close

Вход

Забыли?

вход по аккаунту

?

ESCA Electron Spectroscopy for Chemical Analysis.

код для вставкиСкачать
[30] a) F . Kgtle and A. H. Efler, Chem. Ber. 102, 3071 (1969): b)
F. Vdgtle, Tetrahedron Lett. 1968, 3623.
[31] R. W GrifJin jr., R. W Baughman, and C . E. Ramey, Tetrahedron
Lett. 1968. 5419.
[32] D.J . Cram and R . H. Bauer, J. Amer. Chem. SOC R l , 5971 (1959).
[33] R. Paionr and W Jenny, Helv. Chim. Acta 52, 2041 (1969).
[34] R. H. Mitchell and I.: Boekelheide, Tetrahedron Lett. 1970, 1197:
Chem. Commun. 1970, 1555.
[35] a) ( 5 ) was first prepared by W Baker, F. Glockling, and J . F. W
McOmie, J. Chem. SOC. 1951, 1118; b) R. W G r f j n jr. and N . Orr,
Tetrahedron Lett. 1969,4567.
[36] F. E g t l e , R . Schafer. L. Schunder, and P. Neurnann. Liebigs Ann.
Chem. 734, 102 (1970).
[37] W Jenny and R. Peter, Angew. Chem. 77, 1027 (1965); Angew.
Chem. internat. Edit. 4,979 (1965): R. Peter and W Jenny, Helv. Chim.
Acta 49, 2123 (1966).
[38] F . Egtle and H. A . Staab, Chem. Ber. 101,2709 (1968).
[39] F . Irdgtle. Liebigs Ann. Chem. 728, 17 (1969).
[40] F. Vdgtle and R. Lichtenthaler, Chemiker-Ztg. 94, 727 (1970).
[41] H . E. Winberg, F . S . Fawcert, W E. Mochel, and C. W Theobald,
J. Amer. Chem. SOC. 82,1428 (1960).
[42] H . H. Wasserman and D. 7: Bailey, Chem. Commun. (970, 107.
[43] B. Kamenar and C . K . Prout, J. Chem. SOC. 1965,4838.
[44] D.K . Lonsdale, H . J . Milledge, and K . l! Krishna Rao, Proc. Roy.
SOC. A 255, 82 (1960).
[45] a) W Baker, J . F. W McOmie, and J . M . Norman, Chem. Ind.
(London) 1950. 77; b) J. Chem. SOC1951, 1114.
[46] W Jenny and H . Holzrichter, Chimia 22, 247 (1968)
[47] W Baker, F. Glockling, and J . F . W McOmie, J. Chem. SOC.1951,
1118.
[48] Far fewer articles have been published on electrophilic substitutions on [2.2]metacyclophanes than on similar reactions on [2.2]paracyclophanes
[49] 7: Sato, M . Wakabayashi, Y Okamura, T Amada, and K . Hata.
Bull. Chem. SOC. Japan 40,2363 (1967).
[SO] D. J . Cram, C. S . Montgomery, and G . R. K n o x , J. Amer. Chem.
SOC. 88, 515 (1966).
[51] H . H . Wasserman and R. Kitzing, Tetrahedron Lett. 1969. 3341.
[52] V Boekelheide and J . A. Lawson, Chem. Commun. 1970, 1558.
[53] 7: Sato, E. Yamada, Y Okamura. T Amada, and K . Hata, Bull.
Chem. SOC. Japan 38.1049.2225 (1965).
[54] H. Blaschke and !l Boekelheide, J. Amer. Chem. SOC. 89. 2747
(1967).
[55] C. E. Ramey and l! Boekelheide, J. Amer. Chem. SOC. 92, 3681
(1970).
[56] R. H . Mitchell and 1.: Boekelheide, J. Amer. Chem. SOC.92, 3510
(1970).
[57] H.-R. Blattmann, D.Meuche, E. Heilhronner, R. J . Molyneux, and
1.: Boekelheide, J. Amer. Chem. SOC.87, 130 (1965).
[58] V Boekelheide and W Pepperdine, J. Amer. Chem. SOC. 92, 3684
(1970).
[59] J . Reiner and W Jenny. Helv. Chim. Acta 52, 1624 (1969).
[60] I.: Boekelheide and R . A . Hollins, J. Amer. Chem. SOC. 92, 3512
(1970).
