Beruflich Dokumente
Kultur Dokumente
AND CLUSTERS IN
CHEMISTRY AND CATALYSIS
INDUSTRY-UNIVERSITY COOPERATIVE CHEMISTRY
PROGRAM SYMPOSIA
ORGANOMETALLIC COMPOUNDS
Edited by Bernard L. Shapiro
HETEROGENEOUS CATALYSIS
Edited by Bernard L. Shapiro
NEW DIRECTIONS IN CHEMICAL ANALYSIS
Edited by Bernard L. Shapiro
FUNCTIONAL POLYMERS
Edited by David E. Bergbreiter and Charles R. Martin
Edited by
John P. Fackler, Jr.
Texas A&M University
College Station, Texas
This book contains a series of papers and abstracts from the 7th
Industry-University Cooperative Chemistry Program symposium held in the
spring of 1989 at Texas A&M University. The symposium was larger than
previous IUCCP symposia since it also celebrated the 25 years that had
elapsed since the initial discovery by F.A. Cotton and his co-workers of
the existence of metal-metal quadruple bonds. Cotton's discovery
demonstrated that multiple bonding in inorganic systems is not governed
by the same constraints observed in organic chemistry regarding s and p
orbital involvement. The d orbitals are involved in the multiple bonding
description. The quadruple bond involves considerable d orbital overlap
between adjacent metal centers. Part I of this series of papers focuses
upon the impact of this discovery and describes further contributions to
the development of the field. Multiple metal-metal bonding now is known
to permeate broad areas of transition metal chemistry.
In Part III the papers examine the role clusters play in describing
and understanding solid state materials. Clusters exist throughout solid
state chemistry. Modern structural tools now enable us to examine the
properties of these materials. Part IV presents papers which relate to
the procedures we use today to understand the clusters themselves, the
bonding theories and spectroscopy.
v
and John Smegal. Professor Arthur Martell was a great help to me on
various aspects of the programming and planning. His wife, Mary Martell,
also gave special attention to the details. My students also helped with
transportation and the audio-visual aids. I am especially grateful to
Mrs. Carol Dissen who organized the paperwork associated with the
symposium and made sure that the manuscripts were properly treated. Her
organizational skills made the whole activity painless for the editor. I
am indebted to Mrs. Sherri Sanford for her efforts as the book neared
completion.
CLUSTERS IN CATALYSIS
vii
Mechanistic Features of Carbonyl Cluster Rearrangement 141
Brian F. G. Johnson, Adrian Bott, Robert E. Benfield,
Dario Braga, Elisabeth A. Marseglia and Alison Rodger
CLUSTERS IN MATERIALS
ABSTRACTS
viii
Facile Exchange of Terminal, Doubly-Bridging, and
Quadruply Bridging Carbonyl Ligands in Solution:
Crystal Structure and Solution Dynamics
of LWM3 (C0) 12H, L=C 5H5 , C5Me 5 and M=Os,Ru 312
Yun Chi, Sue-Lein Wang and Shie-Ming Peng
ix
Bimetallic Hydroformylation Catalysis . . 328
Scott A. Laneman and George G. Stanley
252 cF-Plasma Desorption Mass Spectra of Very
Large Clusters . . . . . 329
J. P. Fackler, Jr., C. J. McNeal and R. E. P. Winpenny
Contributors 337
Index . . . . 339
X
A QUARTER-CENTIJRY OF METAlrMETAL MULTIPLE BONDS
F. A. Cotton
Department of Chemistry
Texas A&M University
College Station, Texas 77843
The early growth of the field was slow, but soon revealed itself to
be exponential, as shown in Fig. 1, for close to two decades. In the
past few years activity has leveled off at about 140 publications per
year.
2
150
140
130
I
120
110
100
90 ~
I
0
80 ti
70 ~
m
60 ~
50 LL
0
40 ffi N = e xp LQI57(Y-1952)]
200
Mo
100 w
80 Rh
Re
60
40 Ru
pt
Os
20
10
Tc
5
3
Rh 2 4+ to Rh 2 5+ As a real-world, practical chemist, I justify their
inclusion because their electronic structures (and molecular structures)
are simply logical extensions of those in which the M-M bond orders are
higher, as shown by the following progression:
The element chromium has not been included in Fig. 2, partly be-
cause it is hard to determine a temporal point of departure. Prior to
the present era a number of papers dealt with Cr 2 (0 2 CR) 4 L2 compounds.
However, it was with the discovery of the supershort Cr-Cr bonds
(<2.0 A) in 1977, 9 that the study of Cr-Cr multiple bonds acquired an
enormously greater interest. These super-short bonds are fully com-
parable in electron density with their Mo and W cogeners, and, indeed,
surpass them in this respect. This extraordinarily high bond electron
density has lead to very formidable problems in calculating their
electron<c structures. 10
L 0 f course, t h ere are even greater problems
for the "simple" diatomic Cr 2 itself, and for neighboring ones such as
V2 , Nb 2 , Mo 2 and Ru 2 , where electron correlation problems are what
Texans call humongous.
Until very recently, chromium was the only element in the first
transition series to form M-M bonds analogous to those of its heavier
congeners. In general terms, we can understand the reluctance of the
other first series elements to do so. Their d orbitals are small and
thus interatomic overlaps are relatively poor while intraatomic d-d
interactions are strong, thus making it energetically costly to reach
the appropriate valence states. Recently, however, we have obtained a
4+ 11 4+
Co 2 analog to certain Rh 2 species and it contains the shortest
Co-Co bond ever reported, 2.265A. This encourages us to believe that
with judicious choice of ligands we may succeed in forming other M-M
bonds in the first series, such as V=V 4+ triple bonds, Fe-Fe 4+ double
6+
bonds, and even Mn•Mn quadruple bonds.
4
The region of the d-block in which multiple M-M bonds along 3- or
4-fold molecular axes may be formed has also been extended in other
directions. The first Ir-Ir 4+ compound that is isostructural with a
rhodium analog has been made 12 and compounds containing Nb•Nb 4+ and
Ta=Ta4+ triple bonds have been synthesized and found to be extremely
stable. 13
REFERENCES
3. E. Peligot, C. R. Acad. Sci., 1844, 19, 609; Ann. Chim. Phys., 1844,
12, 528.
4. E. W. Abel, A. Singh and G. Wilkinson, J. Chern. Soc., 1959, 3097;
T. A. Stephenson, E. Bannister and G. Wilkinson J. Chern. Soc., 1964,
2538.
5. D. A. Edwards and G. W. A. Fowles, J. Less-Common Metals, 1962, ~.
512.
6. H. Blondel, Ann. Chim. Phys., 1905, !L 110.
7. J. D. Eakins, D. G. Humphreys and C. E. Mellish, J. Chern. Soc.,
1963, 6012.
B. V. G. Tronev and S. M. Bondin, Dokl. Akad. Nauk SSSR, 1952, 86, 87.
5
9. F. A. Cotton, S. A. Koch and M. Millar, J. Am. Chern. Soc., 1977, 99,
7273.
10. SeeR. D. Davy and M. B. Hall, J. Am. Chern. Soc., 1989, 111, 1268
for the most recent calculations.
11. F. A. Cotton and R. Poli, Inorg. Chern., 1987, 26, 3652.
12. F. A. Cotton and R. Poli, Polyhedron, 1987, ~. 1625.
13. F. A. Cotton, M. P. Diebold and W. J. Roth, J. Am. Chern. Soc., 1987,
109, 5506.
6
THE MULTIPLE METAL-METAL BOND: TWENTY-FIVE YEARS OF SYNTHETIC SERENDIPITY
Richard A. Walton
Department of Chemistry
Purdue University
West Lafayette, IN 47907
INTRODUCTION
Q Re Q Cl or Br
The close relationships that can exist in some instances between these
different species is illustrated by the metathesis reaction between W2(0-
~-Bu)6, a complex which contains a W=W bond, and aliphatic acetylenes to
give (~-Bu0)3W=CR.7
8
Fig. 1. The structure of RezCls(dth)z. (Refs. 10 and 11).
9
THE TRINUCLEAR HALIDES Re3X9 REVISITED.
3 4
QRe Qc1 or Br Q)py
10
Fig. 2. The structure of ReJClg(dppm)J. (Ref. 20)
This sequence of reactions involves the reduction in the metal core from
Re 26+ to Re 25+ to Re24+, with an accompan2in~ change in electronic config-
uration from u2~462 to u2~ 4 626*1 to u2~ 4 6 6* i.e. the Re-Re bond order
*We have shown recently that when PMe3 is used, the complex Re2Cl4(PMe3)4
is formed.
11
p
In the previous section it was pointed out that the Rez4+ core is
very easily oxidized in two one-electron steps to Re 25+ and Re 26+. The
extent to which this redox behavior can be utilized in the reduction
12
0 0
0 0
-5
-
6
13
C2
Br2
Cl
Fig. 4 Structure of the [Re2Br3(~-HN2C2Me2)(~-dppm)2(NCMe)]+
cation. (Ref. 35).
14
a-Re2Cl4(dppbe)2 into the imido complexes Re(NCH2R)Cl3{dppbe) upon
reaction with organic nitriles in the presence of cone. HCl. Second, as a
consequence of our recent discovery of a new class of dirhenium complexes
that can be formulated as 'mixed-valent' Re(IV)-Re(II) species. While we
were investigating the reactions of the quadruply bonded his-acetate
dirhenium{lll) complexes Re2{02CCH3)2X4L2 (X = Cl or Br; L = H20, py, DMF
or Me2SO) with monodentate tertiary phosphines in alcohol solvents {ROH;
R = Me, Et, n-Pr or i-Pr) we isolated alkoxide complexes of stoichiometry
Re2X4(0R)2(PPh3)2 in the case of PPh3. 37 With other phosphines {PMe3,
PMe 2Ph and PMePh2) the reactions resulted in the reduction of the
dirhenium core and the formation of the now familiar triply bonded com-
pounds Re2X4(PR3)4. 37 However, rather than Re2X4(0R)2(PPh3)2 having a
structure that resembles those of the well characterized Re(Ill}-Re(III)
derivatives Re2X6(PR3)2 (X= halide), 38 they are in reality the Re(IV)-
Re(II) species (R0)2X2ReReX2(PPh3)2. This has been established in the
case of Re2Cl4(0Et)2{PPh3)2 by a crystal structure determination (Fig. 5).
The very short Re-Re bond (2.23l(l)A) and an eclipsed rotational geometry
is in accord with the retention of a Re-Re quadruple bond and, formally,
with one component of this bond being dative in character in the sense
Re(l)~Re(2} i.e. Re*==Re. This is similar to the situation that would
hold in the case of Re2Cls{dth)2, if we formulate it as a Re(IV)-Re(I)
dirhenium complex (vide supra).
C2
Cl4
C4
15
A final development that should be mentioned is the recent synthesis
and characterization of Re2Cl5(dto)2 (dto ~ 3,6-dithiaoctane), 4 0 a close
relative of Re2Cl5(dth)2. The reaction of (n-Bu4N)2Re2Cl8 with dto in
ethanol proceeds via (n-Bu4N)Re2Cl7(dto); this has been structurally
characterized <z). Interestingly, if a similar intermediate is formed in
the reaction between [Re2Cl8]2- and dth, then (n-Bu4N)Re2Cl7(dth) could
react with a further equivalent of dth in one of two ways. If attack is
at the rhenium atom that does not already contain a coordinated dth
ligand, we can expect Re2Cl6(dth)2 to be formed. Attack at the other
rhenium could lead to Re2Cl5(dth)2, as has been suggested by Heyen and
Poweu. 4 0
ACKNOWLEDGEMENTS
I would like to thank the National Science Foundation for the support
of much of our work in this area, and my talented coworkers whose
contributions are cited throughout this article.
REFERENCES
16
8. F. A. Cotton and R. A. Walton, Inorg. Chern., 5:1802 (1966).
9. F. A. Cotton, C. Oldham and R. A. Walton, Inorg. Chern., 6:214 (1967).
10. M. J. Bennett, F. A. Cotton and R. A. Walton, J. Am. Chern. Soc.,
88:3866 (1966).
11. M. J. Bennett, F. A. Cotton and R. A. Walton, Proc. Roy. Soc.,
303:175 (1968).
12. W. K. Bratton, F. A. Cotton, M. Debeau and R. A. Walton, J. Coord.
Chern., 1:121 (1971).
13. R. Colton, R. Levitus and G. Wilkinson, J. Chern. Soc., 4121 (1960).
14. D. G. Tisley and R. A. Walton, Inorg. Chern., 12:373 (1973).
15. H. D. Glicksman and R. A. Walton, Inorg. Chim. Acta, 19:91 (1976).
16. See for example, V. I. Spitsyn, A. F. Kuzina, A. A. Oblova and S. V.
Kryuchkov, Russ. Chern. Revs., 54:373 (1985).
17. R. A. Wheeler and R. Hoffmann, J. Am. Chern. Soc., 108:6605 (1986).
18. See ref. 1, p 275.
19. F. A. Cotton and J. T. Mague, Inorg. Chern., 3:1094 (1964).
20. S.M.V. Esjornson, P. E. Fanwick and R. A. Walton, unpublished results
(1988).
21. J. R. Ebner and R. A. Walton, Inorg. Chern., 14:1987 (1975).
22. P. Brant, D. J. Salmon and R. A. Walton, J. Am. Chern. Soc., 100:4424
(1978).
23. F. A. Cotton, K. R. Dunbar, L. R. Falvello, M. Tomas and R. A.
Walton, J. Am. Chern. Soc., 105:4950 (1983).
24. F. A. Cotton, B. A. Frenz, J. R. Ebner and R. A. Walton, Inorg.
Chern., 15:1630 (1976).
25. T. R. Ryan and R. E. McCarley, Inorg. Chern., 21:2072 (1982).
26. R. A. Walton, A.C.S. Symp. Ser., 155:207 (1981) and references cited
therein.
27. R. A. Walton, Polyhedron, in press (1989).
28. A. C. Price and R. A. Walton, Polyhedron, 6:729 (1987) and references
cited therein.
29. F. A. Cotton, K. R. Dunbar, A. C. Price, W. Schwotzer and R. A.
Walton, J. Am. Chern. Soc., 108:4843 (1986).
30. F. A. Cotton, K. R. Dunbar, L. R. Falvello and R. A. Walton,
Inorg. Chern., 24:4180 (1985).
31. L. B. Anderson, F. A. Cotton, K. R. Dunbar, L. R. Falvello, A. C.
Price, A. H. Reid and R. A. Walton, Inorg. Chern., 26:2717 (1987).
32. P. E. Fanwick, A. C. Price and R. A. Walton, Inorg. Chern., 27:2601
(1988).
33. T. J. Barder, F. A. Cotton, L. R. Falvello and R. A. Walton,
Inorg. Chern., 24:1258 (1985).
34. D. Esjornson, P. E. Fanwick and R. A. Walton, Inorg. Chern., 27:3066
(1988).
35. D. Esjornson, D. R. Derringer, P. E. Fanwick and R. A. Walton,
Inorg. Chern., submitted for publication.
36. M. Bakir, D. Esjornson, P. E. Fanwick, K. S. Jones and R. A. Walton,
unpublished results (1989).
37. A. R. Chakravarty, F. A. Cotton, A. R. Cutler and R. A. Walton,
Inorg. Chern., 25:3619 (1986).
38. See ref. 1, p 49.
39. M. H. Chisholm, J. C. Huffman and W. G. VanDer Sluys, J. Am. Chern.
Soc. ,109:2514 (1987).
40. B. J. Heyen and G. L. Powell, Polyhedron, 7:1207 (1988).
17
THEORETICAL STUDIES OF DINUCLEAR COMPOUNDS WITH MULTIPLE METAL-METAL BONDS
Department of Chemistry
The Ohio State University
Columbus, OH 43210
Introduction
The last twenty-five years have been witness to the birth and growth
of metal-metal multiple bond chemistry. During this time a wide variety
of compounds containing multiple bonds of order up to four have been
synthesized, and a wealth of data concerning their structure, properties,
and reactivity has been collected. 1 An exciting aspect of research in this
field is the central role that both qualitative and quantitative
electronic structure theory have played in developing a rational framework
for understanding the physical structures and properties of these
complexes. Approximate molecular orbital theory was applied to these
systems almost as soon as they were discovered, and its use has progressed
in unison with experimental work, so that the two are intimately
intertwined. The synthesis of a new compound containing a metal-metal
multiple bond is almost always combined with an interpretation of its
properties and reactivity in terms of some form of molecular orbital
theory. Conversely, the metal-metal multiple bond problem provides a good
test of the usefulness of an electronic structure method.
This synergism between theory and experiment that has existed in the
metal-metal multiple bond field could not have developed without the
advent of theoretical methods that are flexible enough to treat very large
molecules. For instance, the prototypical quadruple-bond-containing
complex, the octachlorodirhenate(III) dianion (1), contains 288 electrons,
Cl Cl iz-
1 ,,....,_Cl I ....,...,_Cl
Rci"-·-4-R{
Cl~~Cl Cl/1Cl
1
a9(z2)
Xx M-L<J
D
D D
M-X<J
Xx
D M-X<J
D
M
x•... M .••• x X,,,,M,,,,L
x-' ""x L,. ""x
Figure 1. Qualitative molecular orbital scheme for a cationic transition
metal (a), a D4h ML4 fragment (b), a D4h MX4 fragment (c), and a
D2 h ML2 X2 fragment (d).
20
Cotton was able to propose a qualitative description of the bonding
between the two metals which explained the structure. 4 The rhenium
orbitals necessary for metal-ligand bonding are subtracted from the metal
valence atomic orbital set, and then the remaining metal-based orbitals
are combined to form rhenium-rhenium bonding and antibonding molecular
orbitals of a, ~. and 6 symmetry with respect to the Re-Re vector. The
eight metal-based electrons are just enough to fill all the bonding
orbitals, and a net quadruple bond results.
21
particular ligand and metal, but a reasonable net ordering is a 18 (d.z) < e 8
(d,.., ely.) < b 28 (d,cy) (Figure lc). 7 The stronger the w-donating ability of
the ligands, the greater the splitting in energy of these orbitals. In
actual compounds with the M2X8 stoichiometry, the ligands are bent slightly
away from one another on opposite sides of the molecule, so that the M-M-X
angles are greater than 90" (103.7" in 1 3 ). This bending is due to
nonbonded repulsions between the ligands, and results in a more complex
mixing of the ligand and metal orbitals. The electronic consequences are
minor, however, and it is equally informative to treat the MX 4 units as two
square planes.
a 19(a)
....,X I ,...,X
X X
IM·~M···
x/1 x/1
X X
Figure 2. Qualitative molecular orbital diagram of the interaction of
two MX4 fragments to generate the four components of an M-M
quadruple bond.
22
in this level, the two square planar fragments must orient themselves in
an eclipsed conformation. The combination of these components results in
a formally quadruple bond between the two Re atoms.
23
~
X P p
I~ \~
L X X L
lb) in that two u-only donor ligands are replaced by two u and ~ donor
ligands. The ~ (b 18 ) orbital is destabilized by ~ interaction, but by a
lesser amount than in MX 4 because there are only half as many ligand ~
orbitals to interact with. We also see that only one of the dzz (b 28 ) and
dy. (b 38 ) orbitals is oriented properly to receive ~ electron density from
the ligand set, so that the degeneracy of these two is lifted. Thus, if
the ~ donating ligands are oriented along the y axis, the dy. orbital is
destabilized with respect to the ~•• and the former is hybridized away
from the metal more than the latter. The metal-based electrons are now
divided among four non-degenerate fragment orbitals of mostly d metal
character.
I ~MCJI -~Cl
Cl PMe3
M~~Mo.•
MeJP
~ICI CI;1PMe3
5 6
24
going from a D 4h to a Dzd molecule does not have a large effect on the
overall bonding in Mo 2 (PMe 3 )Cl 4 .
(eV)
10e(11')
·2.0
(eV)
·1.0
-3.0
·2.0
-4.0
5b2(11)
·3.0
5tJ:! (yz)
5b1 (11)
~-I'•F'T'"T'~(8)
7a 1 (a) -5.0
-5.0
-7.0
-6.0
·7.0 7a 1 (a)
4a1 (a)
-8.0
HO PH
1,..-PH•I.J>H 9_.,,PH,IPH}:.I
Mo...i-Mo Mo...i-Mo
HO OH'IPH, H,J>'Cl
H,P"I :l Cl.,1PH,
Figure 3. Molecular orbital correlation diagram for Mo 2 (0H) 4 (PH3 ) 4 (left)
and Mo 2 Cl 4 (PH 3 ) 4 (right) showing the interactions important for
metal-metal bond formation. The energy scale has been offset
to make the mid-points of the HOMO-LUMO gaps for the two
molecules coincident. (From ref. 17).
