
]
2
(1)
(F
5
C
6
)Xe
II
Cl A C
6
F
5
Cl + Xe
0
(2)
(F
5
C
6
)
2
Xe
II
A (C
6
F
5
)
2
+ Xe
0
(3)
If one considers the limiting ionic Lewis formulas of the
native Xe
II
compounds, in which both ligands are negatively
charged, then processes described by eqn (1)–(3) are nothing
but ‘simple’ redox reactions: Xe
II
is (partially or completely)
reduced in a (formally) 2e
2
process, while two nonmetal–based
ligands L
2
are oxidized in two 1e
2
processes, and then they
form the L–L bond. Thus stability of nonmetal–Xe bonds is
essentially governed by the electron transfer reactions, just as
in the case of hydrides,
37
oxides, and many other families of
compounds with high oxidation states of chemical elements. In
practice, these and similar reactions are most troublesome in
the synthesis of new compounds, forcing chemists to carry out
their preparations at relatively low temperatures, down to
270 uC; once prepared, the Xe–N, Xe–C and Xe–Cl bonds
may be quite stable kinetically and thermally (Table 1).
The occurrence of redox reactions analogous to those of
eqn (1)–(3) (and the natural preference of Xe
II
for only
two coordinating Lewis bases) results in an inherent instability
of Xe bonds to elements that are any less electronegative
than those discussed above, such as Si, P, As, S, Se, Br and I.
Here, a spontaneous depopulation of the nonmetal’s
orbital (lone pair) takes place; electrons are transferred for
good to Xe
II
. Compounds which contain bonds between the
above-mentioned elements and Xe, have never been prepared
in large amounts in a chemist’s flask (despite numerous
attempts...), and their existence is limited to low concentra-
tions of molecules embedded in the noble gas matrices, at
very low temperatures. We will describe this exotic chemistry
in the present contribution, but first let us show several
representative examples of old reaction types for new
compounds.
2.4 Exploring new chemistry
It turns out that when an additional reducing agent (like
metallic mercury or C
6
F
5
I, which are poor reducers, indeed) or
an oxidizing one (like I
2
) are added to selected organoxenon
species, electron-transfer reactions proceed even easier at very
low temperatures (sometimes vigorously even below 280 uC!)
(eqn (4) and (5)), and with nearly quantitative yields:
(F
5
C
6
2
)
2
Xe
II
+ I
2
A 2 C
6
F
5
I + Xe
0
(4)
(F
5
C
6
)
2
Xe
II
+ Hg
0
A (C
6
F
5
)
2
Hg
II
+ Xe
0
(5)
[(F
5
C
6
2
)Xe
+
](BF
4
2
) + C
6
F
5
I A
[(F
5
C
6
2
)
2
I
+
](BF
4
2
) + Xe
0
(6)
More complex redox reactions (eqn (7)), are also possible:
[(F
5
C
6
2
)Xe
+
](AsF
6
2
) + C
6
H
5
CF
3
+ Me–CMN A
F
5
C
6
–C
6
H
4
(CF
3
) + Xe
0
+ (Me–CMNH
+
)(AsF
6
2
)
(7)
The majority of reactions involving Xe–nonmetal com-
pounds exemplify a ‘simple’ Lewis acid–base chemistry (eqn
(8)):
[(F
5
C
6
2
)Xe
+
](AsF
6
2
) + Me–CMN A
[(F
5
C
6
2
)Xe
+
(NMC–Me)](AsF
6
2
)
(8)
including metathetic reactions (eqn (9)):
2 (F
5
C
6
2
)Xe
II
(F
2
) + Cd(C
6
F
5
)
2
A
2 (F
5
C
6
2
)
2
Xe
II
+ CdF
2
,
(9)
Fig. 3 Illustration of the nature of a hypervalent bond: the
experimental R
1
vs. R
2
dependence for several F
2
…
Kr
II
…
F
2
species.
This journal is
ß
The Royal Society of Chemistry 2007
Chem. Soc. Rev.
, 2007, 36, 1632–1655 | 1635
but sometimes they may take much less expected course, like
in the following group-transfer and hydrogenation reactions
(eqn (10) and (11)):
[(cyclo-1,4-C
6
F
7
)Xe
+
](AsF
6
2
) + NaF A
(cyclo-F
6
C
5
)LCF
2
+ Xe
0
+ NaAsF
6
(10)
[(CF
3
CMC)Xe
+
](BF
4
2
) + 3 HF A
CF
3
CHLCHF + XeF
2
+ [HBF
4
]
(11)
New chemistry (and particularly when it’s originating from
the formation of new types of chemical bonds) is always great
experience...!