ESCA : Electron Spectroscopy for Chemical Analysis[**]
By Carl Nordling‘’]
A substance on which X-rays fall emits photoelectrons and Auger electrons. The energy spectra
of the electrons emitted procide information about the electi.onic structure in the specimen, ranging
from the innermost atomic lenels and their dependence on the chemical enuironment to the molecular orbitals of the valence electrons and the band structure in solids. Electron spectra of this
nature can now be recorded with high-resolution instruments; their analysis offers new aspects
for investigation of chemical composition. The method of electron spectroscopy developed for this
purpose, which has now been developed to a high degree of perfection, will be referred to in the
following discussion as ESCA (Electron Spectroscopy for Chemical Analysis).
1. Introduction
For spectroscopic investigations on the structure of atoms
and molecules and on the dynamics of atomic processes,
one can either analyze the electromagnetic radiation
emitted or one can measure the kinetic energy of the electrons emitted as a result of various processes in the electron
clouds of the atoms. In certain cases, e.g. in NMR and
Mossbauer spectroscopy, the atomic nucleus is also used
as a source of information about the electron cloud. Whereas
[*] Prof. Dr. C. Nordling
Uppsala Universitet
Fysiska Institutionen
Box 530
S-75121 Uppsala 1 (Sweden)
[**I Based on a lecture to the “Solid State Chemistry’’ Group of the
GDCh on October 2, 1970, in Bonn.
Angew. Chem. internat. Edit. Vol. 11 (1972)
No. 2
the spectroscopy of electromagnetic radiation has been
widely used since Newton’s time, the behavior of the electrons liberated from the electron clouds has only rarely
been investigated.Thedeve1opment ofprecision instruments
for the analysis of electron spectra has nevertheless led to
very notable results during the past few years, which show
that a spectroscopy based on the direct observation of
electrons is a very valuable method for the study of atomic
and molecular systems and of solids. Electron spectroscopy
also yields information that is not obtainable by other
methods, and so offers numerous interesting possibilities
e. g. for applications in chemistry.
The following articIe is a report on the ESCA method
(Electron Spectroscopy for Chemical Analysis) and on the
results obtained with its aid by our research group at
Uppsala. ( A more detailed report covering the situation
up to 1969 may be found in refs. [1. 21.)
83
2. Excitation of the Electron Spectra
as well as the molecular orbitals in free molecules and the
energy bands in solids.
Electron spectra can be induced in various ways. The main
agent used for this purpose is characteristic X-rays, though
UV light and electron bombardment are also employed.
The energy analysis of the emitted electrons is carried out
in double-focusing spectrometers of the electrostatic or
Excitation with U V light has the advantage of a small
natural line width, but this process is restricted to the
outermost parts of the electron clouds.
3Detector
Electron bombardment can transfer the system under investigation into a highly excited state, which then decays
with emission of an electron. This process, which is known
as autoionization, thus produces a positively charged ion
and an electron, whose kinetic energy is given by the formula
In cases where the electron bombardment itself leads to
ionized initial state, the system becomes doubly ionized
on decay. This transition from a singly ionized to a doubly
ionized state with emission of an electron is known as the
Auger effect, and the energy of the emitted Auger electron
is given by
iui
spectrometer
Fig. 1. Methods of exciting electron spectra for recording in highresolution instruments.
magnetic types. The experimental arrangement is shown
schematically in Figure 1, with an indication of the various
modes of excitation. One of the magnetic instruments, with
the ancillary electronic equipment, is shown in Figure 2.
Auger electrons can be liberated by high-energy light
quanta. When X-rays act on a specimen, therefore, photoelectrons and Auger electrons are obtained together.
Fig. 2.View of an ESCA instrument
In photoionization, the kinetic energy of the photoelectrons
is equal to the difference between the energy hv of the
stimulating light quantum and the binding energy (ionization energy) E , of the liberated electrons.
Ekj, = h v
-
Eb
(1)
X-rays of sufficiently high energy can liberate electrons
from all levels, i. e. the inner atomic shells can be investigated
84
3. Chemical Shifts in the Spectra of the Inner
Electrons
The binding energies E , are not purely atomic quantities,
but change slightly with the chemical environment. This
property is manifested in the kinetic energies of the emitted
electrons; it can be detected in the electron spectra by
modern high-resolution instruments.
Angew. Chern. interncir. Edir. / Vol. I 1 (1972) No. 2
Figure 3 shows an ESCA spectrum of N-(Z-pyridyl)-pnitrobenzenesulfonamide. On excitation of the sample
with MgKzradiation, photoelectrons were emitted, inter
alia, from the K shells of the nitrogen atoms. Each of the
three nitrogen atoms in the molecule has a different
chemical environment ;three different lines therefore appear
in the electron spectrum.