25
are not currently known to exist. The 6 bond is also diminished by the
rotation, such that it disappears when the two fragments are exactly
staggered. 21 The degree of twist, and the resultant loss of 6 bonding, can
be correlated with an increase in the metal-metal separation and a
decrease in the 6 ~ 6* transition energy. 22
n :: :-
I *,p I . ,p I
RO PPh3 OR
,.oR
......~
...... ~
' M"a-3-Mi,.
.....:..; 4
Re--Re
Cl~~ Cl~~PPh3 '.~pOR
. .,.IOR
RO '
PJ
7 8
26
OH
I ~PH'I ~oH
H~
""'illil ~...
M6"-·-M~. .
(eV) ,I ~1
Hy> Hy> OH OH
-1.0
10e(7t")
-2.0
-3.0
-4.0
4a, (z2j
-5.0
-6.0
are not great enough to prevent formation of the 6 bond. These two
examples point out the important role that metal-ligand interactions have
in tuning metal-metal bonding interactions_
27
bond has not been destroyed. The quadruply bonded [Tc 2 Cl 8 ] 2 - ion has also
been characterized, and, interestingly, the Te-Te bond is longer than that
found in [Tc 2Cl 8 ] 3-. 26 The reason, as suggested above, most likely rests in
the lower metal oxidation state of [Tc 2 Cl 8 ] 3-, leading to commensurately
more diffuse 4d AOs, and hence stronger Te-Te w and 6 bonds. 27 No accurate
calculational results have been reported to verify this hypothesis.
L L L
I \f
M -3- M
I\L
L
I
L
case, but the symmetry of the interactions dictates that fewer metal-based
orbitals remain available for metal-metal bond formation. Thus, the
maximum metal-metal bond order in the M2Ls molecules is less than that in
M2La.
28
(a) (b) (c)
E
0
M-Lo
M-Lo
D
M"'+ L'"'"·M-L
L_.... L~7M......_L
Figure 5. Qualitative molecular orbital diagrams for a cationic
transition metal (a), a planar (D 3h) M~ fragment (b), and a
pyramidal (C 3v) ML3 fragment (c).
29
E
a1a (a)
,
L L L L L L
I
M I \"'
M-'-l
,~
!
M
I\L
L L
....
'i.
I
L
I
L
30
These methods all predict nearly free rotation about the metal-metal bond
in the absence of ligand-ligand interactions. The qualitative conclusion
which was reached above, that the d,a-ya and d,y orbitals are utilized
exclusively for metal-ligand bonding, is modified somewhat by the
inclusion of obtuse M-M-L angles, but is still basically accurate.
OH HO HO
/\I I
Mo--Mo
3 ''
OH OH HO
12
(aV)
-4
-5
-6
Mo-MoCJ Mo-MoCJ
-7
-8
D
D D
01t
-9
-10
-11
C-HCJ
Mo-NCJ
D
Mo-OCJ
10 11 12
Figure 7. Comparison of the valence level molecular orbital schemes for
Mo 2 (CH 3 ) 6 (10), Mo 2 (NH2 ) 6 (11), and Mo 2 (0H) 6 (12). (From ref.
34).
31
d,..y orbitals through 1r interactions. Thus, in 11, where strong Mo-N
d,..2-y2,
1rinteractions are present, there is no o• contribution to the Mo-Mo 1r
bonding MO, but in 10, in which metal-ligand 1r interaction is minimal, a
significant (27.8%) admixture of o· character is present in the
corresponding level. These three calculations point out the wide
variation in metal-metal character that can be induced by the ligand set.
While the basic electronic structures of 10, 11, and 12 all are similar to
that proposed qualitatively in Figure 6, each is unique due to the
characteristics of the surrounding ligands.
5 (eV)
-1
-2
-3
-4
-5
Mo
32
1.93±0.02 A. 37 This dimer has the shortest Mo-Mo bond known, a strong
indicator that the metal-metal bonding is of a high order. Likewise, Cr2
has an extremely short (1.68±0.01 A) bond, 38 much shorter than that
observed in any ligated complex of dichromium.
33
disagreement between this value and the prior experimental
determination. 37 " The bond length and spectroscopic properties were
calculated with much better agreement to experiment by using the GMO
MGSGF-GI treatment. 41 The highest quality calculation, including bond-
centered basis functions and 3212 different configurations, yields final
natural orbital populations of (la 6 ) 1 · 88 (l:lru) 3 · 78 (16 6 ) 3 • 42 (2a6 )1. 92 (lau) 0 · 08
(16u) 0 · 58 (l1rg) 0 · 22 (2au) 0 · 12 , corresponding fairly closely to that predicted
by less rigorous methods. The greater correlation of the 6 orbitals
compared to the 2ag is suggestive of the greater importance of the sa bond
in decreasing the Mo 2 bond length, just as was concluded from the Xa-SW
results.
~----~
#!-~~----~
u-·--o-~
~:·----(j)
34
actinide rather than lanthanide metals since the actinide Sf orbitals are
considerably more diffuse that the 4f orbitals of the lanthanides. We
anticipate this to be an exciting area for synthetic chemists as well;
there are currently no known complexes that contain direct metal-metal
bonds between two f-element metal atoms, in spite of some genuine attempts
to achieve this. 46 Actinide diatomic molecules do exist as transient
species in the gas phase, however, and, for example, the bond dissociation
energies of Th 2 and U2 have been measured mass spectrometrically as 28S±21
and 218±21 kJ mol- 1 , respectively. 47
-1 (eV)
Uli
-2
·3 Ua
-4
-5
, ';,
a 1g (a)
-7
H1C H1C CH1 CH1 CH1 CH1
I . u--u
I l" ,~
I
ru
H1C H,'t
~
u,c
....~
H,'t
I
en,
u
I
CH,
35
ligands to remove some of· the orbitals from the metal-metal bonding
manifold, as was the case in M2L8 and M2 L6 complexes? Given that there are
some similarities between the chemistry of Mo (d 6 ) and U (f6 ), we have
explored the possibility of forming multiple uranium-uranium bonds in the
hypothetical U(III) dimer U2Me 6 . 49
Acknowledgments
References
(2) (a) Johnson, K. H. Adv. Quantum Chern. 1973, l, 143-18S. (b) Slater,
J. G. "Quantum Theory for Molecules and Solids: The Self-Consistent
Field for Molecules and Solids"; McGraw~Hill: New York, 1974; Vol. 4.
(4) (a) Gatton, F. A.; Curtis, N. F.; Harris, G. B.; Johnson, B. F. G.;
Lippard, S. J.; Mague, J. T.; Robinson, W. R.; Wood, J. S. Science
1964, 145, 1305-1307. (b) Gatton, F. A. Inorg. Chern. 1965, ~. 334-
336.
36
(5) Throughout this paper, L will refer to a neutral, two-electron
donating ligand, such as a phosphine, amine, or thioether. X will
refer to an anionic two-electron donating ligand, with possible w-
donating ability, such as a halide, alkoxide, or amide. R will refer
to an anionic two-electron donating ligand with no w-donating ability,
such as an alkyl.
(10) Mortola, A. P.; Moskowitz, J. W.; Rosch, N.; Cowman, C. D.; Gray, H.
B. Chern. Phys. Lett. 1975, 32, 283-286.
(12) Trogler, W. C.; Ellis, D. E.; Berkowitz, J. J. Am. Chern. Soc. 1979,
101, 5896-5901.
(15) Chisholm, M. H.; Felting, K.; Huffman, J. C.; Tatz, R. J. J. Am. Chern.
Soc. 1984, 106, 1153-1154.
(18) Abbott, E. H.; Bose, K. S.; Cotton, F. A.; Hall, W. T.; Sekutowski, J.
C. Inorg. Chern. 1978, 17, 3240-3245.
(20) Agaskar, P. A.; Cotton, F. A.; Derringer, D. R.; Powell, G. L.; Root,
D. R.; Smith, T. J. Inorg. Chern, 1985, 24, 2786-2791, and references
therein.
(21) Best, S. A.; Smith, T. J.; Walton, R. A. Inorg. Chern. 1978, 17, 99-
104.
(22) (a) Campbell, F. L.; Cotton, F. A.; Powell, G. L. Inorg. Chern. 1985,
24, 4384-4389, and references therein. (b) Hopkins, M. D.; Zietlow,
T. C.; Miskowski, V. M.; Gray, H. B. J. Am. Chern. Soc. 1985, 107, 510-
512.
(23) (a) Chakravarty, A. R.; Cotton, F. A.; Cutler, A. R.; Tetrick, S.M.;
Walton, R. A. J. Am. Chern. Soc. 1985, 107, 4795-4796. (b)
37
Chakravarty, A. R.; Cotton, F. A.; Cutler, A. R.; Walton, R. A. Inorg.
Chern. 1986, 25, 3619-3624.
(26) Cotton, F. A.; Daniels, L.; Davison, A.; Orvig, C. Inorg. Chern. 1981,
20, 3051-?.
(27) Cotton, F. A.; Davison, A.; Day, V. W.; Fredrich, M. F.; Orvig, C.;
Swanson, R. Inorg. Chern. 1982, 21, 1211-1214.
(28) Bursten, B. E.; Cotton, F. A.; Fanwick, P. E.; Stanley, G. G.; Walton,
R. A. J. Am. Chern. Soc. 1983, 105, 2606-2611.
(29) Chisholm, M. H.; Cotton, F. A. Ace. Chern. Res. 1978, 11, 356-362.
(31) Chisholm, M. H.; Cotton, F. A.; Frenz, B. A.; Reichert, W. W.; Shive,
L. W.; Stults, B. R. J. Am. Chern. Soc. 1976, 98, 4469-4476.
(33) (a) Albright, T. A.; Hoffmann, R. J. Am. Chern. Soc. 1978, lQQ, 7736-
7738. (b) Dedieu, A.; Albright, T. A.; Hoffmann, R. J. Am. Chern. Soc.
1979, 101, 3141-3151.
(34) Bursten, B. E.; Cotton, F. A.; Green, J. C.; Seddon, E. A.; Stanley,
G. G. J. Am. Chern. Soc. 1980, 102, 4579-4588.
(35) (a) Hall, M. B. J. Am. Chern. Soc, 1980, 102, 2104-2106. (b) Kok, R.
A.; Hall, M. B. Inorg. Chern. 1983, 22, 728-734.
(36) Cotton, F. A.; Stanley, G. G.; Kalbacher, B. J.; Green, J. C.; Seddon,
E.; Chisholm, M. H. Proc. Natl. Acad. Sci. USA 1977, ~. 3109-3113.
(37) (a) Efremov, Yu. M.; Samoilova, A. N.; Kozhukhovskii, V. B.; Gurvich,
L. V. J. Mol. Spectr. 1978, 73, 430-440. (b) Hopkins, J. B.;
Langridge-Smith, P.R. R.; Morse, M. D.; Smalley, R. E. J. Chern. Phys.
1983, 78, 1627-1637.
(38) (a) Michalopoulos, D. L.; Geusic, M. E.; Hansen, S. G.; Powers, D. E.;
Smalley, R. E. J. Phys. Chern. 1982, 86, 3914-3916. (b) Bondybey, V.
E.; English, J. H. Chern. Phys. Lett. 1983, 94, 443-447.
(39) (a) Norman, J. G.; Kolari, H. J.; Gray, H. B.; Trogler, W. C. Inorg.
Chern. 1977, 16, 987-993. (b) Bursten, B. E.; Cotton, F. A.~
Faraday Soc. 1980, 14, 180-193.
(40) Wood, C.; Doran, M.; Hillier, I. H.; Guest, M. F. Symp. Faraday Soc.
1980, 14, 159-169.
(41) Bursten, B. E.; Cotton, F. A.; Hall, M. B. J. Am. Chern. Soc. 1980,
102, 6348-6349.
38
(42) Goodgame, M. M.; Goddard, W. A. G., III J. Phys. Chern. 1981, 85, 215-
217.
(43) Goodgame, M. M.; Goddard, W. A. G., III Phys. Rev. Lett. 1982, 48,
135-138.
(44) (a) Delley, B.; Freeman, A. J.; Ellis, D. E. Phys. Rev. Lett. 1983,
50, 488-491. (b) Goodgame, M. M.; Goddard, W. A. G., III Phys. Rev.
Lett. 1985, 54, 661-664. (c) Delley, B. Phys. Rev, Lett. 1985, 55,
2090. (d) Painter, G. S. J. Phys. Chern. 1986, 90, 5530-5535.
(45) (a) Kok, R. A; Hall, M. B. J. Phys. Chern. 1983, 87, 715-717. (b)
Walch, S. P.; Bauschlicher, C. W.; Roos, B. 0.; Nelin, C. J. Chern.
Phys. Lett. 1983, 103, 175-179. (c) Baykara, N. A.; McMaster, B. N.;
Salahub, D. R. Mol. Phys. 1984, 52, 891-905. (d) Sundholm, D.;
Pyykko, P.; Laaksonen, L. Finn. Chern. Lett. 1985, Sl-55.
(46) (a) Cotton, F. A.; Marler, D. 0.; Schwotzer, W. Inorg. Chim. Acta
1984, 85, L31-L32. (b) Cotton, F. A.; Marler, D. 0.; Schwotzer, W.
Inorg. Chern. 1984, 23, 4211-4215. •
(50) (a) Bursten, B. E.; Fang, A. J. Am. Chern. Soc. 1983, 105, 6495-6496.
(b) Bursten, B. E.; Novo-Gradac, K. J. J. Am. Chern. Soc. 1987, 109,
904-905. (c) Bursten, B. E.; Strittmatter, R. J. J. Am. Chern. Soc.
1987, 109, 6606-6608. (d) Bursten, B. E.; Rhodes, L. F.;
Strittmatter, R. J. J. Am. Chern. Soc. 1989, in press.
39
USES OF METAL CLUSTERS IN HOMOGENEOUS AND HETEROGENEOUS CATALYSIS
Donald J. Darensbow:g
DepanrnnentofCherrnstty
Texas A&M University
College Station, Texas 77843
Molecular metal carbonyl clusters can serve as pivotal metal derivatives for a large
variety of chemical processes. These are illustrated In Scheme 1, where a tetranuclear
cluster, the majority of which are 3-dimensional and tetrahedral, is Indicated.
Scheme 1
Metal clusters may be defined rather liberally as compounds containing more than
one transition metal atom directly bonded to each other. However, tetranuclear clusters
represent the simplest metal cluster where a pronounced similarity with the metallic
state is anticipated. That is, these species possess a 3-dimensional network of metal
atoms held together by metal-metal bonds, and where each metal atom forms at least
three different metal-metal bonds. Indeed the impetus for examining the role of metal
clusters in homogenous catalysis was prompted by Muetterties formulation of the
"surface cluster analogy". I Nevertheless. for the most part the work presented herein
will involve derivatives of the 2-dimensional trinuclear, triangular metal carbonyl
cluster, Ru3(CO) 12·
This contribution will attempt to illustrate by examples from our research
program in organometallic chemistry how metal clusters have impacted on these
various areas. Our entry into defining the mechanistic aspects of catalytic processes
involving transition metal carbonyls commenced with an investigation of the water-gas
shift reaction (hereafter referred to as WGS).2 The organometallic chemistry pertinent
to this process is indicated In skeletal form In Scheme 2.
The WGS reaction Is used to Increase the amount of H2 In CO /H2 mixtures obtained
from fossil fuels (eq. 1). Presently, catalysts consisting of Iron-chromium oxide mix-
tures are used at 350-400°C to reduce the CO content to 3-4%. When lower levels of CO (-
0.1%) are required, a mixture of Cu-Zn oxides which Is sulfur-sensitive but active at 190-
2600C Is used. For reaction (1). 6.G 0298 = 28.5 kJ/mole, or lfH20ill Is considered,
(1)
6.G 0298 =-19.9 kJ/mole. Hence, the free-energy change favoring H2 production decreases
with increasing temperature.
Soluble group 6 and group 8 metal complexes are attractive, homogeneous,
altematlves as WGSR catalysts since they are potentially resistant to poisoning by
sulfur and also are active at low temperatures, where very little CO is present at
equilibrium. The most thoroughly investigated catalytic system is that derived from
ruthenium dodecacarbonyl in aqueous ethoxyethanol solutions in the presence of
KOH.3-5 This catalyst system operates between 90-110'C and 0.5-2.0 atm CO, with the
mature catalyst providing a tumover number of approximately 3 moles C02/H2 per
mole Ru3(COl12 per day. During catalysis the,prtmary ruthenium species In solution are
HRu3(COl11- and H3Ru4(COl12-. Since these two clusters are linked together by the
equilibrium process indicated in eq. (2). as H2 accumulates the tetranuclear species
become predominant. 6
It has been proposed that the catalytic process can be divided into two cycles
1
(Scheme 3) involving separately the tetranuclear and trinuclear ruthenium clusters,
H2 Ru3(C0)12 " \ H -
CO C0 2
Schfme3
42
with the latter species being catalytic more active. Consistent with this implication
when the H2 is continuously removed from the reaction mixture, only HRu3(CO) 11- was
observed and concomitantly the turnover frequency increased by fourfold over that seen
when H2 was allowed to build up in concentration.
,0
[MtH + C02 [MtO-C-H (3)
co,-( ~H 2
~co
Schane4
HCOOH CO
HCOOMe j
+ J___MeOH
HP
43
Consistent with Scheme 4 the addition of CO retards alkyl formate production,
strongly implying C02 as the primary source of the carboxylic carbon atom in HCOOR
This was verified by carrying out reactions in the presence of HC02W(l3co)5- which
provided only H 12cooR after short reaction times. However, in the absence of C02 and
H2 the anionic metal hydrides were observed to be effective catalyst precursors for con-
verting CO and methanol Into methyl formate. In this connection it is important to
recall that a combination of reactions (1) and (4) leads to reaction (5), i.e., the formation
of methyl formate from the carbonylatlon of methanol.
Cf\OH
+
co
In totally analogous processes anionic ruthenium carbonyl clusters serve as
catalysts or catalyst precursors for the hydrogenation of carbon dioxide in alcohols
(eq.4) or the carbonylation of alcohols (eq.5) to produce alkyl formates. There are
however some subtle differences with regard to reactivity and reaction pathway
obseiVed when employing these multinuclear metal carbonyls as catalysts. Table 1
contains a portion of the data obtained for the catalytic hydrocondensatlon of C02 and
molecular hydrogen by anionic ruthenium clusters In the presence of methanol to
provide methyl formate.14 The conditions for the reaction were moderate pressures of
carbon dioxide and hydrogen (250 psi of each at ambient temperature) and a temperature
of 125'C.
44
Table 1. Formation of Methyl Formate from the Oxides
of Carbon Using Anionic Ruthenium Carbonyl Clusters
As noted in Table 1 the HC02RU3(CO) 10- anion has a slightly greater activity for
alkyl formate production than the HRu3(COl 11- anion, by about one turnover. This
metalloformate derivative has been synthesized independently from either Ru3(COh2
and [PPNII02CHI (eq.6) or C02 insertion into HRu3(COl11- (eq.7), and it has been fully
characterized.15 Indeed, the formate ruthenium cluster would be antlclpated to be an
(7)
intermediate in the catalytic cycle originating with HRu3(COJ 11- on the basis of equa-
tion 7. However, unlike the mononuclear group 6 metal carbonyl hydrides which read-
ily react with carbon dioxide the reaction of HRu3(CO) 11- with C02 is energetically less
favorable, occurring only at moderate pressures (60-400 psi) of carbon diOXide. In fact
reaction (7), or analogous process involving an alternative polynuclear metal carbonyl
hydride species (vide infra), most likely represents the rate-determining step in the cat-
alytic production of alkyl formates.
45
description the rate of 13co ligand exchange with HRus(CO) 11- Is faster than the rate of
C02 insertion. Concomitantly. the insertion process is inhibited by the presence of
carbon monoxide. resulting in a decrease in catalytic activity in the hydrocondensation
of C02 (Table 1, entry 5).
SchemeS
Scheme 6 assumes the involvement of two metal centers in the carbon dioxide
insertion reaction. i.e .. intermediate A instead of B. Although species B is possible, as Is
required for mononuclear metal complexes, there appears to be less molecular rear-
rangement required for the formation of A. This is seen upon modelling the decarboxy-
lation process by simply bond-opening one end of the bridging formate ligand in the
structure ofHC02Rus(C0)1Q- and rotating about the alternate C-O bond (eq.S).
- -
i i
o, . .