2.5 Is white stable? i.e. prospect for colourful compounds
All Xe–C and Xe–N compounds synthesized so far are white
or (seldomly) slightly yellowish. This indicates that the first
allowed electronic transition corresponds to UV or deep violet
radiation, and thus requires at least 3 eV. Only
(XeCl
+
)(Sb
2
F
11
2
) and [F
5
TeN(H)Xe][AsF
6
]
38
break this
monotony: they are orange, so should absorb in the blue
(y2.5 to 2.9 eV). Taking into account the products of the
thermal decomposition reactions for a variety of Xe–C and
Xe–N compounds, and applying chemical intuition to redox
reactions, one might anticipate that the lowest energy
electronic transition corresponds to the vertical (nonmetal-
to-Xe or Lewis base-to-acid) charge–transfer (CT) excitation.
Such an electronic transition would thus be a herald of the
‘full’ redox reaction which takes place if the temperature is
raised.
However, analysis of the electronic structure of Xe
II
derivatives (exemplified here by the quite stable, colourless
XeF
2
, Fig. 4 and Table 3) suggests that situation is more
complex. Calculations show three spin-forbidden transitions to
the triplet states and two transitions to singlet states (the
longer wavelength one is symmetry-forbidden) in the photon
energy range from 0 to 6 eV. Thus, XeF
2
does not get the
chance for colour: the first allowed transition is at 5.64 eV.
Nevertheless, the lowest energy transition to
3
P
g
is at energy as
small as 3.56 eV. This indicates that:
- thermal decomposition of XeF
2
to Xe and F
2
may occur
through the bending of an isolated molecule (p
u
mode),
through the coupling between HOMO21 (p
g
*) and LUMO
(s
u
*);
- the white color of XeF
2
is delusive, as the only potentially
colour-providing spin- and symmetry-allowed transition is well
above the lowest energy excitation (the one which may be
activated thermally, leading to decomposition).
For Kr
II
compounds, of course, the F
2
-to-Kr
II
charge
transfer process occurs much more easily than for Xe
II
(and
even in the fluoride environment): yet KrF
2
is still colourless.
39
Lack of colour is premonitory here, as it warns of the presence
of an ‘invisible’ singlet A triplet transition at the low energy of
2.9 eV. This excitation testifies to a significant fragility of the
thermodynamically unstable KrF
2
.
Fortunately, the large energies of the vertical CT
transitions indicate that barriers for the Xe–nonmetal bond
rupture (in a redox fashion) are not as small as one might
think. Indeed, as we have seen from Table 1, some Xe–
nonmetal compounds are surprisingly kinetically stable.
Hypothetically new bonds of even less electronegative elements
to Xe might treat our eye to a palette of colors, but they will be
more unstable thermally—and difficult to synthesize. Despite
this danger—and as long as events reside in the imaginative
minds and skillful hands of an excellent pedigree of synthetic
fluorine chemists—compounds containing new Xe–nonmetal
bonds, such as Xe–Br, Xe–S, Xe–P, Xe–Si, are just a matter of
time.
40
Fig. 4 Electron density integrated over (A) the first unoccupied, (B)
the uppermost occupied, band of solid XeF
2
(cell with Z = 2), which
correspond, respectively, to (C) s
u
* LUMO and (D) p
u
* HOMO of an
isolated XeF
2
molecule. Note, HOMO is doubly degenerate, and only
one orbital is shown here. (E) s
g
HOMO22 of XeF
2
molecule. Note,
HOMO22 is more F- and less Xe-based than LUMO, therefore the
allowed HOMO22 A LUMO (s
g
A s
u
*) electronic excitation has
F-to-Xe charge transfer character. DFT calculations for solid and
molecular XeF
2
.
Table 3 List of the lowest energy electronic excitations for the isolated XeF
2
molecule in its
1
S
g
ground state (B3LYP-TD results). The transition
energy, E, and oscillator strength, f, are shown, along with the predominant orbital contribution to the transition. All transitions to triplet states are
spin-forbidden (provided that heavy-atom effects are neglected), and many transitions to singlet states are symmetry-forbidden. HOMO and
HOMO21 of XeF
2
are p
u
, HOMO22 is s
g
, while HOMO23 and HOMO24 are p
g
; LUMO is s
u
.
E/eV f/1 Excited state Orbitals contributing E/eV f/1 Excited state Orbitals contributing
4.33 0
1
P
g
p
u
A s
u
* 3.56 0
3
P
g
p
u
A s
u
*
5.64 0.006
1
P
u
p
g
A s
u
* 3.78 0
3
S
u
s
g
A s
u
*
7.28 0
1
P
g
p
u
A s
u
* 5.00 0
3
P
u
p
g
A s
u
*
7.73 0.786
1
S
u
s
g
A s
u
* 6.40 0
3
P
g
a
a
Deeper occupier p
u
orbitals are involved.
1636 |
Chem. Soc. Rev.
, 2007, 36, 1632–1655 This journal is
ß
The Royal Society of Chemistry 2007