1500
-*
.
0
N 1000
u
Y
I
.-
I
n
c
a
J
+
c
500
\”
Nls
Electron binding energies and chemical shifts can be
obtained more easily from Hartree-Fock calculations with
the aid of Koopmans’ theorem. In this case the total energy
of the neutral and the ionized systems is not calculated, but
the binding energies are equated to the individual orbital
energies for the neutral system; thus it is additionally
assumed that the other electrons in the ion can be described
by the same orbital wave functions as in the neutral molecule. The binding energies obtained in this way are systematically 10-20 eV higher than those found experimentally. They are also strongly dependent on the magnitude of the basis set used to describe the wave functions
and on its optimization. However, they can be used for a
comparison of theoretically calculated shifts in the binding
energies and those found experimentally. Systematic
calculations based on Koopmans‘ theorem are now available for the chemically induced shifts of the energies of the
inner electrons; we have correlated them with the shifts
observed experimentally121.
Figure 4 shows a comparison of the measured chemical
shifts for the I s level of carbon in some small molecules
with the orbital energy shifts from nb initio MO-LCAOSCF calculations13J. The method of least squares gave a
straight line having a slope of 1.09. It was found in this case
that the agreement of the calculated and the experimental
shifts improves, i.e. the slope of the line becomes closer
to 1, with increasing proximity to the Hartree-Fock limit.
0.
O
?
0
I
I
840
850
00
0-0 0
8
-
Kinetic energy [eV] +
eV
I
I
410
-Binding
LOO
energy [ell
Fig. 3. ESCA spectrum for nitrogen (the I s electrons) in N-(2-pyridyl)-pnitrobenzenesulfonam~de.
It is important to note that the electrons involved here are
not the valence electrons, i. e. the electrons that are actually
involved in the chemical bonding, but the electrons of the
innermost parts of the electron cloud, which are usually
“forgotten” by the chemist.
3.1. Calculation of Chemical Shifts
L
I
0
The binding energies E , of the inner electrons in free molecules and their chemical shifts can be determined very
accurately from quantum-mechanical calculations. In the
best possible approximation method of the ub initio type
for the solution of the Schrodinger equation, i.e. the
Hartree-Fock (SCF) approximation, the binding energy
is obtained as the difference between the total energies for
the neutral molecule and the molecule-ion with a vacancy
in the inner electrons. The shifts calculated so far by this
method correspond to an accuracy of a few tenths of an
electron volt, even though correlation effects and relativistic
effects were disregarded. However, since the latter are
independent of changes in the chemical environment in the
case of the inner electrons, they do not affect the calculated
chemical shifts.
I
r
-
5
10
Experimental chemical shift
eV
Fig. 4. Correlation between calculated orbital energy shifts for the 1 s
level of carbon in some small molecules and the experimentally determined shift.
3.2. The Potential Model
A simplification of the theoretical calculation of shifts in
the binding energy can be achieved with an electrostatic
potential model[’.’’. In this model, the shifts in the binding
energy are correlated with the electron distribution in the
neutral molecule. The model is entirely classical, but can
also be described and applied in quantum-mechanical
terms.
85
In the potential model, the chemical shift for the energy E ,
of an inner electron is determined by a change in the potential of the inner electrons. This potential is obtained as the
superposition of two potentials, one due to the electron
distribution in the atom in question and one due to the
charge distribution in the remainder of the molecule. The
latter contribution, which we call the molecular potential,
is easy to calculate if the charges of the other atoms in the
molecule are imagined to be concentrated in point charges
q i (see Fig. 5). The contribution to the shift that is due to
the charge distribution in the atom under consideration
can be mainly attributed to the valence electrons. If we
assume that these form an electron shell of radius r, and
if the effective charge q of the atom is placed in this shell,
the corresponding contribution to the potential is given
by q/r. A theoretically better estimate of this contribution
is given by the electrostatic interaction integral J between
the inner electron orbital under examination and a valence
electron orbital in the same atom. The expression for the
chemical shift in the energy level of the inner electron then
becomes :
AE = J q
+V+
The charges obtained from the ab initio wave functions are
Mulliken’s “gross atomic charges”. The constants J and 1
were determined in both cases by fitting the least square
of E - I/ to J q + 1; the values for J , 18.3 eV from the ab
initio calculations and 23.5 eV from the CND0/2 calculations, agree relatively well with the repulsion integrals J
calculated for the Is-2p electrons in a carbon atom.