/c, /c\ (8)
I ~
Ru--Ru
~
Ru--Ru
A B
This operation places the formate hydrogen atom much closer to the metal center adja-
cent to the one bearing the monodentate formate ligand than to the metal center bonded
directly to the formate ltgand, 2 .33 A vs 3 . 16 A.
Thus far we have not quantified the kinetic parameters for the carboxylation or
decarboxylation processes involving a bridging hydride moiety as in these cluster
derivatives, however, the reaction indicated in eq.9 has been investigated kinetically for
mononuclear metal hydrides employing isotopically labelled C02. 18,19 The details of
46
one such Investigation is summarized In Scheme 7, where the energetic for the barriers
were obtained from reaction (9) and CO ligand substitution in HC02Cr(COls-.20 As
would be anticipated based on Scheme 7 the decarboxylation reaction Is retarded In the
presence of carbon monoxide.
C0 2 Insertion Mechanism
~ e
,c,
(1) 0 H :
I ,.; ::
!
HCr(CO); + C02 ~ -Cr-
II""' I (5) :
........ -- --.--------------------------------------------- ------ ~
l
Scheme7
Rxn Coordinate
Retumlng to Table 1 It is apparent from the last entry that HRU3(CO) 11- is an
effective catalyst for methanol carbonylation (eq.5). In the presence of excess carbon
monoxide HRu3(CO) 11- In hot methanol readily affords Ru3(CO) 12 and hydrogen (vide
supra), in addition to an equivalent quantity of OMe-. Indeed a comparably active
catalyst system is provided by Ru3(COh2 and one equivalence ofKOMe.12 The proposed
methoxycarbonyl intermediate in the methyl formate synthesis (eq.9) is observable by
IR (vCO: 2072(w), 2015(s), 1990(s), 1965(m), and 1603 cm-1) and 13c NMR (206.6 ppm@
ambient temperature) spectroscopies In TIIF/MeOH.21,22 Reaction (10) has as well
47
remain on catalytic processes incorporating carbon dioxide. In particular an
investigation of supported metal catalysts for the hydrogenation of C02 to methane,
methanation (eq.ll). 23
1) THF or hexane
Al203
(12)
Catalysts prepared similarly from RuCIJ produce a different surface species upon
activation under hydrogen at 200'C followed by exposure to carbon monoxide. The
infrared spectrum of this species exhibits a three band pattern with peaks at 2140, 2075,
and 2013 cm-1. which matches the ..J(CO) infrared pattern of [RuC12(COl31n or
[RuC12(COl41n. The species produced by the two types of supported ruthenium complexes
exhibit different reactivities and selectivities as well. It was shown that the ruthenium
carbonyl clusters are generally more active catalysts precursors than the ruthenium
chloride derived catalyst (Table 2). They are also much more selective, producing
methane and water exclusively. The ruthenium chloride derived catalysts, on the other
hand, also produce carbon monoxide and water (WGS chemistry).
RuClj3Hp 1.0
Ru(C0)5 4.7
Ru3(C0)12 12.6
H4 Ru4(C0\ 2 14.6
CRu6(C0\ 7 21.7
48
Although all the surface species produced by the various ruthenium carbonyl
derivative appear to be identical, it is obvious upon inspection of Table 2 that these vary
significantly in catalytic actiVity toward methanation. Indeed a correlation is noted
between metal cluster size and reactiVity, with the higher nucleartty clusters exhibiting
greater reactivity. This is best seen in a plot of metal dispersion as determined from 02
chemisorption measurements and precursor, cluster nuclearity (Figure 1). Since the
metal particles are generally greater than 50 A as determined by electron microscopy, it
emerges that the degree of metal aggregation upon decarbonylation (or activation) is
inversely proportioned to cluster nucleartty.
The surface reactions of the activated cluster catalysts were studied by diffuse
reflectance FTIR spectroscopy in an effort to identify the surface species which leads to
methane production. 26 Figure 2 shows the diffuse reflectance spectra of the surface of
an activated catalyst under a mixture of C02:H2 of 1:2.5, as is used in catalysis. Under
this atmosphere the original carbonyl bands maintain their band pattern but are
shifted slightly to lower energies at 2036 and 1954 cm-1. A formate species is also
present on the surface under catalytic conditions as evidenced by the symmetric and
antisymmetric ..J(C02) stretches at 1591 cm-1 and 1375 cm-1 and also the C-H stretch at
2904 cm-1. A 150'C, no methane is observed: however, when the sample is heated to
200'C, methane production commences as is seen by the appearance of the band at 3016
cm-1. This band can be assigned to the C-H stretching frequency of gaseous methane
which is trapped in the DRIFTS cell, and so is observed in the IR band pattern. Methane
production increases with increasing temperature, as can be seen by the amount of
methane produced at 250'C in Figure 2.
I.IRVENUI"18EAS CM-l
49
Independent experiments indicate that when a catalyst pretreated with C02 and
H2 at 130'C to afford surface formates is evacuated and retreated with hydrogen,
methane production occurred without an appreciable diminish of the infrared bands due
to the formate species. Hence, hydrogenation of the surface formates does not appear to
be the source of methane. This taken along with changes noted in the terminal CO
region of the infrared spectra during catalysis leads us to propose a mechanism for
methane formation which proceeds via metal carbonyl and RuO(ads) to RuC(surface
species). Hydrogenation of RuO(ads) and RuC(surface species) ultimately result in
methane and water production.
50
The catalytic activities of the two catalyst precursors upon activation under
hydrogen or helium at 200'C for 8-10 hrs are reported as plots of percent conversion of
C02 to CH4 per unit weight of ruthenium vs temperature (Figures 4 and 5). In both cases
selectivity to methane was greater than 95%, with a trace of CO at temperatures at or
above 300'C being the only other observable carbon-based product. In each case, the
catalytic activity is compared to that of the previously discussed Ru3(COl12/A1203
precursor. As is evident in Figures 4 and 5 the best catalyst precursor for methanation of
C02 is the one in which Ru3(CO) 12 was deposited from the vapor phase onto the outer
surface of the zeolite. For the zeolite supported catalysts the metal dispersion was
greater for the helium activated (>60"Al) than for the hydrogen activated catalyst (see
Figure 6). Indeed in the case of the intrazeolite catalyst, helium activation appears to
result in little metal aggregation, as is apparent in that the catalyst retains its yellow
color upon activation, rather than decomposing to a grey solid (suggestive of the growth
of metal particles) which is the case for the H2-activated catalyst.
IDO
I
I
I
B
I
10 I
I
::> I
..,
a:
I
I
I
~ 50 I
-.. I
I
~ I
I
~
I
40 I
w I
>
~
u 20
~
100
B
cf!BO
""
~50
z
.0
~
~ 40
z
8
..:zo
350 400
51
28
24
a::" A
"'
o-
w 20
(D
a::
0
(fJ
0 16
<{
0
u B
(fJ 12
w
_j
0
:::;:
':'
Q
><
10 15 20 25
PRESSURE (mm Hgl
Contrary to the behavior of the vapor phase deposited Ru3(CO) 12 catalyst, which
suffered a slight loss in catalytic activity with time, presumably due to a buildup of
inactive carbon on metal sites, the behavior of the intrazeolite ruthenium catalyst
displayed an increase in activity with time under catalytic conditions. Nevertheless,
during this time period the metal dispersion decreased significantly (see Figure 6). For
example, an increase in catalytic activity of >300% was exhibited by a helium-activated
system after 72 hrs of catalysis at 300"C. This is most likely due to the high volatility of
the very stable ruthenium oxides which migrate and agglomerate outside the zeolite
supercages,31,32 where the ruthenium has been shown to be catalytically more active.
The low activity initially manifested by the highly dispersed ruthenium metal in the in-
trazeolite catalysts is evidently a mass transport problem, even though the reactants
and products are themselves all small entities.
References
1.) E. L. Muetterties, Bull. Soc. Chim. Bel~ .. 84, 959 (1975); .!b!d... 85, 451 (1976).
52
3.) R M. Lalne, R G. Rinker, and P. C. Ford, J. Am Chern Soc., 99, 252 (1977).
6.) J. C. Bricker, C. C. Nagle, and S. G. Shore, J. Am. Chern. Soc., 104, 1444 (1982).
7.) M. M. Taqui Khan, S. B. Halligudi, and S. Shukla, Angew. Chern Int, Ed Engl.,
'1:7, 1735 (1988).
9.) (a) D. J. Darensbourg and R. A. Kudaroski, Adv Orgauomet. Chern., 22, 129
(1983). (b) D. A. Palmer and R. Van Eldik, Chern. Rey., 83, 651 (1983). (c) D.
Walter, Coord Chern. Rey., 79, 135 (1987). (d) A. Behr, Angew, Chern. Int. Engl.
:M.... 27, 661 (1989).
10.) D. J. Darensbourg and C. Ovalles, J, Am, Chern Soc , 106, 3750 (1984).
11.) D. J. Darensbourg, R L. Gray, C. Ovalles, and M. Pala, J. MoL Catal., 29, 285
(1985).
12.) D. J. Darensbourg, R L. Gray, and C. Ovalles, J. Mol. Catal., 41, 329 (1987).
14.) D. J. Darensbourg, C. Ovalles, and M. Pala, J. Am· Chern. Soc .. 105, 5937 (1983).
16.) G. R Steinmetz, A. D. Harley, and G. L. Geoffroy, Inorg. Chern., 19, 2985 (1980).
28.) T. Bein and P. A. Jacobs, J. Chern. Soc .. Faraday Trans.. 79, 1819 (1983).
29.) N. Herron G. D. Stucky, C. A. Tolman, Inorg. Chim. Acta. 100, 135 (1985).
53
31.) J. J. Verdonck, P. A. Jacobs, M. Genet, and G. Poncelet, J. Chern. Soc. Faraday I.
76, 403 (1980).
54
REACTIVITY OF DINUCLEAR AND TETRANUCLEAR CLUSTERS OF MOLYBDENUM AND
TUNGSTEN
Malcolm H. Chisholm
Department of Chemistry
Indiana University
Bloomington, IN 47405
ABSTRACT
INTRODUCTION
56
Conversion of (M=M) 6 + to (MiM) 4+
(a)
(b) (1)
22"C
~
benzene
(2)
( 3)
57
22"C
benzene
(4)
22"C
(i)
benzene
( 5)
(ii)
o·c
hexane
...
(6)
58
atom transfer, reductive elimination of alkane and alkene followed by
oxidative addition of ButOH and insertion of alkene into a metal
hydride derived from the R'OH hydrogen atom. Support for this general
scheme came from the alcoholysis reaction shown in eq. 7 where the
Mo~Mo alkoxide was isolated and subsequently structurally
characterized. Presum~bly in reaction 7 the bulky alkene, Me 2C-CH 2
does not compete effect~vely for access to the metal center.
oac
(7)
hexane
(8)
22°C
-----1·~ W4(H)2(0Pri)14 +
hexane
(9)
59
undergoes successively (a) ,8-H atom transfer to give a W2 (JL-CH2 -NMe)
containing species, 45 (b) CH 2 -NMe bond cleavage to give CH 2 and NMe
ligands and (c) cluster condensation with further C-H bond activation
to yield the final product W4 (p. 4-C) (p.-NMe) (0Pri) 12 . In any event the
major product at short reaction times is the Lewis base adduct
W2 (0Pri) 6 (HNMe 2 ) 2 . At first we suspected that this might be an
intermediate in the formation of W4 (C)(NMe)(OPr 1 ) 12 but this was soon
ruled out. Furthermore, and to our great surprise, we found that
starting with pure W2 (0Pri) 6 (HNMe 2 ) 2 dissolved in.hexane or toluene and
adding Pr 1 0H failed to give yields of W4 (H) 2 (0Pr 1 ) 14 approaching those
obtained in reactions starting with W2 (NMe 2 ) 6 , eq. 9, that were allowed
to proceed for ~ 12 hours at room temperature.
(i)
(10)
(ii)
22°C
hexane
60
0(39~
Addition of Me 2 NH/Cl- to a solution of NaW 2 (H) (OPri) 8 generates
the tetratungsten dihydride according to eq . 12. 43
22°C
hexane/THF
(12)
W4 (H) 2 (OPri) 14
a-olefins ------------------------~..
~ internal olefins (13)
hexane/benzene, 22oc
61
hydrogenation occurs, though the details of latter reaction have not
yet been investigated in any detail. It is sufficient to note that the
system seems poised for catalysis based on reversible C-H, W-C and W-H
bond forming reactions.
The fact that M-M bonds are generally weaker than M-L bonds
renders dinuclear compounds with M-M multiple bonds susceptible to
cleavage by w-acid ligands such as C=O, RNC and NO. See eqs. 14 49 and
15.50
22•c
hexane
... (14)
M = Mo, W, R = But or Pr 1
22•c
(15)
hexane ...
22·c
MeOH, PF 6 -
... 2Mo(CNR)~+ (16)
-----1··
M=M and C=C bonds in the cleavage reaction shown in eq. 17. 52 • 53
o·c
hexane 2(But0) 3W=CR
( 17)
R- Me, Et, Pr 1
(18)
62
The Reductive Coupling and Selective Cross-Coupling of Ketones and
Aldehydes
22"C
(19)
hexane
63
Fig. 1. A ball-and-stick drawing o~ the W2 (0CH 2But) 6 (JL-CMe 2 )(0)(py)
molecule. W-W = 2. 705(1) A, W-O(oxo) = 1.684(4) A and W-
C(alkylidene) = 2.12(1) A (ave).
0°C
hexane •
(20)
64
listed in Table 1. In a typical reaction 300 to 500 mg of
W2 (0CH 2 But) 6 (py) 2 was allowed to react with the ketone (3 equiv) in
hexane at ca 22 •c for 24 h with stirring. Water was then added to
destroy the tungsten alkoxide and the aqueous layer was extracted with
ether and the organic product separated by column chromatography (Si0 2 ,
hexane/Et 2 0). The olefins were identified by NMR, mass spectrometry
Aldehyde/Ketone Olefin
51%
21%
o-CHO 3 34%
0< 4 44%
66%
5
b
6 36%
~b
~H H
7 18%
65
and melting points. The isolated yields given in Table 1 are not
optimized. From following the course of various reactions by NMR
spectroscopy employing 13 C labelled ketonic carbon atoms the idealized
yield should be very much higher in most cases.
66
bonding is favored in the W2 (11-CO) compounds relative to Mo 2 (11-CO)
compounds because the orbital energy separation is smaller for
tungsten.
0 o-
1
~ c
/"-..
M=M
+
M- '
/- M
22•c
~ W4 (C0) 2 (0R) 12 + ButOH
-,-----1...
hexane
(21)
(22)
(ii) -+
67
+ M=M -+
(23)
22"C
hexane
(24)
·~t\?l
l; c.·'w./
0 .·
r l '?l f' i·.
o::::::._~0 ·w ,./ o
j~'\??'w
wtz:J-w·/
68
carbonyl supported tetranuclear carbido clusters, e.g. Fe 4 (J.1 4 -
C)(C0)13.71 In this geometry the carbido ligand is strongly bonded to
the wingtip metal atoms but more weakly bonded to the two backbone
metal atoms of the butterfly. For a certain W4 (JJ- 4 -C) moiety lacking
any element of symmetry there are two large, W-wingtip-C, and two
small, W-backbone-C, values of 1 Jtasw_13c· For a W4 (C)(O)(OR) 12
molecule having virtual C2v symmetry only two values of 1 Jtasw_13c will
be seen, one large and one small, but the relative intensities of the
two satellites will be twice that observed for a W4 (C)(O)(OR) 12
molecule lacking any element of symmetry.
22"C
•
(25)
hexane
How can we be sure that this carbonyl C-0 bond has been cleaved?
Could the 13 C NMR data correspond to a W4 (JJ- 4 -CO) containing compound of
the type characterized by Shriver? 74 To address this problem Charles
Hammond 75 employed further 13 C NMR spectroscopic studies using a 60:40
tsctGo: tsctso mixture. Using high field and high resolution 13 C NMR
spectroscopy there is a readily measurable isotope chemical shift
effect. The 13 C resonances for the mixture of 13 C16 0: 13 C18 0 employed
in these studies are shown in Figure 3. The chemical shift separation
is 0.050 ppm. The compounds W2 (0But) 6 (JJ.-CO), W4 (JJ.-C0) 2 (DCH 2But) 12
and W4 (C) ( 0) ( OCH 2R) 12 were prepared from the 60:40 mixture of
tsctGQ;Isctso and their 13 C spectra recorded in the region of interest.
The presence of the C-0 bond is readily apparent in the spectra of
W2 (0But) 6 (J.1-CO) and W4 (JJ.-C0) 2 (0CH 2But) 12 but for the W4 (C)(O)(OCH 2R) 12
compounds the 13 C spectra appeared identical to those obtained earlier.
See Figures 2 and 4.
69
1 ~C( 18 0) = 0.048 ppm
~~-T-r,-,~~~-r•-r,-~~-T-r,-.-~~-r~
184.550 184.500
PPH
70
0
II
/c""
W=W
CONCLUDING REMARKS
ACKNOWLEDGMENTS
REFERENCES
71
2. It is no accident that in nature the Fe 4 S4 cluster in ferredoxin
serves as an electron transfer reservoir.
3. In the limiting case this can be taken to be a dinuclear early-
late transition metal containing complex and a considerable
expenditure of research effort has been recently made in this
area for the activation of molecules such as C=O: see C. P.
Casey, R. E. Palermo, and A. L. Rheingold, J. Am. Chem. Soc.
108:549 (1986); C. P. Casey, R. E. Palermo, R. F. Jordan, and
A. L. Rheingold, J. Am. Chem. Soc. 107:4597 (1985); C. P.
Casey, R. F. Jordan, and A. L. Rheingold, J. Am. Chem. Soc.
3:504 (1984).
4. J. T. Yates and M. R. Albert, in: "The Surface Scientists Guide
to Organometallic Chemistry," ACS Publishers, Washington,
D.C. (1987).
5. Sachtler coined the term "ensemble requirement" to explain results
obtained from alloy catalysts having an inactive component.
The ensemble requirement of a catalytic reaction is the
number of contiguous metal atoms needed for that specific
reaction to occur: W. M. H. Sachtler, p. 434, Chemtech, July
(1983). The term ensemble effect has also extended to
certain carbonyl cluster reactions: D. F. Shriver and M. J.
Sailor, Ace. Chem. Res. 21:374 (1988).
6. Y. Doi, K. Koshizuka, and T. Keii, Inorg. Chem. 21:2732 (1982).
7. G. SUss-Fink, J. Organomet. Chem. 193:C20 (1980).
8. G. SUss-Fink and J. Reiner, J. Mol. Catal. 16:231 (1982).
9. J. A. Smiejla, J. E. Gozum, and W. L. Gladfelter, Organometallics
5:2154 (1986).
10. G. SAs-Fink, Angew. Chem. Intl. Ed. Engl. 21:73 (1982).
11. R. L. Pruett and W. W. Walker, Union Carbide Corp. U.S. Patents 3
833 634 (1974), 3 957 857 (1976); J. L. Vidal, Z. C. Mester,
and W. Walker, Union Carbide Corp. U.S. Patent 4 115 428
(1978).
12. J. L. Vidal and R. C. Schoening, Inorg. Chem. 21:438 (1982).
13. G. C. Demitras and E. L. Muetterties, J. Am. Chem. Soc. 99:2796
(1977).
14. H.-K. Wang, H. W. Choi, and E. L. Muetterties, Inorg. Chem.
20:2661 (1981).
15. J. C. Bricker, C. C. Nagel, and S. G. Shore, J. Am. Chem. Soc.
104:1444 (1982).
16. J. C. Bricker, C. C. Nagel, A. A. Bhattacharayya, and S. G. Shore,
J. Am. Chem. Soc. 107:377 (1985).
17. M. W. Payne, D. L. Leussing, and S. G. Shore, J. Am. Chem. Soc.
109:617 (1987).
18. P. C. Ford, Ace. Chem. Res. 14:31 (1981).
19. R. D. Adams, H. Kim, and S. Wang, J. Am. Chem. Soc. 107:6107
(1985).
20. R. D. Adams, J. E. Babin, and M. Tasi, Inorg. Chem. 25:514 (1986).
21. See also Chapters by R. D. Adams and W. L. Gladfelter in this
volume.
22. For example see syngas and HDS catalysts derived from
Cp~Mo 2 Fe 2 S 2 (C0) 8 and Cp~Mo 2 Co 2 S 3 : M. D. Curtis, J. E.
Penner-Hahm, J. Schwank, 0. Baralt, D. J. McCabe, L.
Thompson, and G. Waldo, Polyhedron 7:2411 (1988).
23. For a novel low temperature route to MoW alloys see A. K.
Cheetham, Nature 288:469 (1980).
24. A. B. Brignole and F. A. Cotton, Inorg. Syn. 13:81 (1972).
25. M. H. Chisholm, D. A. Haitko, and C. A. Murillo, Inorg. Syn. 21:51
(1982).
26. M. H. Chisholm, D. A. Haitko, and J. C. Huffman, J. Am. Chem. Soc.
103:4046 (1981).
27. M. J. Chetcuti, M. H. Chisholm, K. F0 lting, D. A. Haitko, J. C.
Huffman, and J. Janos, J. Am. Chem. Soc. 105:1163 (1983).
72
28. M. H. Chisholm, M. J. Hampden-Smith, J. C. Huffman, and K. G.
Moodley, J. Am. Chem. Soc. 110:4070 (1988).
29. M. H. Chisholm, M. J. Hampden-Smith, K. A. Stahl, J. C. Huffman,
J. D. Martin, and K. G. Moodley, Polyhedron 7:1991 (1988).
30. M. H. Chisholm and M. W. Extine, J. Am. Chem. Soc. 99:782 (1977);
idem, ibid, 99:792 (1977).
31. M. J. Chetcuti, M. H. Chisholm, K. Folting, D. A. Haitko, and J.
C. Huffman, J. Am. Chem. Soc. 104:2138 (1982).
32. J. K. Kochi in "Organometallic Mechanisms and Catalysis,"
Academic Press, Chpt. 11 and references cited therein (1978).
33. M. J. Chetcuti, M. H. Chisholm, K. Folting, J. C. Huffman, and J.
Janos, J. Am. Chem. Soc. 104:4684 (1982).
34. M. J. Chetcuti, M. H. Chisholm, H. T. Chiu, and J. C. Huffman,
Polyhedron 4:1213 (1985); H. T. Chiu, Indiana University,
Ph.D. Thesis (1986).
35. M. H. Chisholm, H. T. Chiu, and J. C. Huffman, Polyhedron 3:475
(1984).
36. M. H. Chisholm, D. L. Clark, J. C. Huffman, and W. G. Van Der
Sluys, J. Am. Chem. Soc. 109:6817 (1987).
37. M. H. Chisholm, and R. J. Tatz, Organometallics 5:1590 (1986).
38. M. H. Chisholm, K. Folting, J. C. Huffman, and R. J. Tatz, J. Am.
Chem. Soc. 106:1153 (1984).
39. F. A. Cotton and R. A. Walton in "Multiple Bonds Between Metal
Atoms," Wiley Publishers, Chpt. 3, Sect. 3 .1. 5, pp. 99-105 and
references therein (1982).
40. M. H. Chisholm, C. C. Kirkpatrick, and J. C. Huffman, Inorg. Chem.
20:871 (1981).
41. M. Akiyama, M. H. Chisholm, F. A. Cotton, M. W. Extine, D. A.
Haitko, D. Little, and P. E. Fanwick, Inorg. Chem. 18:2266
(1979).
42. M. H. Chisholm, J. C. Huffman, C. C. Kirkpatrick, and J. Leonelli,
J. Am. Chem. Soc. 103:6093 (1981).
43. M. H. Chisholm, J. C. Huffman, and C. A. Smith, J. Am. Chem. Soc.
108:222 (1986).
44. M. H. Chisholm, D. L. Clark, J. C. Huffman, and C. A. Smith,
Organometallics 6:1280 (1987).
45. K. J. Ahmed, M. H. Chisholm, K. Folting, and J. C. Huffman, J. Am.
Chem. Soc. 108:989 (1986).
46. M. H. Chisholm, D. L. Clark, K. Folting, J. C. Huffman, and M. J.
Hampden-Smith, J. Am. Chem. Soc. 109:7750 (1987).
47. M. H. Chisholm, D. L. Clark, and M. J. Hampden-Smith, J. Am. Chem.
Soc. 110:574 (1989).
48. M. Akiyama, M. H. Chisholm, F. A. Cotton, M. W. Extine, D. A.
Haitko, D. Little, and J. Leonelli, J. Am. Chem. Soc. 103:779
(1981).
49. (a) M - Mo: M. H. Chisholm, F. A. Cotton, M. W. Extine, and R.
L. Kelly, J. Am. Chem. Soc. 100:3354 (1978).
(b) M- W: M. H. Chisholm, F. A. Cotton, M. W. Extine, and R.
L. Kelly, Inorg. Chem. 18:116 (1979).
50. M. H. Chisholm, F. A. Cotton, M. W. Extine, and R. L. Kelly, J.
Am. Chem. Soc. 101:7645 (1979).
51. R. A. Walton, ACS Symp. Ser. 155:207 (1981) and references
therein.
52. R. R. Schrock, M. L. Listemann, and L. G. Sturgeoff, J. Am. Chem.
Soc. 104:4291 (1982).
53. R. R. Schrock and M. L. Listemann, Organometallics 4:74 (1985).
54. M. H. Chisholm, D. M. Hoffman, and J. C. Huffman, J. Am. Chem.
Soc. 106:6794 (1984).
55. M. H. Chisholm, B. K. Conroy, B. W. Eichhorn, K. Folting, D. M.
Hoffman, J. C. Huffman, and N. S. Marchant, Polyhedron 6:783
(1987).
73
56. M. A. Harmer, T. R. Halbert, W. -H. Pan, C. L. Coyle, S. A. Cohen,
and E. I. Stiefel, Polyhedron, 5:341 (1986).
57. T. P. Blatchford, M. H. Chisholm, K. Folting, and J. C. Huffman,
J. Chern. Soc., Chern. Commun. 1295 (1984).
58. J. E. McMurry, Ace. Chern. Res. 16:405 (1982). A similar
reactivity is found in a reagent prepared by the reduction of
WC1 6 with LiBun: K. B. Sharpless, M. A. Umbreit, M. T. Nieh,
and T. C. Flood, J. Am. Chern. Soc. 94:6538 (1972). More
recently a variety of active metals of the actinides have
been employed as reagent for similar reductive couplings: B.
E. Kahn and R. T. Riecke, Chern. Rev. 88:733 (1988).
59. F. A. Cotton, R. A. Walton, D. DeMarco, and L. R. Falvello, J. Am.
Chern. Soc. 104:7375 (1982).
60. M. H. Chisholm and J. A. Klang, J. Am. Chern. Soc. llO:xxx (1989).
61. R. R. Schrock, J. Am. Chern. Soc. 98:5399 (1976).
62. T. C. Bryan and J. M. Meyer, J. Am. Chern. Soc. 109:7213 (1987).
63. M. H. Chisholm, D. M. Hoffman, and J. C. Huffman, Inorg. Chern.
22:2903 (1983).
64. M. H. Chisholm, J. C. Huffman, J. Leonelli, and I. P. Rothwell, J.
Am. Chern. Soc. 104:7030 (1982).
65. R. C. Benson, W. H. Flygare, M. Oda, and R. Breslow, J. Am. Chern.
Soc. 95:2772 (1973).
66. E. M. Kober and D. L. Lichtenberger, J. Am. Chem. Soc. 107:7199
(1985).
67. M. H. Chisholm, D. M. Hoffman, and J. C. Huffman, Organometallics
4:986 (1985). See also F. A. Cotton and W. Schwotzer, J. Am.
Chem. Soc. 105:4955 (1983).
68. D. F. Shriver and C. P. Horwitz, Adv. Organometal. Chem. 23:219
(1984).
69. M. H. Chisholm, Polyhedron 2:681 (1983).
70. P. J. Blower, M. H. Chisholm, D. L. Clark, and B. W. Eichhorn,
Organometallics 5:2125 (1986).
71. J. S. Bradley, G. B. Ansell, and E. W. Hill, J. Am. Chem. Soc.
101:7417 (1979).
72. M. H. Chisholm, K.· Felting, C. E. Hammond, J. C. Huffman, and M.
J. Hampden-Smith, J. Am. Chem. Soc. 110:3314 (1988).
73. M. H. Chisholm, C. E. Hammond, and M. J. Hampden-Smith, J. Am.
Chem. Soc., submitted.
74. E.g., as in Fe 4 (C0) 13 -: M. Manassero, M. Sansoni, and G. Longoni,
J. Chem. Soc. Chem. Commun. 919 (1976). See Ref. 5 for
reactivi.ty studies.
75. For a review for 13 C18 0 isotope shifts see: P. E. Hensen, Ann.
Rep. NHR Spectrosc. 15:105 (1983).
74
CLUSTERS AND THEIR IMPLICATIONS FOR CATALYIS
Richard D. Adams
Department of Chemistry
University of South Carolina
Columbia, SC 29208
INTRODUCTION
H H
'c
l
R R
c/
' /
u/ 'u
,(~\-/) M
M---M
DI
A B
R R
\ "
'
' /\"t' /
I
c--c ' Os--Os
I
-l~~:}Y\-
llO"C
-~~~'l\' (I)
,,;'
........
...... 'Y . . . .
-CO R
,c,l /c'
,' W R
H Cp
Cp
' / \ "
-2 co
''?f~'fJ
Os-
.' . . I, '
/ 'I __
'-- c,
- -Os
( 2)
_os~ v i s , Ph
I c
I
Ph
,...cu,tol
o--c
\
'os-1
~~W~p\
/1\. · llO"C (3)
"" I ". / '\ .,.... -2 co
-Os---os
I\ I'
76
,1/ R
Os 1
,\ /I"-. l_.c~H
Os Os~ \\ R
/1 /I ~c~
H
us•c j-co
\/ l
-os-c
,\/
Os
k~ 1\..-a
os- '
/I !\" H
R=H;
-CO
1
\
\co
R:alkyl
o._
...._I_.
//\"'-~
,\os:----..... /
___ o._
/'\,H,/"
c-c
a"' 'a
2 3
77
~pj) 0 I
)t::_
P,
'I
(4)
f /~I, so•c I/
/Ru:..-
:R~<. iu~ -c,n, '\ / Ru~
I Ru-
/ \ -Ru' \
/I'-
4 5
78
cluster. The carbon atom is bonded to one metal atom and the
nitrogen atom is bonded to the other. Interestingly, the C-N
vector is not parallel to the Os-Os bond, but it forms an angle
that brings the CH 2 group closer to the axial CO ligand on Os(l).
While the origin of this twist may be as simple as the
differences in steric interactions between the NMe 2 and CH 2 group
with the axial CO ligand on Os(l), the shift of the CH 2 group
toward Os(l) is believed to favor the transformation that occurs
when atom Os(l) is decarbonylated.
When heated to 97°C, compound 6 is transformed to two
products. 18 One of these is os 3 (C0) 9 [~ 3 -C(H)NMe 2 J(~-H), 7 formed
by the loss of CO and the activation of one of the CH bonds on
the CH 2 group. The second product is os 3 (C0) 10 (p-C=NMe 2 )(p-H), 8
formed by the activation of both CH bonds on the CH 2 group and
the elimination of one formula equivalent of H2 . ORTEP diagrams
of the molecular structures of 7 and 8 are shown in Figures 3 and
4, respectively.
In compound 8, the iminium ligand was transformed into a
triply bridging dimethylaminocarbene ligand. The carbon atom
bridges two of the osmium atoms while the nitrogen atom is
coordinated to the third osmium atom. The formation of 7 is
believed to occur by the loss of a CO ligand from the CO-rich
Os(C0) 4 group of the cluster and is followed by the activation of
one of the CH bonds of the CH 2 group at the resultant vacant
site, Eq. ( 5).
' v.····"··,~/
HH MeMe
Me Me
'le--N\/ H \/
'_.-N
' ,_
/o'~s/"- --"1\,. \----~05-
l~n~// -os-
(5)
H~
/,,Os
/i'
7
6
79
Figure 3. An ORTEP diagram of os 3 (C0) 9 [p 3-HCNMe 2 J(p-H) 2 , 7.
Reprinted from Organometallics 1988, 7, 963.
Copyright (1988) American Chemical Society.
80
the Os-N bond in 7 to yield a vacant site to permit cleavage of
one of the CH bonds of the CH 2 group and formation of a C-bridged
carbene ligand intermediate. This intermediate could be induced
to eliminate H2 by coordination of the nitrogen atom of the amino
group, see Scheme I. The mechanism of the activation of the
second CH bond of the CH 2 group is less clear, but could occur at
a vacant site on the third metal atom formed by cleavage of the
Os-N bond, and be followed by a shift of the dimethylaminocarbyne
ligand from a triply-bridging to an edge-bridging position.
These latter transformations would require appropriate
repositioning of some of the carbonyl ligands.
Scheme I
81
Figure 5. An ORTEP diagram of os 3 (CO)g[C(H)NMe 2 ](p-SPh)(p-H),
9. Reprinted with permission form J. Am. Chern. Soc.
1987, 109, 1414. Copyright (1987) American Chemical
Society.
82
9
83
process. A multicenter activation and shift process was proposed, 1 ~
see Scheme III. The cleavage of the carbon-sulfur bond is believe<
to occur by an oxidative addition to the third osmium atom, and
may resemble the phosphorus-carbon bond cleavage process observe'
by Carty for 4, Eq.(4). 16 Since 10 does not contain any vacant
coordination sites, and since the formation of 14 does not
require a ligand elimination, it is proposed that the s-c
oxidative addition is promoted by the cleavage of one metal-
metal bond. The unobserved intermediate c would contain a tripl:
bridging sulfide ligand and a sigma-bonded phenyl group, see
Scheme III. The shift of the phenyl group to the carbon atom of
the carbyne ligand could be driven by a reformation of the
cleaved metal-metal bond.
()
"'/VI
:s-Os-...........___ /
H\v !'
Os--C
/H?,oi
12s•c
/\ '\-R
I
R
10 c 14
Scheme III
84
C • H Activation in NEt3
\ HI NEt 2
' / II
Os-e-e
'1/,\/
os........_ /H 'u
,..., I -.. . . . . II
Os
/I'
Et 2N
' e-e\//
,
H
I_.._.. /,.....u'
I" __
{)s-
',.....os~
os
u-- /\'
17 16
Scheme IV
85
Figure 7. An ORTEP diagram of Os 3 (C0) 9 [p 3-HC 2NEt 2 J(p-
H)2, 17. Reprinted with permission of
Organometallics 1988, 7, 2241. Copyright (1988)
American Chemical Society.
(6)
86
In a related study, we found that the sulfur-containing
osmium cluster os 3 (co) 10 (p 3-s) reacts with NMe 3 at 125°C to yield
the new cluster complex os 3 (co) 8 [c(H)NMe 2 J(p 3-s)(p-H) 2 , 19 that
contains a terminally coordinated secondary dimethylaminocarbene
ligand. 24 It is structurally analogous to 14. The carbene
ligand was apparently formed by the activation of two CH bonds on
one of the methyl groups of an NMe 3 molecule. The hydrogen atoms
were transferred to the metal atoms to become bridging hydride
ligands. It was found that 19 will react both with secondary
amines and with tertiary amines to give exchange of the amino
group of the carbene ligand and form the new complexes
n
Os 3 (C0) 8 [C(H)NR 2 J(p 3-S)(p-H) 2 , R=Et, Pr . The reactions with
tertiary amines also yield NMe 2 R formed by transfer of one alkyl
group to the NMe 2 group of the carbene ligand. At 145°C,
solutions of 19 will produce tertiary amine metathesis
catalytically, but this catalysis is complicated by the presence
of the anion [os 3 (co) 9 (p 3 -S)(p-H)]-, 20, a decomposition product
derived from 19 by loss of the carbene ligand, loss of a proton
and the addition of one CO ligand. Our independent
investigations of 20, prepared by the reaction of tertiary amine
with os 3 (co) 9 (p 3-s( )p-H) 2 have shown that it is also an active
catalyst for tertiary amine metathesis. Studies of catalysis
using 19 and 20 conducted in co 3oo solvent have shown deuterium
incorporation into the alkyl groups of the mixed amine products
and imply that metal-induced CH activation processes accompany
the alkyl exchange reaction. Further studies are in progress.
Acknowledgement
These studies were supported by the u.s. Department of
Energy under Grant No. DEFG84ER13296.
References
87
6) Ugo, R.; Psara, P. J. Mol. Catal. 1983, 20, 53.
7) Ford, P. c. Accts. Chern. Res. 1981, 14, 31.
8) Mark6, L.; Vizi-Orosz, A. in "Metal Clusters in Catalysis",
Gates, B. C.; Guczi, L.; KnOzinger, eds. Elsevier,
Amsterdam, 1987, Ch.S.
9) SUss-Fink, G. Polyhedron 1988, 7, 2341.
10) Park, J. T.; Shapley, J. R.; Churchill, M. R.; Bueno, c. J.
Am. Chem. Soc. 1983, 105, 6182.
11) Fernandez, J. M.; Johnson, B. F. G.; Lewis, J.; Raithby, P.
R. Acta Cryst. 1978, B34, 3086.
12) Shapley, J. R.; Strickland, D. S.; St. George, G. M.;
Ziller, J. W.; Beanan, L. R. J. Am. Chern. Soc. 1984, 106,
1144.
13) a) Ponec, v. Cat. Rev. Sci. Eng. 1978, 18, 151.
b) Herrmann, w. A. Angew Chern. Int. ·Engl. 1982, 21, 117.
c) Henrici-Olive, G.; olive, s. Angew. Chern. Int. Engl.
1976, 15, 136.
d) Muetterties, E. L. Chern. Rev. 1979, 79, 479.
14) a) Bergman, R. G-i Seidler, P. F •; Wenzel, T. T. J. Am.
Chern. Soc. 1985, 107, 4358.
b) Stoutland, P. 0.; Bergman, R. G. J. Am. Chern. Soc.
1985, 107, 4581.
c) Silvestre, J.; Calhorda, M. J.; Hoffmann, R.;
Stoutland, P. 0.; Bergman, R. G. Organometallics 1986,
5, 1841.
d) Boncella, J. M.; Green, M. L. H. J. Organomet. Chern.
1987, 325, 217.
15) a) Johnson, B. F. G.; Lewis, J. Inorg. Chim. Acta 1979,
109, 271.
b) Deeming, A. J. in "Transition Metal Clusters", Johnson,
B. F. G., ed., Wiley & Sons, Chichester, 1980, Ch. 6.
16) MacLaughlin, S. A.; Carty, A. J.; Taylor, N. J. Can. J.
Chern. 1982, 60, 87.
17) Adams, R. D.; Babin, J. E.; Kim, H. s. Polyhedron 1988, 7,
967.
18) Adams, R. D.; Babin, J. E. Organometallics 1988, 7, 1963.
19) Adams, R. D.; Babin, J. E.; Kim, H. s. J. Am. Chern. Soc.
1987, 109, 1414.
20) Shapley, J. R.; Tachikawa, M.; Churchill, M. R., Lashewycz,
R. A. J. Organomet. Chern. 1978, 162, C39.
21) Adams, R. D.; Tanner, J. T. Organometallics 1988, 7, 2241.
88
22) a) Raithby, P. R.; Rosales, M. J. Adv. Inorg. Radiochem.
1985, 29, 169.
b) Sappa, E.; Tiripicchio, A.; Braunstein, P. Chern. Rev.
1983, 83, 203.
23) a) Shvo, Y.; Laine, R. D. J. Chern. Soc., Chern. Commun.
1980, 753.
b) Wilson, R. B. Jr.; Laine, R. M. J. Am. Chern. Soc. 1985,
107, 361.
24) Adams, R. D.; Kim, H. S.; Wang, s. J. Am. Chern. Soc. 1985,
107, 6107.
89
METAL CLUSTERS IN THE SOLID STATE
Robert E. McCarley
INTRODUCTION
In the chain structure it is essential that the van der Waals radius
of oxygen be comparable to the metallic single bond radius 17 of Mo since
the M-M and 0-0 spacings must be the same along the chain. If the
required anion spacing is too large to permit good M-M bond formation
along the chain, then break-down of the chain structure would be expected.
for Ta and W, and ~G">O for Nb and Mo. Cheetham and his coworkerslB
discuss this problem with reference to their lack of success in the
preparation of Zn 2w3o8 , the tungsten analogue of the well-known zn 2Mo 3o8
structure.