,
1
0
where the molecular potential is given by
lr(85BsI
5
10
Calculated Shift 118.3q+V+301
eV
Fig. 6. Correlation between measured chemical shifts for the 1 s level of
carbon and the shifts calculated from the potential model, the charges
being obtained from ab initio calculations
and the constant 1 depends on the choice of the reference
level. It is found that J is approximately equal to the expectation value ( l / r ) for a valence electron.
Figures 6 and 7 illustrate the correlation between experimental chemical shifts for the I s level of carbon and shifts
calculated from the potential model with ab initio wave
functions and with C N D 0 / 2 wave functions re~pectively[~!
q1
I
I
I1‘
I
0
i
;‘3
I
II
I
I
4 q3
AE ( l s l = q / r + E q i / r i + l
Fig. 5. In the potential model, the chemical shift of E, for an inner
electron (e.g. 1s) is divided into a contribution from the interaction with
the valence electrons of the atom in question and a contribution from
the other atoms of the molecule. If the valence electrons of the atom in
question are assumed to be concentrated in a spherical shell of radius r
and having a net charge q, the first contribution is 4 / r . If the (net) charges
of the other atoms are regarded as point charges q , at distances r, from
the nucleus of the atom under Investigation, the second contribution,
i. e. the molecular potential, is Z ~ J ’ ; .
86
5
10
Calculated shift I235q+V+O 221-
eV
Fig. 7. Correlation between measured chemical shifts for the 1 s level of
carbon and shifts calculated from the potential model, the charges being
obtained from CNDO/2 calculations. (For the assignment of the numbers, cf. [3].)
It can be seen with the aid of the simple potential model
that the binding energy E , of an inner electron increases
(and the electron thus becomes increasingly difficult to
remove from the electron cloud) with increasing positive
charge on the atom. This can be demonstrated by the ESCA
spectrum for carbon and oxygen in acetone (Fig. 8). The
most strongly positive carbonyl C atom has the highest
binding energy (left-hand line). The methyl C atoms (rightAngew. Chem. internat. Edil. 1 Vol. I 1 (1972) 1 No. 2
hand line) are more nearly neutral. It can also be seen that
the intensities of the lines reflect the relative abundances of
the two different types of carbon atoms in the molecule.
The intensities of the electron lines are proportional to the
abundances of the elements, and can therefore be used for
a quantitative elemental analysis, while information on the
valences of the elements is obtained simultaneously.
The carbon atoms in ethyl trifluoroacetate (Fig. 9) carry
substituents having different electronegativities. The increasingly positive charge is clear from the positions of the
lines in the ESCA spectrum of the compound. The lines,
in order of excitation energy, are due to the four C atoms
of the molecule in accordance with the notation used in
Figure 9.
CIS
3.3. Charge Correlations
When ESCA is applied to chemical problems, it is often
desirable to be able to interpret the chemical shifts in simple
language that is understandable to any chemist. This has
been found possible with the aid of the electronegativity
concept. We have empirically established correlations between measured chemical shifts and an atomic charge
parameter qp. This charge parameter is obtained as the
sum of the formal charges on the atom and the partial ionic
characters of its bonds with neighboring atoms:
4,
=
charge
+
Q
ZI
formal contribution of
charge ionic character
energy CeVl
-Binding
Fig. 8. ESCA spectrum for the 1 s electrons of carbon and of oxygen in
acetone.
The ionic bonding component, according to Pauling, can
be expressed as a function of the electronegativity difference
between the atoms of a bond.
I,,
=
1 - exp[-0.25(~,
- xB)’]
The procedure can be illustrated e. g. by the calculation of
the charge on the nitrogen in an alkylammonium ion, which
carries a positive formal charge :
Q
/
4, = 1.00
b
\
c Is
0
I
1190
1195
Kinetic energy [eV]-
I
295
I
290
285
Binding energy [eVl
Fig. 9. ESCA spectrum for the 1 s electrons of carbon in ethyl trifluoroacetate.
Angew. Chem. iniernat. Edit. j Val. I 1 (1972) No. 2
1.oo
\
+ 2(-0.30) + 2(-0.15)
=
f0.10
In this way it was possible to correlate shifts found in the
ESCA spectra for a large number of compounds having
very different structures with a calculated charge parameter
qp. Figure 10 shows the result for a series of sulfur groups,
based on measurements of the 2p energy for about 100
sulfur compounds. Similar correlations are found for other
elements, and this confirms the hypothesis of the chargedependence of the binding energies. They have proved to
be particularly useful in the application ofthe ESCA method.
Some examples of the application of ESCA to special
problems in chemistry are described in the following section.
87
eV
I
I
I
I
I
I
0
t
-s-s-
-s-s4
.1+
00
0
Disulfoxide S-Thiosulfonate
Analysis of the IR and NMR spectra had yielded no definite
answer regarding the occurrence of these isomers. The
ESCA spectrum of the dioxide clearly shows, on the other
hand, that the compound has a thiosulfonate structure
(Fig. 11).
ci2oos
2500
-05
0.0
-
05
1.0
15
Charge l q p l
20
2.5
2wo
Fig. 10. Correlation of the ESCA shifts for the 2 p level of sulfur in
various sulfur groups with a charge parameter calculated as the sum
of the partial ionic contributions in the bonds. (For assignment of the
numbers, cf. [4].)
6I
I
c
1500
C/LO s
c
4. Selected Examples
v)
-:2000
S2pIAI,,l
0
4.1. Polarity of Chemical Bonds
bl
The charge-dependence of the shifts makes electron spectroscopy particularly suitable for the study of the polarity
of chemical bonds. It is possible in this way to investigate
the polarity without knowing the bond lengths and bond
angles.
S-0 bonds in many sulfur compounds have been intensively studied[4]. The polarity has been found to be fairly
structure-dependent for sulfoxides. Its average value lies
roughly between the values calculated on the basis ofmodel
substances and charge correlations for the two limiting
structures.
A comparison with the carbonyl group shows that the
sulfinyl group has approximately the same polarity. These
experimental results help to answer the question whether
sulfur expands its valence shell by making use of d orbitals.
4.2. Oxidation of Cystine
The example of doubly oxidized cystine can be used to show
how the chemical shifts in the ESCA spectra can help to
solve structural problems[51.The cystine molecule contains
two equivalent sulfur atoms and two possible structures
HOOC - CH-CH, - S-S - CH,
I
N H2
- YH-COOH
N Hz
had to be considered for the dioxide, i.e. disulfoxide and
S-thiosulfonate.
88
1500
0
I
0
I
1000
1310
1320
1315
1325
Kinetic energy IeVI-
,
170
-Binding
I
165
160
energy lev1
Fig. 11. Structure determination by ESCA. A comparison of the electron
spectra from the 2 p level of sulfur in cystine .$,.$-dioxide (a) and in
unoxidized cystine (b) shows that the dioxide has a thiosulfonate structure.
Instead of a single line, as in the case of unoxidized cystine,
two lines are observed for the 2 p level of sulfur, one of these
showing no shift and the other a shift of 4.0 eV to a more
highly oxidized state. The disulfoxide would have given
only one line, with the relative intensity doubled and with
a shift of = 2 eV.
More complicated structural problems have also been
elucidated by the ESCA technique, and it is likely that the
further development of the instruments used will lead to
increasingly valuable conclusions.
4.3. Valence Problems
The problem of the electron transfer between the metal and
the carbon in transition metal carbides was disputed for
a long time. For lack of reliable experimental data, no
decision could be reached in favor of any of the various
theories proposed. ESCA measurements of the energy
Angew. Chem. internal. Edit.
Val. I 1 (1972)
No. 2
shifts for the inner electrons in such compounds and other
model substances have now shown that an electron transfer
takes place from the metal to the carbon in these carbides
(Fig. 12)r61.
1
1
c/60 s
Cls
#$,Ti
1
complexes in which the metal is present in a low formal
oxidation state[s1. On the basis of the ESCA data, it was
possible to assign relative oxidation states to the central
metal, and the investigation confirmed that the coordination
of “neutral” ligands can lead to a considerable charge
CllOS transfer from the metal to the ligands. Figure 14 shows the
position of the 4f7,, level of platinum in such complexes.
r
metal
..
-Binding
energy
kV1
Fig. 12. ESCA spectra of T i c and Ti metal. The line shifts yield information on the charge transfer between the metal and the carbon in the
carbide.
In an investigation of the reduced ternary phases in the
W-V-0
system, the problem of the oxidation state of the
tungsten and vanadium atoms was examined. This is
associated with the question of the order or disorder of the
two types ofmetal atoms in the crystal structure. The ESCA
spectra showed the following distribution of the oxidation
states in W,V,O,, and WV,O,. which had already been
:
suggested on the basis of various
(Ph3 PJ2PtiPhCzPhl
i Ph3PIz Pt I P Ph3)2
74
73
c- Binding
72
energy kV1
71
Fig. 14. Chemical shiftsfor the4f,,, level ofplatinum insomecompiexes.