92
Table 1. Structure Types with Infinite Chains of
Condensed Octahedral Cluster Units
Known Examples
M - Li, Zn 14.5-15.0
M - Y, Lnc 14.0
x- 5.45, M- Ca
The best known and most abundant examples of M-M bonded oxide systems
are the several structure types involving infinite chains of trans edge-
shared octahedral metal cluster units. These are listed in Table I, and
views of the construction of the infinite chain and oxide framework are
shown in Figure 1. Basically in each structure type the molybdenum oxide
chains are interlinked in a different way through Mo-0-Mo bonding to
create the particular oxide framework. Within each framework the ternary
metal ions are located in 0-coordinated sites of geometry best suited for
the particular metal ion. For example, the metal ions in the ~Mo 4 o 6
93
a
94
~[Mo 4 o 6 ] chains, but the geometry of the coordination varies from cubic 8-
coordinate (M-Na, K, Ba), 19 • 20 to highly distorted cubic 8-coordinate
(Pb), 21 to square pyramidal 4 coordinate (In), 22 to square planar (Sn). 23
Only for M-Na, K and In are all of these ternary metal sites filled to
provide the stoichiometry MxMo 4o6 with x-1. When M-Ba, Pb and Sn, all
divalent metal ions, vacancies occur in the ternary metal sites with
resulting compositions x-0.62, 0.75 and 0.9, respectively. In the case of
the Ba 20 and Pb 21 compounds the vacancies are ordered in super cells
having 8 times and 4 times the dimension of NaMo 4o6 in the unique
direction (c-axis). Both size of the ternary metal ions and total
electrons transferred to the Mo 4o6 chains appear to be involved in this
nonstoichiometry. Cation-cation repulsions between the large divalent
ions in the fully loaded structure and instabilities resulting from
transfer of 2 electronsfMo 4o6 unit favor lower x-values in the observed
order.
95
Fig. 2. Two views of the condensed octahedral cluster
chains as found in Gd4Mo 4 o11 . The top view shows
the pairwise mating of the Mo atoms along the chain
direction and the bottom view shows the alternating
long and short bonds on the shared edges. Only Mo
atoms are shown in these views.
The discovery of the compound In 11Mo 40 o62 and the elucidation of its
fascinating structure 26 brought with it the realization that it is
possible to dissect the chains of the type shown in Fig. 1 and thereby
obtain discrete oligomeric cluster units. In the wonderful example of
this compound, not one, but two oligomeric clusters were displayed in the
same structure, and as a bonus two different oligomeric I~+ chains were
also found! Both kinds of oligomeric units are represented in the
formulation (Ins7+) (In 6B+>< Mo 18 o28
7->< Mo 22 oa-
34 ). The linear cation chains can
be understood as constructed from n-2 internal In+ atoms and two terminal
In 2+ atoms to give the oligomeric general.formula In~n+ 2 )+. Presumably
the I~n+ 2 )+ cations are selected to stabilize this structure because they
almost perfectly fill the spaces between the anionic cluster units and
provide just the needed MCE count by electron transfer to the anions. A
representation of this structure illustrating only the Mo 18 , Mo 22 , Ins and
In 6 units is shown in Fig. 3.
96
Fig. 3. The structure of In11Mo 40 o62 with oxygen atoms
omitted. Molybdenum atoms are shown as filled
circles and indium atoms as open circles. Note the
presence of the In 6 chains between the Mo 22
clusters and In5 chains between the Mo 18 cluster
units.
Remarkably, within the last year two papers have appeared which
report the synthesis and structure of two additional members of this
oligomeric series. The first of these, 28 BaMo 6o10 , represents not only a
member with n-1, but also the first ternary oxide having individual
octahedral clusters, of the Mo 6o12 type, as the essential structural unit.
Based on other known clusters of the M6x12 type we might expect that the
MCE of the cluster unit in BaMo 6o10 would be 14 to 16. 7 The observed MCE
of 18 can be rationalized on the basis that additional electrons are
needed to form the intercluster Mo-Mo bonds of low bond order, as shown in
Fig. 4. By using the bond distances given by Wang, Wang and Lii, 28 bond
order sums 9 can be used to estimate that ca. 14.5 electrGns participate in
intracluster Mo-Mo bonding, and £A. 3.5 electrons compose the intercluster
Mo-Mo bonds. As illustrated in Fig. 4, the intercluster bonds knit the
97
Fig. 4. Representation of Mo-Mo bonding in and between
octahedral cluster units of BaMo 6o10 . Only Mo
atoms are illustrated. Double lines represent
intracluster bonds and single lines represent
intercluster bonds. Closed circles at one level
and open circles at level above or below.
Mo 6o10 units together to form infinite chains which run parallel to the
orthorhombic b-axis. Because of the intercluster Mo-Mo bonding we might
expect that this compound will exhibit low resistivity in the b direction,
perhaps even metallic character.
98
Fig. 5. 3- cluster anion found in
Structure of the Mo 8o14
NdMo 8o14 . Heavy filled lines represent Mo-Mo
bonds and unfilled lines represent Mo-O bonds .
The Mo 8 cluster core is shown with 24 ligating
oxygen atoms derived from sharing of oxygen atoms
between cluster units in the crystal lattice.
99
Although the Mo 8 cluster unit may be viewed in a number of ways, e.g.
octahedron plus two face capping atoms or two rhomboidal cluster units
placed one on top of the other, the most fruitful view, we believe, is to
consider it as two Mo 4 butterfly cluster units coupled together front to
back. The butterfly units can be thought of as the repeat units in the
infinite chains like those shown in Fig. 1 for the trans edge-shared mode
of condensation of octahedral clusters. If two such butterfly repeat
units can be condensed to form a discrete cluster unit, we might expect
that it also should be possible to couple 3 or more units together in the
same way. Thus we have the basis for a new oligomeric series (Mo 4 )n
formed from the coupled butterfly units. NdMo 8o14 is therefore a member
with n-2 of a possible oligomeric series with the general formula
~M 0 4n°6n+2"
The Mo-Mo bond distances in the Mo 8 cluster range from 2.66 to 2.84A
with the average of 2.726A. From the formula we deduce that there should
be 23 electrons available for Mo-Mo bonding. From the Pauling bond order
sum for all Mo-Mo bonds in the cluster unit, ~n- 12.1, we calculate an
average bond order n (ave.) - 12.1/18- 0.67 for the 18 Mo-Mo intracluster
bonds. This compares favorably with n (ave.) - 23/36- 0.64 calculated
for 23 electrons distributed over 18 bonds. This comparison indicates all
23 cluster electrons are indeed involved in Mo-Mo bonding. However it
does not appear that the odd number of cluster electrons gives rise to a
magnetic moment residing on the cluster anions. magnetic susceptibility
measurements indicate that the magnetic moment of NdMo 8o14 , 3.29~B· is
derived solely from the Nd3+, with an expected moment of 3.62~B calculated
from the ground state 4 1 912 term. We conclude therefore that pairing of
the odd electrons must take place by coupling between cluster units. The
nearest intercluster Mo-Mo distance of 3.059(2)A, observed between a
cluster and each of two neighbors, signals electron delocalization and
pairing by band formation as the most likely mode of unpaired electron
coupling.
100
REFERENCES -
101
25. K.-H. Lii, R.E. McCarley, S. Kim and R.A. Jacobson, J. Solid State
Chern. 64, 347 (1986).
26. H. Mattausch, A. Simon, E.-M. Peters, Inorg. Chern. 25, 3428
(1986).
27. R.A. Wheeler and R. Hoffman, J. Am. Chern. Soc. 110, 7315 (1988).
28. S.L. Wang, C.C. Wang and K.-H. Lii, J. Solid State Chern. 11. 407
(1988).
29. S.J. Hibble, A.K. Cheetham, A.R.L. Bogle, H.R. Wakerley and D.E.
Cox, J. Am. Chern. Soc. 110, 3295 (1988).
30. P. Gougeon, C.D. Carlson and R.E. McCarley, to be published.
102
NATURE OF BIMETALLIC CLUSTERS
J.H. Sinfelt
INTRODUCTION
104
catalytic activity of a binary alloy compr~s~ng a combination of metals
from groups VIII and IB is markedly lower than the activity of the Group
VIII metal alone (21-26). In contrast, the catalytic activity of a
Group VIII metal for reactions of carbon-hydrogen bonds, i.e.,
hydrogenation and dehydrogenation reactions, is affected relatively
little when it is alloyed with a Group IB metal. In fact, the activity
for such reactions may actually increase (22, 23). The discovery of
this selectivity phenomenon stimulated much interest in bimetallic
catalyst systems in the early 1970s.
105
that the component of the alloy with the greater affinity for the gas
may be drawn to the surface (36).
106
Ru-Cu
22 22.5 23 23.5
ENERGY, KeV
K ~ (2mE)l/2j b (1)
107
0.4
Ru-cu
2o-.------.,
0.2
"">< 0.0
"'
0 10
"" X
-0.2 :;
C<:
0
LJ..
Vl
-0.4 z
<[
C<:
1--
2o-.-------.
LJ..
0
0.2 w
0
=>
1-
~ z
..... 0.0
l:)
<[ 10
>< :;
""
-0.2 v
0
2 4 b 8 10
K, _A-1
12 14 1b 0 2
.
R, A
4 b
108
bimetallic clusters, including the use of data on appropriate reference
materials, can be found in our early papers (43, 44).
109
0.50 Cu
0.0
-0.50
-
>l
.....
':- 0.4
"' Cu-Ru
0.2
0.0
-0.2
-0 .44---.--,.......,......,..--.-"'T"'...,.....T""""""'~~--.-...,.....-j
2 4 6 8 10 12 14 16
K, A-1
Fig. 3. Comparison of the copper EXAFS
(K absorption edge) of copper
and ruthenium-copper clusters
at lOOK. The EXAFS shown is
the part due to nearest
neighbor metal atoms (~).
110
in the clusters. Consequently, with a 1:1 atomic ratio of copper to the
other component, the surface could not have consisted solely of copper.
The Re-Cu clusters presented an extreme case in which the average
coordination number of the atoms was only four. Such a low value implies
that the clusters were extremely small, and one can readily visualize
structures in which all of the atoms are surface atoms.
CONCLUDING REMARKS
REFERENCES
1. Sinfelt, J.H. Chern. and Eng. News 1972 (July 3 issue), 50, 18.
2. Sinfelt, J.H. J. Catalysis 1973, 29, 308.
3. Sinfelt, J.H., Ace. Chern. Res. 1977, 10, 15.
4. Sinfelt, J.H. Bimetallic Catalysts: Discoveries. Concepts. and
Applications; Yiley: New York, 1983.
5. Sinfelt, J.H. Ace. Chern. Res. 1987, 20, 134.
6. Sinfelt, J.H. U.S. Patent 3.953.368; 1976.
7. Cotton, F.A. Inorg. Chern. 1965, ~. 334.
8. Spenadel, L.; Boudart, M. J. Phys. Chern. 1960, 64, 204.
9. Sinfelt, J.H.; Yates, D.J.C. J. Catalysis 1967, ~. 82.
10. Yates, D.J.C.; Sinfelt, J.H. J. Catalysis 1967, ~. 348.
11. Yilson, G.R.; Hall, Y.K. J. Catalysis 1970, 17, 190.
12. Sinfelt, J.H. Annual Review of Materials Science 1972,z. 641.
13. Hildebrand, J.H.; Scott, R.L. The Solubility of Nonelectrolytes, 3rd
ed.; Reinhold: New York, 1950; p.335.
14. Schwab, G.M. Disc. Faraday Soc. 1950, ~. 166.
15. Dowden, D.A.; Reynolds, P. Disc. Faraday Soc. 1950, ~. 184.
16. Couper, A.; Eley, D.D. Disc. Faraday Soc. 1950, ~. 172.
17. Best, R.J.; Russell, W.W. J. Amer. Chern. Soc. 1954, 76, 838.
18. Hall, ~.K.; Emmett, P.H. J. Phys. Chern. 1959, 63, 1102.
19. Hall, W.K.; Emmett, P.H. J. Pbys. Chern. 1958, 62, 816.
20. Mott, N.F.; Jones, H. Theory of the Properties of Metals and Alloys;
Oxford University Press: Oxford, 1936.
21. Sinfelt, J.H.; Barnett, A.E.; Dembinski, G.W. U.S. Patent 3,442.973;
1969.
111
22. Sinfelt, J.H.; Barnett, A.E.; Carter, J.L. U.S. Patent 3,617,518;
1971.
23. Sinfelt, J.H.; Carter, J.L.; Yates, D.J.C. J. Catalysis 1972, 24,
283.
24. Sinfelt, J.H.; Lam, Y.L.; Cusumano, J.A.; Barnett, A.E. J. Catalysis
1976, 42, 227.
25. Ponec, V.; Sachtler, W.M.H. J. Catalysis 1972, 24, 250.
26. Beelen, J.M.; Ponec, V.; Sachtler, W.M.H. J. Catalysis 1973, 28,
376.
27. Balandin, A.A. Z. Physik Chern. 1929, B2, 289.
28. Balandin, A.A. Z. Physik. Chern. 1929, B3, 167.
29. Boudart, M. Chern. Technol. 1974, ~. 748.
30. Dowden, D.A. Proceedings of the Fifth International Congress on
Catalysis, Vol. 1; North-Holland: Amsterdam, 1973, p. 621-631.
31. Sachtler, W.M.H.; Van Santen, R.A. Advances in Catalysis 1977, 26,
69.
32. van der Plank, P.; Sachtler, W.M.H. J. Catalysis 1967, l. 300.
33. Cadenhead, D.A.; Wagner, N.J. J. Phys. Chern. 1968, 72, 2775.
34. Williams, F.L.; Boudart, M. J. Catalysis 1973, 30, 438.
35. Bouwman, R.; Lippits, G.J.M.; Sachtler, W.M.H. J. Catalysis 1972,
25, 350.
36. Sinfelt, J.H. Progress in Solid-State Chemistry 1975, 10(2), 55.
37. Hansen, M. Constitution of Binary Alloys, 2nd. ed.; McGraw-Hill:
New York, 1958; pp. 607, 619, 620.
38. Sayers, D.E.; Lytle, F.W.; Stern, E. Phys. Rev. Lett. 1971, 27,
1204.
39. Lytle, F.W.; Sayers, D.; Stern, E. Phys. Rev. B: Solid State 1975,
11, 4825.
40. Stern, E.; Sayers D.; Lytle, F.W. Phys. Rev. B: Solid State
975, 11, 4836.
41. Kincaid, B.M.; Eisenberger, P. Phys. Rev. Lett. 1975, 34, 1361.
42. Lytle, F.W.; Via, G.H.; Sinfelt, J.H. Am. Chern. Soc .. Div. Petr.
Chern. Preprints 1976, 21(2), 366.
43. Sinfelt, J.H.; Via, G.H.; Lytle, F.W. J. Chern. Phys. 1980, 72, 4832.
44. Sinfelt, J.H.; Via, G.H.; Lytle, F.W.; Greegor, R.B. J. Chern. Phys.
1981, 75, 5527.
45. Sinfelt, J.H.; Via, G.H.; Lytle, F.W. J Chern. Phys. 1982, 76, 2779.
46. Meitzner, G.; Via, G.H.; Lytle, F.W.; Sinfelt, J.H. J. Chern. Phys.
1983, 78, 882.
47. Meitzner, G.; Via, G.H.; Lytle, F.W.; Sinfelt, J.H. J. Chern. Phys,
983, 78, 2533.
48. Meitzner, G.; Via, G.H.; Lytle, F.W.; Sinfelt, J.H. J. Chern. Phys,
1985, 83, 353.
49. Meitzner, G.; Via, G.H.; Lytle, F.W.; Sinfelt, J.H. J. Chern. Phys,
1985, 83, 4793.
50. Meitzner, G.; Via, G.H.; Lytle, F.W.; Sinfelt, J.H. J. Chern. Phys.
1987, 87, 6354.
51. Cullity, B.D. Elements of X-ray Diffraction; Addison-Wesley:
Reading, Massachusetts, 1956; pp. 370-372.
52. Schneider, A.; Esch, U. Z. Elektrochem 1944, 50, 290.
53. Stockdale, D. J. Inst. Metals 1930, 43, 193.
54. Betz, F.; Arias, M.; Braun, P. Nucl. Instrum. Methods 1980, 170,
347. •
55. Knock, B.; Star, W.M.; van Rongen, H.J.M.; van den Berg, G.J.
Physica (Amsterdam) 1964, 30, 1124.
56. Raube, E.; Roschel, E. Z. Metallkd. 1969, 60, 142.
112
THERMOCHEMICAL ASPECTS OF ORGANOTRANSITION METAL CHEMISTRY. INSIGHTS
Department of Chemistry
Northwestern University
Evanston, IL 60208-3113
INTRODUCTION
Ln = ancillary ligands
For a given LnM array and X ligand, it can be seen that the approach
of eqs.(3)-(6) yields relative D(LnM-R) values which are anchoredlO to the
corresponding D(LnM-X) parameters. While such information is adequate for
analyzing numerous LnM-R ~ LnM-R' transformations, it would also be desir-
able to obtain absolute values of D(LnM-R) to understand both M and Ln
effects. An approximate approach has been to equate D(LnM-X) values to
the corresponding Dl(MXm) parameters where M is in the same formal oxida-
tion state.6,8-10 While this transferability approximation seems reason-
able, it would be of great interest to quantify the effects of the
ancillary ligand array, Ln· A more rigorous approach to determining
absolute D(LnM-R) values is possible when LnM/LnM-R pairs are available as
shown in eqs.(7)-(ll).ll-13 Such studies have recently been carried out
114
Table 1. Bond Disruption Enthalpy Data for Organo-f-Element
Complexes and the Corresponding Homoleptic Halides
in Kcal/Mol
aReference 12
bReference 13
cReference 11
dReferences 14,15
eReference 16
• Cp',SmX
0 x,smx
Cl Br
Fig. 1. Comparison of measured D(Cp2Sm-halogen) values to the first bond
dissociation energies (refs. 14,15) of the corresponding
samarium trihalides.
115
0 Cp' 2 ln-l
0
s
'-..
60
«i
~ " 40
,....,.
I
ep 20
0
Sm Eu Yb
135
CIA
o'su
125
•
0
s 115
'til Br
•
"
.!< 105 aS"Pr
H
xI 95 • I
;:::
cp NMe 2
• •
85
R
75
40 50 60 70 80 90 100
D(Sm-X), kcaJ/mol
Fig.3, Comparison of measured D(Cp2Sm-X) data (ref.l2) to those of the
corresponding Cp2Th(X)-X series (refs. 8,12 using an approximate
thermodynamic anchor). For Sm, R ~ CH(SiMe3)2 while forTh, R ~
Me. The X= H and snpr data for Th are averages of D[Cp2Th(X)-S]
and D(Cp2Th(OtBu)-X] data, and the X= halide data are from
D1(ThX4) (ref. 16).
116
100
80
0
8
'-
a; 60
C)
.:.:
><I 40
~
0 20
D(Sm-X), kcal/mol
115 Cl•
0 105
8
'-
a;
0
95
."'
><I 85
...
N
q 75
65
40 50 60 70 80 90 100
D{Sm-X), kcal/mol
117
ligands. It can be seen that bonding patterns in these archetypical
early transition element, actinide, and lanthanide systems, having rather
different steric environments about the metal centers, are strikingly
similar.
Two gauges are readily available to help judge whether the above
similarities in metal-ligand bonding patterns extend to middle and late
transition metals. The first approach is to compare key D(M-X) - D(M-X')
parameters as a function of position in the Periodic Table.ll Shown in
Figures 6 and 7 are comparisons of the parameters D(M-H)-D(M-CH3)11-22 and
D(M-I)-D(M-CH3)17-22, respectively. The former quantity is pivotal in
determining the course of p-hydride elimination and C-H functionalization
reactions while the latter plays an important role in a variety of
oxidative addition/reductive elimination and metal-halogen interchange
processes. It can be seen that D(M-H)-D(M-CH3) rises almost monotonically
as the Periodic Table is traversed to the right while D(M-I)-D(M-CH3)
falls almost monotonically. Clearly, these bonding patterns vary greatly
and, to an appreciable extent, systematically with position in the
Periodic Table.
While portions of the above trends have been explained on the basis
of promotion energies,23 bond polarity effects 24 and orbital repulsion
effects,25 an equally appealing and complementary qualitative explanation
can be derived from electronegativity24,26-29 considerations. In
particular, the perturbation theory-based formalism of Matcha30 (eq.(l2))
30
Pal'tX.
0 25
E
:::::..
0
_,.u 20
.-::.
:I:
u
.
I 15
3Cl
10
.......
:I:
I
::IE
c 5
118
(12)
25
cp".smx
0 20
E Cp' 1Th(OR)X
::::-0 Cp',U(OR)X
0
.Y.
15 Cp' 1 ZrX, Cp 1loloX,
,...::. Cp',H!X, Cp,WX,
~
:I:
u
I
:::; loln(CO),X
0 10
Cp'lrPX,
-::::-
1
~ 5
0
119
12
0
E
~
c 8
u
.Yo
,......