5. Electron Spectra of Free Molecules and Solids
Figure 13 shows the spectra of vanadium (2p,,,) and of
tungsten (4f5,,.,,,) in W,V,O,,. The positions of the lines
obtained with model substances for various oxidation
numbers are indicated. It can be seen that tungsten is
present exclusively in the hexavalent state, whereas
vanadium is present both in the tetravalent and in the
pentavalent states.
c12oos
,..6oS
t
1500
3000
c
2500
a,
1000
A
c
j00
730
735
”
1210
Kinetic energy lev]
-
1215
Fig. 13. I t can be seen from this ESCA spectrum of W,V,O,
that tungsten is present as Wvl and vanadium in two oxidation states, i.e. as
V”’ and Vv.
In the study of many catalytic reactions, the main point of
interest is the bonding of the metal in organometallic
complexes. We have recently investigated some platinum
Angew. Chem. inrernar. Edir. 1 Vol I 1 (19721 1 No. 2
As was mentioned in Section 2, if X-rays are used to excite
the electron spectra, the behavior of both the inner electrons
and the valence electrons can be investigated. Figure 15
shows an ESCA spectrum of gaseous tetrafluoromethane
stimulated with Mg,, radiation. The two lines for binding
energies of 696 and 302 eV correspond to the I s electrons
of the fluorine and of the carbon respectively. The K-Auger
\pectrum of fluorine is observed between kinetic energies
of 630 and 660 eV. Molecular orbitals of the valence electrons, which are formed mainly from the atomic 2s and 2 p
orbitals of the carbon and the fluorine in the CF, molecule,
‘ire “seen” in the range of binding energies below 50eV
(kinetic energy 1200-1254 eV). The 32 valence electrons
of CF, are distributed over 16 molecular orbitals, but
because of the high symmetry of the molecule, several of
the orbitals are degenerate, so that only 7 ionization energies
can be distinguished in the spectrum of the valence electrons. Three of these states were investigated earlier on the
basis of UV-excited spectra ;they can be essentially assigned
to non-bonding molecular orbitals.
For a deeper understanding of the chemical bonding, it is
necessary also to study the bonding orbitals. These are
shown in Figure 15. We used semiempirical (CNDO) and
ab inirio calculations of the valence electron configuration
and of the atomic parentage of the molecular orbitals for
the interpretation of the M O spectra. A property of the
X-ray excited electron spectra that is particularly useful
89
in this connection is that the cross section for photoemission
from a given orbital is strongly dependent on the symmetry
of the orbital.
i
Air
01s
Nls
Ar 2p
An interesting line splitting is observed in the spectra of the
ZlOOG
*
800 -
inner electrons of paramagnetic molecules. This splitting
is shown in Figure 16 for molecular oxygen in air. The
~
c)
2
5 600
-
- LOO 0)
5
200
-
0-
715“840
8L5
1000 1005
Kinetic energy rev1 +
710
”
Fig. 16. ESCA spectrum of oxygen, nitrogen, and argon In air. For
oxygen, which is paramagnetic, one observes a splitting of the 1s line
into two components.
I
650
640
695$
-
It-LL-cI
950 955
1210 1220 1230 12LO
Kinetic enerav [eV]
{&$
-Binding
I
305 300
energy [eVl
LO
I
20
30
10
Fig 15. ESCA spectrum of CF,. This spectrum shows both the inner
electron levels F l s and C l s and the energy levels for the molecular
orbitals of the valence electrons. It also shows the K-Auger electrons
of fluorine.
spectrum contains the I s lines of the oxygen and of the
nitrogen in the air, as well as the 2 p spin doublet of argon,
which had a partial pressure of
torr in this case. The
1slineofoxygenissplit into twocomponents, withadistance
of 1.1 eV between the lines and with an intensity ratio of
1 :2. This splitting can be explained as follows. On photoemission of an electron from the 1s level, it is necessary to
take into account the interaction between a I s state with
spin i and the M O state ( 7 ~ ~of2 oxygen,
~)
which is responsible for its paramagnetic properties.
1
5L5
-Binding
540
energy IeVl
Fig. 17. 1s lines of oxygen in 0, and H,O. The line splitting observed
for the paramagnetic oxygen molecule disappears when the oxygen
is chemically bonded in water vapor. This is accompanied by a considerable chemical shift of the 1s level.