:I:
.. 4
' '
(J
I
= 2~~--- -
:::::!l 0
a XcH,
-4 From Matcha Formalism
I'.
I
a:::::!l -8
L,,-~-r,-r,-ro~~-r~-r,-r,-r
1.20 1.45 1.70 1.95 2.20
xt.A
Fig. 8. Correlation between D(M-H) - D(M-CH3) and XM calculated using
eq.(l3) and metal Pauling electronegativities.
0 20
E
.........
0 16
u
~
..-....., 12
I
(J
I 8
:::::!:
.........
Q
4 From Matcha Formalism
I 0
:::::!:
......... 1.20 1.45 1.70 1.95 2.20
Q
xt.A
Fig. 9. Correlation between D(M-1) - D(M-CH3) and XM calcuated using an
expression analogous to eq.(l3) (footnote 31) and metal Pauling
electronegativities. XcH 3 is taken to be 2.3.
120
D(M-X) - D(H-X) + l/2[D(M 2 )-D(H 2 )] + 105(e-0· 219 (XH-XX) 2
-e
-0.219(xwxx)2 ) (14)
(15)
130
0 120
s
'-..
<0 110
~
u
.
H
100
xI CH(TMS) 2
* NMe
ES
Q
90
*
2
1r-C3Hs
BO
PEt,
70
30 40 50 60 70 80 90 100
D(Sm-X), kcal/mol
121
Table 2. Product and Turnover Frequency Data for Organolanthanide-
Catalyzed Hydroaminations/Cyclizations.
1,1
1 H 2N~
z. '6 1 13(25°C)
140(60°C)
1a
la
1,1
2 H2N~ :!.
'+ .8. 125(25°C)
75(80°C)
<1(80°C)
la
lb
lc
H
--(:}
NH2
3 ~ ! .2. 84(25°C) la
4 H2N~
s. l) lJl 5(60°C) la
5 ():)--- l l la
~I
(16)
122
Preliminary mechanistic data, in addition to the aforementioned ggm-
dimethyl effect, are in accord with the scenario in Figure 11 where olefin
insertionjcyclization (1, eq.(lS)) is, i~ most cases, turnover-limiting.
Kinetically, the hydroaminations in Table 2, entries 1, 2, and 4, are
first-order in organolanthanide and zero-order in amino olefin over the
entire course of the reactions (i.e., Nt is independent of olefin and
lanthanide concentration), implying that proton transfer (ii, eq.(l6)) is
the rapid step (as expected36). Also in agreement with this picture is
the relative ordering of catalyst activities for 3~8: (Cp2LaH)2 (la) >
[Me2Si(Me4Cs)2LuHl2 (lb) > (cp 2LuH)2 (lc) (Table 2)--identical to the
p!eviously reported ordering for catalytic propylene oligomerization
activity.33-36 The outcome of the isotopic labelling experiment 3-d2 ~
8-d2 (eq.(l7)) further supports the proposed mechanism, revealing exactly
D
la
(17)
CONCLUSIONS
-{J
Cp'zl.aU H
123
aiding in the design of new types of transformations. While the emphasis
has been on lanthanide, actinide, and early transition element chemistry,
the broad generality of the approFches discussed should be evident.
ACKNOWLEDGMENTS
REFERENCES
124
27. J. E. Huheey, "Inorganic Chemistry," 3rd ed., Harper and Row, New
York (1983) pp. 144-160.
28. B. E. Douglas, D. H. McDaniel, and J. J. Alexander, "Concepts and
Models of Inorganic Chemistry," 2nd ed., Wiley, New York (1983)
Chapt. 24.
29. J. Mullay, Struct. Bond (Berlin) 66:1 (1987), and references therein.
30. R. L. Matcha, J. Am. Chern. Soc. 105:4859 (1983). We utilize the
formulation having an arithmetic mean expression for D(A-B) since
this leads more straightforwardly to useful relationships such as
eq.(l2). Both arithmetic and geometric mean approaches have been
employed in the Pauling formulation,26-28 with the latter preferred
in cases of sign ambiguities.
31. Because D(I2)-D(Me2) is negative (a common artifact when the
arithmetic mean is used in eq.(l2)),28 we arbitrarily set it equal to
zero in the expression analogous to eq.(l3). This in no way affects
the XM dependence of the function.
32. H. E. Bryndza, L. K. Fong, R. A. Paciello, W. Tam, and J. E. Bercaw,
J. Am. Chern. Soc. 109:1444 (1987).
33. H. Mauermann and T. J. Marks, Organometallics 4:200 (1985).
34. G. Jeske, H. Lauke, H. Mauermann, P. N. Swepston, H. Schumann, and T.
J. Marks, J. Am. Chern. Soc. 107:8091 (1985).
35. G. Jeske, L. E. Schock, H. Mauermann, P. N. Swepston, H. Schumann,
and T. J. Marks, J. Am. Chern. Soc. 107:8103 (1985).
36. G. Jeske, H. Lauke, H. Mauermann, H. Schumann, and T. J. Marks, ~
Am. Chern. Soc. 107:8111 (1985).
37. Cp2LaNMe2 catalyzes ethylene polymerization (D. Hedden and T. J.
Marks, unpublished results).
38. P. J. Fagan, J. M. Manriquez, C. H. Vollmer, C. S. Day, V. Day, and
T. J. Marks, J. Am. Chern. Soc. 103:2206 (1981).
39. For the addition of NH3 to H2C=CH2, fiG 0 ~ -4 kcal/mol.
40. M. R. Gagn~ and T. J. Marks, J. Am. Chern. Soc., in press.
41. Y. Tamaru, M. Hojo, H. Higashima, and Z. Yoshida, J. Am. Chern. Soc.
110:3994 (1988), and references therein.
42. J. P. Collman, L. G. Hegedus, J. R. Norton, and R. G. Finke,
"Principles and Applications of Organotransition Metal Chemistry,"
University Science Books, Mill Valley, CA (1987), Chapts. 7.4, 17.1.
43. A. L. Casalnuovo, J. C. Calabrese, and D. Milstein, J. Am. Chern. Soc.
110:6738 (1988).
44. G. Pez, and J. E. Galle, Pure Appl. Chern. 57:1917 (1985).
125
METAL CLUSTERS AND SUPPORTED METAL CATALYSTS
B. c. Gates
Center for Catalytic Science and Technology
Dept. of Chemical Engineering
University of Delaware
Newark, Delaware 19716
INTRODUCTION
128
supports, which allow the formation of anionic carbonyl
clusters. 1 0 The surface organometallic chemistry of formation
of Os and Ru clusters is analogous to the reductive
carbonylation of halide complexes such as [PtC1612- in basic
solution.
129
w
0
Precurs or Prepare d by
Adsorpt ion of [Os 3 (COh2 l co 2-
[0s5C(C0 )14] /MgO
on MgO 400 275°C, 1 atm
4 h
C, 1 atm
He 150°C 4 h
275°C, 1 atrn 2 h
2 h
A ~
[H 3 Os 4 (C0) 12 ]-/MgO
Precurs or Prepare d CO or
from CO + H2
H2 0sC16+ Mg0400 {0Mg} 2 0s (CO) x CO + H2 , 275°C, 11 atrn
200-250° C
1 atrn x = 2 and 3 12 h
131
Top View Side View
2.978 A
~
® Re 10n
MgO 0 surface OH
0
(100l
face 0 atom
0 Mg atom
132
from chemisorption of CO. What was not generally understood was
that CO chemisorption, even at room temperature, can readily
lead to the fragmentation and oxidation of very small supported
rhodium aggregates. The noninnocence of CO as an adsorbate was
demonstrated by Primet , 2 9 who monitored the reactions of CO
with small alumina-supported rhodium aggregates using infrared
spectroscopy. At temperatures above 100 K the Rh(I) geminal
dicarbony1 species are formed at the expense of CO-covered Rh
aggregates. These mononuclear species result from oxidative
addition of surface hydroxyl groups to zerovalent metal
centers.
133
catalyst surface, and experiments with Os aggregates formed
from chloroosmic acid on alumina34 (which is not a strong
enough base to stabilize the "molecular" clusters) were roughly
two orders of magnitude more active than the MgO-supported
osmium carbonyls (Table 1) . Neither this conventional catalyst
nor the supported "molecular" catalyst had an unusual
selectivity for the CO hydrogenation reaction (Table 1) .
134
Table 1. Performance of supported osmium catalysts for CO hydrogenation at low conversions.
w
(11
w
en
Product Approximate
Catalyst Catalyst distribution, mol% catalyst
precursor support T, °C P,atm C1 C2 C3 C4 C5 lifetime, h Ref.
137
Table 3. MgO-supported rhenium catalysts: effects of surface
site nuclearity on catalytic activity measured as
turnover frequency, molecules/(Re atom•s) .28,29
CATALYST
Propylene
hydrogenation 3.5 x 1o-2 2. 1 x 1o-2 -1o-1
at 80°C
Cyclopropane
hydrogenolysis
to give methane 4. 7 X lQ-4 -3 X lQ-2
+ ethane at 200°C
138
We should imagine families of well-defined supported metal
catalysts that can be characterized in a depth rivaling that
attainable for molecular catalysts. From them may emerge
advances in fundamental understanding of metal catalysis to
rival those made possible some 15 years ago by the techniques
of ultrahigh-vacuum surface science applied to single crystals
of metal.
ACKNOWLEDGMENTS
REFERENCES
139
21. P. Fejes, I. Kiricsi, I. Hannus, T. Tihanyi, and A. Kiss,
in "Catalysis by Zeolites," B. Imelik et al., eds.,
Elsevier, Amsterdam (1980) p. 135.
22. L. R. M. Martens, P. J. Grobet, and P. A. Jacobs, Nature
(London), ~ (1985) 568.
23. P. L. Zhou and B. C. Gates, J. Chern. Soc, Chern. Commun,,
(1989) in press.
24. H. H. Lamb, M. Wolfer, and B. C. Gates, to be published.
25. J. Schwank, L. F. Allard, M. Deeba, and B. C. Gates, ~.
~., M (1983) 27.
26. F. B. M. Duivenvoorden, D. C. Koningsberger, Y. S. Uh, and
B. C. Gates, J. Am· Chern. Soc., ~ (1986) 6254.
27. P. S. Kirlin, F. A. DeThomas, J. W. Bailey, H. S. Gold, C.
Dybowski, and B. C. Gates, J. Phys Chern , ..9..Q. (1986)
4882.
28. P. S. Kirlin, F. B. M. van Zon, D. C. Koningsberger, and
B. C. Gates, to be published.
29. M. Primet, J Chern Soc. Faraday Trans.I, 1.1 (1978) 2570.
30. H. F. J. van't Blik, J. B. A. D. van Zon, T. Huizinga, J.
C. Vis, D. C. Koningsberger, and R. Prins, J. Am. Chern .
.fulQ., 1.Q1. (1985) 3139.
31. D. J. C. Yates, L. L. Murrell, and E. B. Prestridge in J.
Bourdon, ed., "Growth and Properties of Metal
Clusters," Elsevier, Amsterdam (1980) p. 137.
32. J. L. Robbins, J. Phys. Chern., ..9..Q. (1986) 3381.
33. H. H. Lamb, T. R. Krause, and B. C. Gates, Proc 9th Int.
Congr. Catal .• Calary. 1988, The Chemical Institute of
Canada, Ottawa, 1988, Vol. 3, p. 1378.
34. E. 0. Odebunmi, B. A. Matrana, A. K. Datye, L. F. Allard,
Jr., J. Schwank, W. H. Manogue, A. Hayman, J. H.
Onuferko, H. Knozinger, and B. C. Gates, J. Catal,, ~
(1985) 370.
35. R. Pierantozzi, E. G. Valagene, A. F. Nordquist, and P. N.
Dyer, J:, MQl, Ca:tal. , .2..l (1983) 189.
36. H. H. Lamb, M. Wolfer, and B. c. Gates, to be published.
37. Y. Iwasawa, AQy, !:;;;a:t;;a1., ll. (1987) 187.
38. Y. Iwasawa, N. Ito, H. Ishii, and H. Kuroda, J:, Ch~m. SQ~.
Cbew CQIDID.l.lll. (1985) 827.
39. P. s. Kirlin and B. c. Gates, Nat.l.li:e ( LQilQQil) , .32..5. (1987)
38.
40. Y. Iwasawa and B. C. Gates, CHEMTECH, in press, March,
1989.
140
MECHANISTIC FEATURES OF CARBONYL CLUSTER REARRANGEMENT
ABSTRACT
INTRODUCTION
MOLECULAR POLYHEDRA
142
b c
0~
d
a
b
c
d
e
f
The convex de~ahedra:
Tetrahedron (4-hedron)
Triangular dipyramid (6-hedron)
Octahedron (8-hedron)
Pentagonal dipyramid (1()-hedron)
12-lledron
14-lledron
G/' 0;_\
f
g 16-hedron
h Icosahedron (20-hedron) g h
143
polyhedra produce unequal bond lengths. For this reason alternative
polyhedra are !Sf, {71, lBf, {91 and t10! in which all M-L bond lengths
are equivalent have been sought. We shall not consider this aspect in
detail but it is worth noting that a complete range of polyhedral forms
in which the vertices lie on the surface of sphere have been derived.
Thus, if we consider eight coordination the square antiprism {10-2) is an
example of a polyhedron in which all M-L bonds are equal but which is not
the most stable form. In coordination geometry we are frequently
presented with the question of whether or not the coordination geometry
approximates to one in which all L-L contacts are equal but the M-L bonds
are not, or an alternative in which all M-L contacts are equal but
longer. There is apparently no hard and fast answer to this question and
examples of both occur.
We shall now consider the effect of breaking two or more edges. Edge
cleavage with retention of the maximum number of edges, will lead to the
results shown in Figure 6 for the octahedron, and in Figure 7 for the
triangulated dodecahedron. For {6f, the successive cleavage of
appropriate edges leads first to (7-1f, then to rs-2), and finally to
{9-3] - the trigonal prism:
In each case the systematic and steady progression across the series
leads to increasingly open polyhedra (i.e. with some square rather than
all triangular faces) and a controlled increase in the M-L bond length.
A decrease in stability will also follow the same progression. This will
be important in our consideration of pseudorotation of coordination
compounds.
PSEUDOROTATION
144
Figure 4: Polyhedral Growth Sequence 2
-
{8-2} {9-3}
{6} { 7-1} (3)
(1)
(2)
145
® {8}
___..
r:i1lJ ~
{9-1 }
{1 0-2 }
\
{ 13-5}
..- {12-4}
- {11-3}
@
'---...
{ 14-6 }
146
transition state or the intermediate. The ease with which these
processes can occur will depend upon the energy difference between the
complementary geometries (the ground state or the transition state) which
constitutes the barrier to the process. These in turn depend on the
relative movements of atoms or groups of atoms within the parent. When
the complementary geometries are energetically close, either because of
the nature of the ligands or because of the inherent properties of
certain coordination numbers, the barrier to pseudorotation will be
small. In many cases the barrier will be increased by the need of the
system to traverse an intermediate in which the ligand distribution or
unfavourable bond angles and/or unfavourable bond distances lead to
instability. The best examples of pseudorotation and stereochemical
non-rigidity will be found for those complexes in which the barrier
between complementary geometries is significantly small than the free
energy of activation for bond fission. The problem usually is on
deciding upon the appropriate, accessible, complementary geometry.
147
In the past it has proved to be relatively more difficult to derive
satisfactory complementary geometries for coordination numbers other than
4, 5 and 6. For larger coordination numbers, e.g. seven or eight, the
availability of several closely related polyhedral forms has usually been
taken as sufficient to explain the generally observed low barriers to
fluxionality and no general scheme appears to exist. In this paper we
wish to present such a general scheme which we believe offers a
convenient approach to these fascinating phenomena and permits a simple
view of suitable complementary geometries.
[6f ...., (7-1) _. [B-2f - {9-31 -+ [B-21 ' ...._ l7-1J' - [6)'
148
geometry available. Single edge cleavage, as illustrated in Figure 4
will lead to a nido-dodecahedron (8-1} which cannot serve as a
complementary geometry since operation of the diamond-square-diamond
process in this instance will not return us to the parent polyhedral form
(7J . Instead the capped-octahedron will be produced. This may be
regarded as unfavourable for a homoleptic ML7 complex since one M-L bond
(viz. that involving a bond to the capping ligand) is required to undergo
a relatively large extension. A further edge cleavage is required to
give the more favourable arachno-tricapped prism, {9-2}, which can serve
as an appropriate complementary geometry. Hence, ligand rearrangement in
this case may be viewed as the five-step process:
149
Figure 9: Bailar Twist Mechanism for 6-coordinate Complexes
COMPLIMENTARY GEOMETRIES
{5} {6-1>
<8> <9-D
{9) <10-1}
150
important factors such as the relative motion of the metal cluster unit
within the ligand polyhedron must also be considered.
So far we have been concerned only with structure in the solid state.
In solution structures are sometimes different and alternative isomeric
forms may exist. A particularly good example is Co2(CO)g. In the solid,
this molecule has a well established structure with two bridging
carbonyls. In solution, however, it has been shown, by IR spectroscopy,
that several other isomeric forms, some without bridges also exist.
Similar observations have been recorded for Fe3(C0) 1 2 which also exhibits
151
Figure 11: Molecular Structures of (a) Co2(CO)a
(b) Fe 2 (CO)g and (c) Mn2(C0) 1 o
152
(a) (b)
Figure 12: The Molecular Structures of (a) Fe3(COl12 and (b) Os3(COl12
Cube- Octahedron.
3: 6: (3)
Figure 13: The Molecular Structure of Ir4(C0) 1 2
153
an IR spectrum that is markedly solvent and temperature dependent
suggesting the existence of alternative· isomeric forms. These and other
similar observations have led to the view that for many carbonyl clusters
interconvert through relatively minor deformations and it is clear that
for a given carbonyl a number of structures exist with energies similar
to that of the ground-state (solid-state) and which can be excited by
thermal energy. As a consequence a classical (localised) bond
description may not be appropriate since the ground-state structures does
not sit in well defined potential in the potential energy surface
governing nuclear motion. The existence of these readily excited states
leads to the dynamic phenomenon of fluxionality - the subject of this
paper.
In this part of the paper we shall deal with the nature of these
isomeric forms, and with fluxionality in both the solid and solution.
Three carbonyl prototypes will be considered, viz. Co2(C0)9, Fe3(C0) 1 2
and Co4(C0)12• but similar arguments may be extended to all other
systems.
Co2(C0) 8
Fluxional Behaviour
154
- -
Figure 14: Molecular Structure of Fe3(COl12
According to the Polyhedral Model
155
The nature of these fluctuating structures in the solid is also
easily visualised. Consider form 2a. Rotation of the Co-Co vector about
its two-fold axis - and this will coincide with the Co-Co vibration -
will initially lead to the formation of two asymmetric bridges and
eventually to the non-bridged form 2b. Hence, to the apparent
equilibration of bridging and terminal groups. In reality, there will be
an oscillation or libration of the CO-Co unit about this two-fold axis
and at no point do ligands actually migrate from one metal atom to the
other at this stage of the fluxional process. However, this libration
will also be expected to perturb the CO polyhedron about the same
molecular c 2 axis and it would not seem unreasonable to associate this
with the distortion of the bicapped trigonal prism {9-1J towards the
square antiprism {10-2) . Eventually, as the temperature is increased,
and the amplitude of the libration becomes sufficiently large a new
isomer with the square antiprismatic geometry will be reached. At this
stage intermolecular CO ligand migration could then occur via the
sequence of pseudorotation processes for eight-coordination as outlined
above, i.e.
Fe 3 (C0) 12
156
Ia
a
b
Square antiprism 3
a
b
[2]
[3]
157
icosahedron of CO groups. Rotation of this triangle by 15° about the
molecular C2 axis which passes through the unique iron atom Fe(l)
produces a second isomeric form of Fe3(COl12 with D3-symmetry (Figure
14) . This is also based on an icosahedron of CO groups but in contrast
to the solid state structure all CO groups are terminally bonded and all
Fe-Fe distances are the same (260 pm) . This rotation of the iron
triangle will also cause the slightly flattened icosahedron of carbonyls
to become regular. Other isomers are possible. The most obvious being
based on the less stable anticuboctahedron of carbonyls with a
pseudo-Os3(COl12 structure (see e.g. ref. 9).
Fluxional Behayiour
Co 4 (C0) 12
158
centre of the opposite triangular face. The molecule itself has only C1
symmetry with three carbonyl groups in edge-bridging positions around a
tetrahedral face, the remaining carbonyls being terminally bound. As
with Fe 3 (C0) 12 the CO-polyhedron is again approximately icosahedral, and
the molecular symmetry is only slightly distorted from C3v (Figure 17).