Ar
18000
4
n-3
3s’ 3PS(?p3A
3s’3p6nd1
3Sz3P5(?Pin)
5
4
r
6 7 8
1
n -3
5
4
I
,
I
I
00
I
6 7 8
co
1
I
-.T
16000
.
v)
c)
ID
u
0
12000
c
.-
-VI
P)
c
c
8000
3s’3p6np’-
1
n =4
,
3s2 3p5(2P1,,)
3 s ’ 3p6np’-3sz
3p5(?p3/2)
,
1
5
6
10 5
110
115
1
12 5
120
Kinetic energy
[eVI
8 910
I
I
1
5
n=L
1 ,
7
6
7
130
,
I
8 910
13 5
Fig. 18. Autoionization electron spectrum of argon, excited by electron bombardment.
90
Angew. Chem. internat. Edit. I Val. I1 (1972) 1 No. 2
co
C-Auger
bI
5000 -
rn
rrm
30 20282L2221)
3000 3b
2000
272
270
268
12
’-LJ
0
I
2 50
260
Kinetic energy [etll
-
I
270
Fig. 19. Auger and autoionization electron spectrum of CO with vibrational structure.
The resultant spin of the two electrons with parallel spin
in this orbital is 1. The total spin of the molecule after the
emission of a Is electron will therefore be $ or $, i . ~ the
.
molecule passes into a doublet or quadruplet state. The
intensity ratio of the lines thus becomes 2:4, and the calculated magnitude of the energy splitting agrees with that
in the electron spectrum (Fig. 16).
c / ms
Be IK-emission11
0 I K-emission1
I
I
al
E 2000
4
al
-
-
This “spin (or exchange) splitting” disappears when the
oxygen is chemically bonded to other atoms in diamagnetic
molecules, as can be seen from the spectra of 0, and H,O
(Fig. 17). As discussed earlier, it is replaced by a “chemical
shift”, which has a magnitude of 3.5 eV in this example
between (the quadruplet state of) molecular oxygen and
the oxygen bound in H,O.
The autoionization and Auger electron spectra (cf. Section
2) offree atoms and molecules have not yet been investigated
very extensively. We have obtained highly resolved spectra
of the inert gases and ofmolecules by electron bombardment
excitation. These spectra contain very large numbers of
lines, and can be compared in part with transitions in UV
spectra. The autoionization spectrum of argon (Fig. 18)
contains several Rydberg series with a series limit that
corresponds to the complete separation of a 3 s electron in
the initial state. The Auger spectrum and the autoionization
spectrum of carbon in the CO molecule are shown in Figure
19. The left-hand part of the spectrum is the Auger portion,
which shows a pronounced vibration structure in the region
around 253 eV. Further to the right in the spectrum, one
finds a number of groups of autoionization lines with
clearly resolved vibrational components.
We have also extended the scope of the ESCA technique to
investigations on solids. Thus the angular distribution of
elastically scattered electrons ejected by Mg,, radiation
from various levels in an NaCl crystal has recently been
Angew. Chem. internar. Edit. I Voi. I 1 (1972) N o . 2
2000“
1000
- 1000
5LO
11858201
530
520
h
120
-Binding
20
30
110
10
0
energy [eVI
Fig. 20. ESCA spectrum of BeO, excited with MgK,. This spectrum
sho’ws both the inner electron levels and the valence bands of the oxide.
A comparison with X-ray emission data is shown in the upper part of
the figure.
-10 -5
0
-10 -5 0
-10 -5
0
-10 -5
0 ell
Fig. 21. Band spectra of the transition metals of the first, second, and
third series.
91
investigated[’’, and the “escape depth” for electrons liberated from a metal by the action ofx-rays has been determined
The valence bands in LiF, BeO, BN,
and graphite have also been studied, and the results compared with X-ray spectroscopic data“ ‘I; see Figure 20.
We have also studied the valence bands in transition metals
and other solids“**‘I. Figure 21 shows electron spectra
from the valence bands of twelve transition metals. After
certain corrections, these ESCA spectra provide a very
good description of the energy distribution in the bands.
It is my pleasure to thank Professor Kai Siegbahn and all the
other members of our research group in Uppsala for their
contribution to the work that led to the results presented here.
Received: October 28.1970 [A 858 IE]
German version: Angew. Chem. 84,144 (1972)
Translated by Express Translation Service, London
[I] K . Siegbahn, C . Nordling, A . Fahlman, R . Nordberg, K . Hamrin,
J . Hedman, G. Johansson, T Bergmark, S.-E. Karlsson, I . Lindgren, and
B. Lindberg: ESCA: Atomic, Molecular and Solid State Structure
Studied by Means of Electron Spectroscopy. Nova Acta Regiae SOC.
sci. Upsaliensis, Ser. IV, Vol. 20 (1967).