The IR spectrum at ambient temperature of both solid and solution samples
is consistent with this symmetry and the same structure is thought to
persist in both phases. Nevertheless, we find that the spectrum does
change with temperature and evidence for a new isomeric form has been
(11). Within the icosahedron of carbonyl ligands at least two other
orientations of the tetrahedron are possible (Figure 17). (12). Rotation
of the Co 4 tetrahedron about the unique C3 molecular-axis which passes
through the apical Co atom and the centre of the basal co 3 face produces
a new form (2) which is also based on the icosahedral of carbonyl groups
but in contrast to isomer (1) all CO groups are equivalent and terminal.
This isomer has Td symmetry. alternatively, rotation of the co 4
tetrahedron about any one of the three C3 molecular axes which pass
through a basal Co atom and the opposite Co3 face produces a third
isomeric form which also contains CO-bridges (3) (12). The high
temperature IR indicates a change in those frequencies associated with
the CO-bridging region and is consistent with D2 d symmetry. Hence,
isomer (3) is the preferred.
Fluxional Behaviour
CONCLUSIONS
159
possible L-L contacts. We have also shown that the observed solid state
n.m.r. data for cluster carbonyls can be rationalised in terms of low
energy librations of the metal polyhedron within the CO-polyhedron,
rather than the higher energy CO intermolecular migrations previously
postulated. The m.s.d. parameters derived from X-ray data provide
evidence for such librations; the thermal ellipsoids of the iron atoms in
Fe 3 (co) 12 indicate motion of the iron triangle about its pseudo
three-fold axis, and those of the cobalt atoms in Co4(COJ 1 2 indicate
motion about the three c 3 axes of the cobalt tetrahedron. Other
concerted motions of the metal atoms and the CO groups must also be
involved in these dynamic processes in order to account for CO migrations
in solution; these concerted motions must be consistent with the symmetry
of the normal modes of vibration of the system and will involve amongst
others pseudorotation of the CO-polyhedron in ways related to those
postulated for simple coordination compounds. The formation of cluster
isomers also follows naturally from these arguments. Similar arguments
may be applied generally to cluster carbonyls but great care must be
exercised in drawing comparisons between the solid and solution states
especially when ionic compounds are under consideration.
REFERENCES
160
SELECTIVE OXIDATION CHEMISTRY ON SOLUBLE OXIDES: A PROGRESS REPORT
acrystalytics Company
Lincoln, Nebraska 68501
Department of Chemistry
University of Nebraska
Lincoln, Nebraska 61801
bDepartment of Chemistry
University of Illinois
Urbana, Illinois 61801
INTRODUCTION
Global
Reactant(s) Product production,
k I r
C2H4 Polyethylene 6 X 10 9
CH 30H, 02 Formaldehyde 5 X 10 9
c3H6 , NH 3 , 0 2 Acrylonitrile 3 X 10 9
E._- Xylene, 02 Phthalic Anhydride 2 X 10 9
_g_-Butane, 02 Maleic Anhydride 4 X 10 8
Finally, the nature of C-H activation over vanadium oxides will be consid-
ered, not in terms of achieving partial oxidation but in terms of synthe-
sizing soluble polyoxovanadates that might have features normally associ-
ated only with solid vanadium oxide surfaces.
METHANOL DEHYDROGENATION
(1)
162
across a molybdenum-oxygen double bond followed by oxidation involving par-
ticipation of a second molybdenum center. Although this mechanism has re-
ceived no experimental support, its key features have been endorsed by
Allison and Goddard 4 • 5 on the basis of general valence bond quantum mechan-
ical computations. These calculations support addition of methanol across
an Mo=O bond followed by an oxidation step that involves an adjacent molyb-
denum center. Unfortunately, there is no experimental precedent for the
formation of stable surface methoxides of the type observed in methanol ox-
idation by addition across Mo=O bonds. Instead, oxomolybdenum(VI) alkoxides
favor dioxomolybdenum structures that could be obtained by addition across
molybdenum-oxygen single bonds. 6 Formation of this type of intermediate ob-
viates the need for a dual-site oxidation process and allows the possibility
of a one-site oxidation mechanism. A mechanism employing these two altern-
ative steps is shown in Scheme III.
Scheme I
0 0 HO OCH3 0
I
Mo--0--Mo
I + \1 Mo--0--Mo
II
HO OCH 3 0 HO OCH OH
\1 II
Mo--0--Mo
\1 2
Mo--0--Mo--
I
Scheme II
163
OCH 3 HO
. . . _ _...o........._ / "'-../ '\./
Mo
Mo Mo + ~ Mo +
#~ #~ #~ #~
0 0 0 0 0 0 0 0
OCH 3 0~
"'-../ "'-..--~cH 2
Mo ~ -•------'~~ Mo
0
#~
0 0
#'\.OH
Scheme III
164
b
3- 7 2
electron-donating than the P3o9 ligand. Formation of a coordinated n-
aldehyde intermediate is consistent with the reluctance of the sterically
crowded pentamethylcyclopentadienyl complex (C 5Me 5 )Mo0 2 (0CH 2CH 3 ), c, to
generate acetaldehyde, even at 125°C in solution.
ETHYLENE EPOXIDATION
(2)
The industrial process for producing ethylene oxide in this fashion util-
izes silver supported on alumina as a catalyst and proceeds as indicated
165
Ag(ox.)
Ag(ox.)I0 2(ads.)
Ag(ox.)IO(ads.)
)
Scheme IV
2-
(3)
166
2-
10
form, and identified as an isomer of e, the oxametallacyclobutane f:
2-
XYLENE/BUTANE OXIDATION
167
~ + 3 02
~ 0
+ 3 H2o (4)
+ 7/2 02
~ 0
+ 4 H2o (5)
168
ACKNOWLEDGMENT
REFERENCES
169
14. M. A. Pepera, J. L. Callahan, M. J. Desmond, E. C. Milberger, P. R.
Blum, and N. J. Bremer, J. Am. Chern. Soc. 107, 4883 (1985).
15. For representative examples, see (a) S-(V0) 2P2o7 : Yu. E. Gorbunova and
s. A. Linde, Sov. Phys. Dokl. (Engl. Transl.) 24, 139 (1979); (b) a-
VP05: B. Jordan and C. Calvo, Can. J. Chern. 51, 2621 (1973); (c) S-
VP05: R. Gopal and C. Calvo, J. Solid State Chern. 5, 432 (1972); (d)
v2o5 : R. Enjalbert and J. Galy, Acta Crystallogr. Sec. C C42, 1467
(1986); (e) K12v18o42 ·16H 20: G. K. Johnson and E. 0. Schlemper, J. Am.
Chern. Soc. 100, 3645 (1978); (f) [NMe 4 J6 rv 15 o36 ]•Cl·4H2o: A. Muller,
E. Krickemeyer, M. Penk, H.-J. Walberg, and H. Bogge, Angew Chern., Int.
Ed. Engl. 26, 1045 (1987); (g) K6 [As 6v15o42 (H 20)]·8H 20: A. Muller and
J. Doring, Angew. Chern., Int. Ed. Engl. 27, 1721 (1988).
16. V. W. Day, W. G. Klemperer, and 0. M. Yaghi, submitted for publication.
170
THE STUDY OF CLUSTERS OF POLYLITHIUM ORGANIC COMPOUNDS AND STRUCTURAL
STUDIES OF POLYLITHIUM ORGANIC COMPOUNDS
Richard J. Lagow
Department of Chemistry
The University of Texas at Austin
Austin, Texas 78712
Tungsten Wire
Vacuum Feedthrouah
Prabe Body
Copper Leads
172
Table 1. Major Ions from the Flash Vaporization Mass Spectrum of CH 3Li
Table 2. Major Ions from the Flash Vaporization Mass Spectrum of CH 2Li 2
Table 3. Major Ions from the Flash Vaporization Mass Spectrum of HCLi 3
173
Table 4. Major Ions from the Flash Vaporization Mass Spectrum of CLi 4
CLi 2 26
CLi 3 33
c2Li 31
c2Li 2 38
c2Li 3 45
c2Li 4 52
c3Li 2 50
c3Li 3 57
c3Li 4 64
c3Li 5 71
(C3Li4)2 128
(C3Li4)3 192
spectrum is that the sample was not highly pure and contained other
species such as dilithiomethane). Note the dilithiomethane gives species
as high as n = 6. The Robert West compound, c3Li 4 , gives interesting
species such as the trimer (C 3 Li 4 ) 3 (Table 5).
It is also noteworthy along these lines that due particularly to the
thermal rearrangement of solid methyllithium to dilithiomethane, 5 the
mass spectrum of methyllithium was unattainable by standard mass
spectrometric techniques. Using our technique it was possible 7 to
compare the spectra, and establish for the first time what was always
suspected--that the vapor of methyllithium should be tetrametric. One
can now compare the spectra of methyllithium with the classic mass
spectral work of Joe Berkowitz and Ted Brown 8 • 9 (see Tables 6 and 7) on
ethyllithium and t-butyllithium. Of course, here the secret is that methyl-
lithium is vaporized before it rearranges to dilithiomethane.
174
Table 6. Positive Ions in the Flash-Heated High-Resolution Mass Spectrum
of Halide-Free Methyllithium
Rel
Calcdb Measd Error, Rel
a Mass ppm Intensityc
Ion Mass
CH b
3
175
As previously stated, polylithium compounds vaporize in clusters
with the dimer or trimer being more intense in the mass spectrum than the
monomer. In almost all cases one observes a "parent ion" for the monomer
as well as higher polymers. It is thought that the exact distribution
and nature of the clusters observed depend on the mass spectrometric
technique which is used to produce the clusters from the three
dimensional solid lattice. In this respect, the flash vaporization mass
spectrometric technique which was developed to see the polylithium
compounds is viewed in our laboratory as an extreme method and was
primarily invented to overcome lack of appropriate equipment at the
University of Texas. A new high resolution spectrometer with
multifaceted techniques has been obtained from the United States National
Science Foundation and it is felt that when such techniques as field
desorption, fast atom bombardment (FAB) and perhaps laser desorption are
applied in this field that even higher molecular weight species will be
generated. In subsequent sections of this manuscript we will discuss
structures and predictions of structures of polylithium organic
compounds. While the structures of monomers appear to be of great
interest and importance, the structures of clusters and oligomers should
be both extraordinarily interesting and unusual. Collaboration is
underway with Professor Josef Michl and his research associates in which
such extraordinarily interesting species are to be vaporized into Nobel
gas matrices and their structures studied by vibrational electronic and a
wide range of other spectral techniques. An apparatus has been designed
where a steady stream of powder drops on a high temperature surface. The
vapor produced is deflected into a matrix as it fragments into monomers,
dimers:
lT
Polyllthium Power Feed
High
Vacuum I
Environment Low Temperature
I Matrix for
~~~rature _)J #
Various
Spectroscopic
Investigations
176
feasibility of such techniques. A schematic for an apparatus designed to
flash vaporize polylithium clusters onto a cryostat is pictured below:
177
Q
cu-Orlrlhromethane
p-OrbrtJI syslem m
d•lrthtOmethane
traiU'·DIIIIhlomethlne
z
enforced reduction of the angle around the planar carbon atom by means of
small rings. 10 • 18 Because it has pi acceptor and sigma donor character
together with its small atomic radius, lithium may be particularly
effective in stabilizing the planar arrangement. Roald Hoffmann had
previously observed that lithium substitution could be very advantageous
for stabilizing planar carbon in a treatise on planar carbon. 18
Ab initio molecular orbital calculations on dilithiomethane (CH 2Li 2 )
-10 19-20
monomer ' have revealed remarkable features: cis-trans isomers 1
and with the cis isomer being the more stable species, and a low
~.
178
H
H-'
lr
Ll
\/'
lr H lr lr
H--e
I~ e--H e \I li
l~\
e
H lr
lr
1\
F H1
H
I, 2- Orlrthraethane Tetr•lrthromethane Orfluarcrdrlrthrornethane I, 1-0rlrthrocyclopropane
5 I 7
•
Unsaturated organolithium compounds may show even greater anomalies
in structure. Hydrocarbon olefins have been generally found to be two-
dimensional, that is, planar. However, the most stable conformation for
the lithium-substituted analog of ethylene is predicted to be a skewed
olefin structure 2 with a 5 kcal/mol rotation barrier holding the vapor
monomer in a skewed configuration with the lithium site orthogonal to the
unsubstituted olefin sites. 17
This calculation has been conducted at higher levels by Fritz
Schaefer 23 with the same conclusion. The skewed olefin effect, one which
one can see we now have the molecules to probe, is one that we in the
Lagow research group really expect to see. The vacant p orbitals on
lithium have a very high probability of interacting with the electron
rich ~-systems and the distances are reasonable.
The most stable configuration of the tetralithioethylene species 10
has been predicted to be a planar structure with "electron-deficient"
three-center bonding leading to a bridging lithium configuration. 12 This
unsual bridging structure has also been postulated for the acetylene
derivative c2Li 2 , ll• rendering it planar rather than linear. 12
H
..."'··... Lr
··.. I/
li
' \ I
I
I \
\
\
H/c
/ \ e c
Lr---e e---li
\ / \ /
\ I \ I
'u' \
li
I
' 10 II
179
Schleyer proposal was that the molecule dilithium (Li 2 ) could act as a
ligand and supply one or two sites on the CLi 4 monomer structure with an
electron:
1---------1-----
Li CLis
----l----z-- Li
Li Li
Li
2 D.,.,h -s180
3 0 3h - 28 120
4 Td - 38 109.47
5 2
03h -l/ 2 ( 8 180 + 28 120) [ 1 + {1 - 8 ( 8 180 8 120 - 38 90)/(S180
+ 2S )2}1/2]
120
6 oh -(S180 + 48 90)
8 oh -( 38 70.53 + 38 109.47 + 8 180)
180
Table 9. SCF Total Energies and Carbon Populations for Some Lithiated Carbon Species CLi , as Calculated with
3-21G and Dunning-Hay (DH) Split Valence Basis Sets,f n
a b c
Present work. 3-21G optimized geometry (2.022 a , 1.999 c ), ST0-3G basis set and geometry (1.979 a , 1.953 c ).
d e f
3-21G optimized geometry (2,094). ST0-3G basis set and geometry (2.063). Note that the charge on carbon is
assessed as being much more negative in the natural population analysis than in the Mulliken population analysis.
(XI
Figure 2. Comparison of 2p NAOs in neutral carbon atom (top) and the
carbon atom of CLi 4 (bottom), showing the much greater
diffuseness of the latter compared to "normal" carbon
orbitals. (The orbitals shown are the "pre-NAOs", preceding
interatomic orthogonalization, whose overlaps allow direct
estimates of NAO Fock matrix elements and are more readily
visualized.) Contour intervals are 0.02 au with outermost
contour at 0.02 au.
Calculated Measured
Ion Mass Mass Source of Ion
182
tetralithiomethane 24 , later in our laboratory we observed a number of
other five coordinate carbocations 27 (see Table 10). The structures of
the first five coordinate carbocations have also been calculated 24 • 25 .
Presently our mass spectrometric facilities do not permit us to determine
whether these species come from the more likely ion molecule reactions in
the vapor phase or the more interesting generation from neutral
precursors such as (CLi 5 )n' (CLi 6 )n' or (CLi 8 )n.
In the recently completed structure of dilithiomethane, it is
possible to see how one might generate by certain decomposition and
vaporization processes gas phase species such as CLi 6 and CLi 8 . Our
laboratory in conjunction with those of Galen Stucky of the University of
California at Santa Barbara and David Cox of Brookhaven National
Laboratory have recently completed the neutron powder diffraction
structure of dilithiomethane and the same collaborators have the
synchrotron microcrystal structure in progress on the same compound along
with the synchrotron microcrystal structures of other polylithium
compounds soon to be studied. A first approximation of the structure of
dilithiomethane, (CLi 2H2 )n' has the antifluoride structure shown below:
0 0
c u
One can see in a perfect antifluoride structure that each carbon atom is
surrounded by four lithium atoms and then the "hypervalent" clusters are
bridged by two lithiums to the next carbon. Thus a network polymer is
observed for dilithiomethane.
Neutron diffraction data were measured at 5, 101 and 293 K on a
triple-axis diffractometer at Brookhaven National Laboratory. All lines
were indexed in a body-centered tetragonal system with lattice constants
at the above temperatures: a=b=5.6216(6), c=10.865(2) @5K a=b=g=90°;
a=b=5.631(1), c=10.901(1) @101K; a=b=5.833(1), c=10.938(1) @ 293K. The
structure was solved by a combination of (1) using Patterson techniques
and estimated peak intensities, (2) packing models based on the relative
intensities of the Bragg reflections and (3) comparisons with known
183
.....
4r\'r-er...
0
o
00
~
0
eLi
c
0
4
,...
iooo..
0
G)
to...13
binary structures of the form AB 2 • The fact that the c unit cell, axis is
approximately 2 x a and the above considerations led to the refined
structure shown in Figure 3 in the space group I4 1;a ·
The structure is a very distorted version of the antifluorite
structure of Li 2o28 and Be 2c29 in which one of the cubic cell edges is
doubled (Figur~ 3) . To a first approximation the carbon atoms are
located at the face centered sites of the parent cubic antifluorite cell
and the lithium atoms are near the (0.25,0 . 25,0 . 25) body diagonal
position. The shortest lithium- lithium atom distance in the refined
tetragonal structure is 2.26(2) K which
can be compared with the Li-Li
distance of 2. 314(1) ~in Li 20, 2. 383(6) Kin [LiC 6H11 ] 2 , 30 and 2.56(1) A
in CH 3Li 31 Both the lithium and carbon atoms were found to be in
ordered sites.
~axis
Qc
u
184
Table 11. Structural Parameters CD 2Li 2 .
Atom X y z U(isotropic)x10
The arrangement of the lithium atoms about the carbon atom is shown
in Figure 4. A crystallographic c2 axis passes through C(l) in the
vertical direction of the figure and only one of the two disorder
orientations for the deuterium atoms is shown. 32 The closest carbon-
lithium approach [2.17(2) !]
is comparable to that in cyclohexyllithium30
[2.172(5) A] or in ethyllithium33 [2.19(1) A]. The shortest Li-D
distances [1.72(3) to 2.17(2) XJ 34 are compatible, but shorter than Li-H
185
distances previously observed for organolithium compounds [e.g. 1.96(1)
K for the shortest alpha carbon hydrogen atom in cyclohexyllithium30 • 35 ].
The puckering of the four atom lithium rings shown in Figure 4 optimizes
the close approach of C(1)-Li(1) [2.17(2) ~] and Li(1)-D(1) [1.72(2) A].
The observed D(1)-C(1)-D(2) angle is 100(1) 0 with C-D distances of
1.09(1) A[C(1)-D(1)] and 1.18(2) A[C(1)-D(2)].
The local geometry about the carbon atom (Figure 4) suggests two
types of C-D coordination if one considers closest neighbor interactions.
The coordination of D(1) can be viewed as (I) below, while that for D(2)
;H· ..
~C-H···M -c-·····M
I \
II
186
interest to study the origin of such mass spectrometrically observed
species to determine whether their precursors were these neutral species.
Another unusual and interesting feature of the structure is the
extraordinary coordination number of the central carbon atom. The
current dilithiomethane structure contains a carbon atom which is
surrounded by eight lithium atoms, six of which are less than 2.4 A away.
If one considers the additional bonds to the 2 deuterium atoms, all
carbon atoms in the structure could be considered to be at least eight
coordinate and possibly even ten coordinate. Although coordination
numbers for carbon of up to six are known in a few transition metal
compounds, a coordination number of eight is the highest coordination
number of carbon yet reported as far as we can determine.
Prior to the neutron diffraction structure obtained by the Stucky,
Lagow and Cox research groups, the solid 13 c spectra of dilithiomethane
was studied. The 13 c spectra of solid CH 2 ( 7Li) 2 , CH 2 ( 6 Li) 2 and CH 36 Li
·
· d us~ng
were o b ta~ne 1H- 13 c cross po 1 ar~zat~on
· · mag~c · ·
· ang 1e sp~nn~ng
(CPMAS) techniques in collaboration with C.S. Yannoni. 39 The weak signal
13 7
observed in the proton-decoupled C spectrum of CH 2 ( Li) 2 at 150 •c
shown in Figure Sa is due to broadening from dipolar coupling with
the 7Li nucleus. This type of broadening was eliminated through use of
.
d ecoup 1 ~ng tee hn~ques,
. 40 t h roug h mag~c . . 40 an d t hroug h
. ang 1e sp~nn~ng
(a)
(b)
187
(b)
(c)
13
Figure 6. Proton-decoupled solid-phase C NMR spectra of
methyllithium obtained at -150 •c: (a) CH 3 7Li;
(b) CH 36 Li, no 6 Li decoupling; (c) CH 36Li, 6 Li
decoupled. Peak is 16 ppm upfield from Me 4Si.