[2] K . Siegbahn, C . Nordling, G . Johansson, J . Hedman, P. F. Heden,
K . Hamrin, U . Gelius, T Bergmark, L . 0. Werme, R . Manne, and Y Baer:
ESCA Applied to Free Molecules. North-Holland, Amsterdam 1969.
[ 3 ] U . Gelius, P. F. Heden, J . Hedman, B. J . Lindberg, R. Manne, R.
Nordberg, C . Nordling, and K . Siegbahn, Physica Scripta 2, 70 (1970).
[4] B. Lindberg, K . Hamrin, G. Johansson, U . Gelius, A . Fahlman,
C . Nordling, and K . Siegbahn, Physica Scripta I , 286 (1970).
[5] G . Axelson, K . Hamrin, A . Fahlman, C . Nordling, and B. J . Lindberg,
Spectrochim. Acta 23, 2015 (1967).
[6] L. Ramqvist, K . Hamrm, G. Johansson, A . Fahlman, and C . Nordling,
J. Phys. Chem. Solids 30, 1835 (1969).
[7] K . Hamrin, C. Nordlmg, and L . Kihlborg, Ann. Acad. Reg. Sci.
Upsaliensis 14 (1970).
[8] C . D. Cook, K . Y Wan, U . Gelius, K . Hamrin, G . Johansson, E . Olson,
H . Siegbahn, C . Nordling, and K . Siegbahn, J. Amer. Chem. SOC.93,
1904 (1971).
[9J I<. Siegbuhn, U . Gelius, H. Siegbahn, and E. Olson, Physica Scripta
I , 272 (1970).
[lo] Y Burr, P F . Heden, J . Hedman, M . Klasson, and C . Nordling,
Solid State Commun. 8, 1479 (1970).
[ I l l K . Hamrin. G . Johansson, U . Gelius, C . Nordling, and K . Siegbahn,
Physica Scripta I , 277 (1970).
[I21 E Baer, P. F. H e d h , M . Klasson, C . Nordling, and K . Siegbahn,
Physica Scripta I , 55 (1970).
The Electronic Properties of Diradicals
By Lionel Salem and Colin Ronland‘‘l
A review of the various possible definitions of diradicals leads the authors to describe these systems
as having two odd electrons in degenerate or nearly-degenerate molecular orbitals. A study of
the wave-function for the two odd electrons shows that its form depends entirely on whether the
diradical is homo- or heterosymmetric. Energy schemes are given in these two cases, as well as in
the intermediate %on-symmetric” case. The extent of zwitterionic character in diradical states is
also investigated. This is followed by a discussion of intersystem crossing between singlet and
triplet diradical states via spin-orbit coupling and other mechanisms. The electronic matrix elements for spin-orbit coupling are calculated and evaluated numerically for various model cases.
It is then possible to establish general rules for favorable (electronic) intersystem crossing. I n
1,3 or 1,4 diradicals its efficiency is estimated to be comparable with that in aromatics. The role
of the electron-nuclear hyperfine interaction in mixing singlet and triplet states, particularly in
CIDNP, is explained.
Finally the question of whether diradicals actually occur as secondary minima on potential
energy surfaces is examined. Recent quantum-mechanical calculations, in contradiction to some
thermochemical and kinetic ecidence, lead to flat singlet surfaces without significant minima.
1. Introduction: What is a Diradical?
The purpose of this article is to study the qualitative and
quantitative properties of those molecular species which
Organic Chemists call diradicals. The recent progress in
[*] Prof. Dr. L. Salem and Dr C. Rowland
Laboratoire de Chimie Theorique
Centre Scientifique d’Orsay
Universite de Paris-Sud, 91-Orsay (France)
The Laboratoire de Chimie Theorique IS also part of the Laboratoire
de Physico-Chimie des Rayonnements, associated with the CNRS.
92
the understanding of concerted reactions“] has been accompanied by an increased interest in the Chemistry of species
with non-concerted behavior‘’]. Although, as we shall see,
it is difficult to rigorously define the boundaries of the “diradical” family, it is possible to specify fundamental electronic properties which are typical of this family (Section 1).
Following that, we shall study successively the wave functions of diradicals and their covalent versus ionic character
(Section 2), general rules for the spin-orbit induced intersystem crossing between singlet and triplet states (as a
function of geometry) and other mixing mechanisms (SecAngew. Chem. internat. Edit.1 Vol. I 1 1197.2) No. 2
Документ
Категория
Без категории
Просмотров
0
Размер файла
985 Кб
Теги
spectroscopy, esca, chemical, electro, analysis
1/--страниц
Пожаловаться на содержимое документа