188
The downfield shift of CH 2Li 2 relative to methyllithium may be
attributed to an increase in charge density at the methylene carbon with
increasing lithium substitution. Electropositive substituents are known
to function as effective o-donors. 18
The very small linewidth for CH 2 ( 6 Li) 2 compared with that observed
for CH 2 ( 7Li) 2 shows that the stratagem of using CH 2 ( 6 Li) 2 and decoupling
6 Li works, and should prove useful for 13 c studied of other organolithium
materials.
Further information about the structure of dilithiomethane and a
great deal of information about the charge distribution in the compound
were obtained in a collaborative study on ESCA spectrum with Mike Hall
and co-workers at Texas A&M University. 42
(a) Methylilth•um
700
4000
Li Is
400
2
ot
<:"'::J !!
a ()
u 0
<:
::J
0 u;
~ (b) Diithiomethone
2500 1000
1500
600
189
0.0
-:c
-1.0
en
"6
.Eu -2.0
"'
.t::
(.)
..!!!
(.) -3.0
-4.0
Calculated qc
C 1s C 1s Calcd Li 1s Calcd
Compound BE, eV shift, eV qc BE, eV qLi
aUsed as reference. b
Average of monomer, dimer, and trimer structures.
Standard deviation in parentheses.
190
In the synthesis area, a very important new development is a high
yield general synthesis for gem 1,1-dilithio compounds. Gem dilithio
compounds are very unusual new reagents and of interest from a structural
and theoretical point of view. 10 • 11
In 1985, our research group 3j modified and refined a Karl Ziegler
pyrolysis and thermal rearrangement of methyllithium generating a high
yield synthesis for dilithiomethanes. This synthesis made
dilithiomethane an accessible reagent and added polylithium chemical to
the repertoire of conventional organic synthetic chemists.
Recently we have established that the Kawa-Lagow-Ziegler synthesis
is a very general and broadly applicable synthetic method. 3j .The method
can be modified so that synthesis of higher 1,1-dilithio compounds such
as 1,1-dilithioneopentane, (trimethysi1yl)dilithiomethane and bis-
(trimethylsilyl)dilithiomethane are possible in 88, 97 and 95% yield; 3j
the synthesis of tetramethyl-1,1-dilithiocyclopropane in 90% yield; 3k
and the synthesis of 9,9-dilithiofluorene. 43 (Figure 9.) Many of
the compounds which have been prepared in the R.J. Lagow research
laboratory by the new Kawa-Lagow-Ziegler synthesis combined with those
prepared by the original Lagow lithium vapor-chlorocarbon replacement
Me3So H Mo3Si Lo
'-./ 170"C ..
""-/
c + IMo3 Sil:! CHz
2
/c'-.. 95%
Me3Si/ "LI
Mo3So Lo
'/
ICH313 c H (CH3 13C u
2 c '
H/ '-u
/ 180"C
88% > c
H/ ' L i
• (CH 314C
I'/
ICH3lzC H
2 I"/ c 180"C
90% .,. /c""- • I '-cHz
(CH.Y~/ '-.Li (CHyzC Ll ICH3lzC/
.~"
if"
Figure 9. Synthesis of gem-dilithium compounds
by the Kawa-Lagow-Ziegler Pyrolysis.
191
reaction has led to a number of interesting compounds which should play a
key role in establishing the accuracy of the Paul Schleyer, John Pople
and Fritz Schaefer forecasts:
Me
Mel> Li
Li
Me
Me
These are all prime candidates for study in the Nobel gas matrix
spectroscopy program joint between the Richard Lagow and Josef Michl
research groups at the University of Texas. These and many more
dilithiomethanes and 1,1-dilithio olefins may reveal the extraordinary
features predicted over the last decade about the structures of monomers.
The preliminary NMR data, some field desorption mass spectrometry data
from our laboratory, and flash vaporization mass spectroscopy done in our
laboratory coupled with an earlier prediction of quadruply bridged
11
lithium by Schleyer and Pople are beginning to create for such dilithio
species a structural picture where there are two orthogonally sets of
bridging lithium. We are finding that very bulky groups prevent the
extensive oligomerization characteristic of the polylithium compounds and
will probably lead to structures (we have both single crystal and
microcrystal synchrotron studies in progress) where the lithium compounds
exist in the crystal lattice as dimers:
R~.--~:7:--- R
R
::.:.....u~
R
Li
Acknowledgement
We are grateful to the National Science Foundation (CHE-8521390) and
Robert A. Welch Foundation (F-700) for support of this work.
192
Dedication
This manuscript is dedicated to my close friend and former
colleague, F. Albert Cotton. I have had the highest respect for Al
Cotton as a person and as a scientist for many years. Al has contributed
to my career in many ways and is a person from whom I have learned much.
He encouraged me early in my career, when there were long odds against
the RJL idea that a new form of matter such as polylithium organic
compounds could exist when there were suggestions and reports in the
literature that such compounds were unstable and impossible to prepare.
Al Cotton is a mentor who supported and encouraged the early studies in
this area. F. Albert Cotton on an earlier occasion was also responsible
for selecting and purchasing the first bottle of outstanding wine ever
consumed by RJL, starting an exciting new hobby.
References
1. J.A. Gurak, J.W. Chinn, Jr. and R.J. Lagow, ~Am. Chern. Soc.
104:2637 (1982).
2. L.A. Shimp, J.A. Morrison, J.A. Gurak, J.W. Chinn, Jr., and R.J.
Lagow, ~Am. Chern. Soc. 103:5951 (1981).
3. (a) C. Chung and R.J. Lagow, ~Chern. Soc., Chern. Commun. 1970
(1972); (b) L.A. Shimp and R.J. Lagow, ~Am. Chern. Soc. 95:1343
(1973); (c) L.G. Sneddon and R.J. Lagow, J. Chern. Soc., Chern.
Commun. 302 (1975); (d) J.A. Morrison, c.-chung, and R.J. Lagow, J.
Am. Chern. Soc. 97:5015 (1975); (e) J.A. Morrison and R.J. Lagow,
Inorg. Chern. 16:2972 (1977); (f) L.A. Shimp, C. Chung, and R.J.
Lagow, Inorg. Chim. Acta 29:77 (1978); (g) K.M. Abraham and R.J.
Lagow, Tetrahedron Lett. 3:211 (1979); (h) L.A. Shimp and R.J.
Lagow, ~Am. Chern. Soc. 101:2214 (1979); (i) L.A. Shimp and R.J.
Lagow, ~ Org. Chern. 44:2231 (1979); (j) H. Kawa, B.C. Manley, and
R.J. Lagow, ~Am. Chern. Soc. 107:5313 (1985); (k) H. Kawa, B.C.
Manley and R.J. Lagow, Polyhedron 19/20:2023 (1988).
4. (a) R. West and P.C. Jones, J. Am. Chern. Soc. 91:6155 (1969); (b)
G.A. Gornowicz and R. West, ~Am. Chern. Soc. 93:1720 (1971); (c) R.
West, P.A. Carney, and I.C. Mim~ J. Am. Chern. Soc. 81:3788 (1965);
(d) G.A. Gornowicz and R. West, J. Am. Chern. Soc. 93:1714 (1971);
(e) W. Priester, R. West, and T.~ Chwang, J. Am. Chern. Soc. 98:8413
(1976). - - -- --
5. K. Ziegler, K. Nagel, and M. Patheiger, ~ Anorg. Allgern. Chern.
282:345 (1955).
6. F.J. Landro, J.A. Gurak, J.W. Chinn, Jr., and R.J. Lagow,~
Organomet. Chern. 249:1 (1983).
7. J.W. Chinn, Jr. and R.J. Lagow, Organometallics 3:75 (1984).
8. (a) J. Berkowitz, D.A. Bafus, and T.L. Brown, ~ Phys. Chern. 65:1380
(1961); (b) T.L. Brown, Ann. N.Y. Acad. Sci. 136:98 (1966).
9. M.Y. Darensbourg, B.Y. Kimura~E. Hartwell, and T.L. Brown, ~Am.
Chern. Soc. 92:1236 (1970).
10. ~Collins, J.D. Dill, E.D. Jemmis, Y. Apeloig, P.v.R. Schleyer,
and J.A. Pople, ~Am. Chern. Soc. 98:5419 (1976).
11. E.D. Jemmis, P.v.R. Schleyer, and J.A. Pople, ~ Organomet. Chern.
154:327 (1978).
12. Y. Apeloig, P.v.R. Schleyer, J.S. Brinkley, J.A. Pople, and W.A.
Jorgensen, Tetrahedron Lett. 43:3923 (1976).
13. E.D. Jemmis, D. Poppinger, P.v.R. Schleyer, and J.A. Pople, J. Am.
Chern. Soc. 99:5796 (1977). --
193
14. G. Rauscher, T. Clark, D. Poppinger, and P.v.R. Schleyer, Angew.
Chern. 90:306 (1978).
15. E.D. Jemmis, J. Chandrasekhar, and P.v.R. Schleyer, ~Am. Chern.
Soc. 101:2848 (1979).
16. A.J. Kos, D. Poppinger, P.v.R. Schleyer, and W. Thiel, Tetrahedron
Lett. 21:2151 (1980).
17. Y. Apeloig, P.v.R. Schleyer, J.S. Binkley, and J.A. Pople, ~Am.
Chern. Soc. 98:4332 (1976).
18. (a) R. Hoffmann, R.G. Alder, and C.F. Wilcox, Jr., J. Am. Chern. Soc.
92:49g2 (1970); (b) R. Hoffmann, Pure~ Chern. 28:1'ii"l(1971).--
19. W.D. Laidig and H.F. Schaefer, ~Am. Chern. Soc. 100:5972 (1978).
20. E.W. Ni1ssen and A. Skancke, ~ Organomet. Chern. 116:251 (1976).
21. S.M. Bachrach and A. Streitwieser, Jr., J. Am. Chern. Soc. 106:5818
- (1984). ------
22. A.J. Kos, E.D. Jemmis, P.v.R. Schleyer, R. Gleiter, J.A. Pople, and
U. Fischback, J. Am. Chern. Soc. 103:4996 (1981).
23. W.D. Laidig an~H~ Schaefer, ~Am. Chern. Soc. 101:7184 (1979).
24. J.W. Chinn, Jr., F.J. Landro, P.v.R. Schleyer, J.A. Pople, and R.J.
Lagow, ~Am. Chern. Soc. 104:4275 (1982).
25. P.v.R. Schleyer, E.D. Jemmis, J. Chandrasekhar, E.-U. Wurthwein,
A.J. Kos, B.T. Luke, and J.A. Pople, J. Am. Chern. Soc. 105:484
(1983). - - -- --
26. A.E. Reed and F. Weinhold, J. Am. Chern. Soc. 107:1919 (1985).
27. J.W. Chinn, Jr. and R.J. Lagow~.~Chern. Soc. 106:3694 (1984).
28. Swanson et al., Nat!. Bur. Std. Mon~25:1 (1962).
29. Stackelberg and Quatram, Z.~s~hern. Leipzig B27:50 (1934).
30. R. Zerger, W. Rhine, and G.D. Stucky, J. Am. Chern. Soc. 96:6048
(1974). - - -- --
31. (a) E. Weiss and A.C. Lucken, ~ Organomet. Chern. 2:197 (1964); (b)
E. Weiss and G. Hencken, ~ Organomet. Chern. 21:265 (1970).
32. Refinement of this model was carried out with the soft constraint
that D(1)-D(2) be separated by 1.78(25) A. This was required
because of the close approach of D(1) to the twofold axis passing
through the carbon atom.
33. E. Weiss, Acta Cryst. 16:681 (1963).
34. G.S. Smith, Q.C. Johnson, D.K. Smith, D.E. Cox, R.L. Snyder, R.-S.
Zhou and A. Zalkin, Brookhaven Synchrotron Light Source Annual
Report, 227 (1988).
35. E. Zintl and A. Harder, ~ Phys. Chern. 28:478 (1935).
36. (a) W. Rhine, D. Groves and G.D. Stucky, J. Am. Chern. Soc. 93:1553
(1971); (b) W. Rhine, S.W. Peterson and G:D.-siucky, J. Am. Chern.
Soc. 97:6401 (1975). -- --
37. ~P.v.R. Schleyer, E.-U. Wurthwein, E. Kaufmann, T. Clark, and
J.A. Pople, ~Am. Chern. Soc. 105:5930 (1983); (b) P.v.R. Schleyer,
"New Horizons of Quantum Chemistry," P.-O. Lowdin and B. Pullman,
eds., D. Reidel, New York (1983), pp. 95-109; (c) E.-U. Wurthwein,
P.v.R. Schleyer, and J.A. Pople, J. Am. Chern. Soc. 106:6973 (1984);
(d) P.v.R. Schleyer, E.-U. Wurthwein~nd J.A. Pople, J. Am. Chern.
Soc. 104:5839 (1982). -- --
38. Private communication between RJL and P.v.R. Schleyer prior to the
Lagow communication (ref. 24).
39. (a) J.A. Gurak, J.W. Chinn, Jr., C.S. Yannoni, H. Steinfink, and
R.J. Lagow, Inorg. Chern. 23:3717 (1984); (b) C.S. Yannoni, J.A.
Gurak, J.W. Chinn, Jr., R.A. Kendrick, and R.J. Lagow, Inorg. Chim.
Acta 96:L75 (1985).
40. W.S. Veeman and E.M. Mengerand, ~ Magn. Reson. 46:257 (1982).
41. L.D. McKeever, R. Waack, M.A. Doran, and E.B. Baker, J. Am. Chern.
Soc. 91:1057 (1969). -- --
42. J.W. Chinn, Jr., G.F. Meyers, M.B. Hall, and R.J. Lagow, ~Am.
Chern. Soc. 107:1413 (1985).
43. H.P.S. Chauhan, H. Kawa, and R.J. Lagow, ~ Org. Chern. 51:1632
(1986).
194
ORGANOMETALLIC CHEMICAL VAPOR DEPOSffiON
Alan H. Cowley
Department of Chemistry
The University of Texas at Austin
Austin, TX 78712
INTRODUCTION
Gallium arsenide and related ternary materials are used extensively in the electronics
industry (Table I). Other III- V semiconductors such as indium phosphide are employed in
the telecommunications industry and current interest in indium antimonide stems from its
potential use in infrared sensors. Broadly speaking, two approaches are available for the
preparation of III- V materials, namely molecular beam epitaxy (MBE) and organometallic
chemical vapor deposition (OMCVD). Of these approaches, the OMCVD method is more
suited to the production of larger quantities of compound semiconductors. Typically, the
OMCVD method is based on the reaction of Me 3Ga or Me 3In with PH3 or AsH3 at 600-
700 ·c. e.g.l
M= Ga, In; E = P, As
196
are of the general type RnMxEyRm'• where M and E are group III or group V elements,
respectively, and Rn and Rm' are substituents that are capable of facile hydrocarbon
elimination. Our precursors feature the additional advantages of ease of purification,
short-term air stability, and relatively low toxicity.
Synthetic Chemist:ty
It is clear from the current review literature2 that, with some notable exceptions,3 little
attention has been paid to the organometallic chemistry of gallium and indium. The initial
stages of our program were therefore concerned with exploration of the range of possible
III-V organometallic compounds. As illustrated in Figure 1, compounds of 1:1, 1:2, 1:3,
and 1:4 stoichiometry have now been isolated and characterized. Our entry into this field
r
was provided by the metathetical reaction of lithium phosphide or arsenide with the
appropriate metal trichloride which afforded compounds of 1:3 stoichiometry, 1.4
t-Bu2
/M~
t-Bu 2E E-t-Bu 2
In the same year, Wells et al.5 described the synthesis ofGa[As(mesityl) 2] 3 via the reaction
of (mesityl)2AsSiMe3 with GaCl3. Compounds of type 1 are useful reagents for the
synthesis of mixed metal compounds and clusters with 1: 1 stoichiometry. For example,
treatment of 1 with trialkylgallium or trialkylindium affords very high yields of 1:1
compounds of the type [R 2ME-t-Bu 2]z (2). 6 These compounds can also be made by
treatment of GaCl3 with one equivalent of t-Bu2ELi (E=P,As) and two equivalents ofRLi(R
=Me, n-Bu).4,6
197
'
t-Bu t-Bu
Me Me
R
' M'\./ R Sb /
'/ '\.I
R-E E-R t-Bu2sb....._ / "\. /'
Me-.M
I I
M-Me
_;~Ja Ga,
Cl " / Sb-t-Bu2
Me" " ' / 'Me Sb
E
/' t-Bu
/ 't-Bu
R R
1:1 1:2
M=Ga,ln;
E=P,As,Sb
1:4 1:3
E-t-Bu2
~
t-Bu2E/ ~E-t-Bu2
Figure 1. Ill-Y Compounds of Various Stoichiometries
198
Alkane or arene elimination also represents a viable synthetic approach to Ill-V cluster
compounds and indeed this method was used in pioneering studies in the 1960's by Coates
and co-workers.3 We and others7 have since used the alkane or arene elimination method to
prepare several 1: 1 compounds, a recent example of which is shown below. 8
At the present time, rather few compounds of 1:2 stoichiometry have been prepared.
(X=Cl,Br; R= mesityl, Me3SiCH2) either by salt or silyl halide elimination, and in our
laboratories we have prepared an indium-antimony compound of the same stoichiometry (3)
by silyl halide elimination.lO
t-Bu t-Bu
CI., 'Sb,
/ ••• Sb-t-Bu2
II,, / " t.''
In In
complexes, 1. 4,6 If the less bulky phosphide or arsenide reagents, Ph2ELi, are employed in
the same stoichiometry, only intractable products can be obtained. However, when the
reactant ratio is changed to 1:4 crystalline "ate" complexes can be isolated.11
199
THF
M = Ga, In; E = P, As
In addition to the foregoing, mention should also be made of the 3:1 stoichiometry
complex, [(THF)Br2Gal)As which was synthesized via the reaction of GaBr3 with
(Me3Si)3As. 12
Structural Information
By means of X-ray crystallography, it has been established that gallium-phosphorus4
and gallium-arsenic4, 13 dimers of 1:1 stoichiometry adopt diborane-like structures, 4. An
approximately tetrahedral geometry is evident at both theM and E centers and the M-E bond
lengths are close to the sums of the respective single-bond covalent radii. The geometry of
each M2E2 ring is planar and there is no indication of M ... MorE ...E cross-ring bonding.
These essential structural features have since been found for several other such dimers and
also for the 2:1 stoichiometry complexes [BrGa{As(CH2SiMe3)2hl9 and
[Clln{Sb-t-Bu2hh.10 In the latter complexes the exocyclic R2'E groups adopt a mutually
trans disposition. The weakly bonded dimer of 1:3 stoichiometry,
[{(Me3SiCH2)2As}3Gah, also possesses structure 4 14. However, in this case the
The trimers exhibit a wide range of conformations for the M 3E 3 rings. For example,
200
conformations respectively. The case of [Me2InAsMe2]3 is particularly interesting because
of the presence of puckered and approximately planar conformations in the same asymmetric
unit.15 Dynamic 1H NMR data for [Me 2InAsMe 2] 3 establish that the In-methyl groups
remain equivalent to -80 ·c. Clearly, it will be of interest to acquire more structural
information on the trimers. However, the available data imply that the various possible
M3E3 ring conformations are close in energy.
Finally, the "ate" anions, [M(EPh2)4L have been found to possess C1 skeletal
structures. II These anions are therefore much less symmetrical than transition metal
analogues which adopt D 2d skeletal geometries.l6
Deposition Studies
At the present time, most of our deposition studies have focussed on 5 as the
single-source precursor. I? Films of gallium arsenide have been grown in a cold-wall
OMCVD reactor.
t-Bu t-Bu
"-/
As
Me ·.,, / ' ..• Me
'Ga/ 'Ga''
Me,;::?' "'-. / ~Me
As
t-Bu
/"'-. t-Bu
This reactor was designed and built in the laboratories of Professor John G. Ekerdt of the
Department of Chemical Engineering at the University of Texas at Austin, and it is a pleasure
to acknowledge this important collaboration. The reaction conditions are summarized in
Table IT, together with a listing of the substrates employed and the substrate cleaning
protocols. The use of the conditions outlined in Table IT resulted in typical fllm growth rates
of 0. 7 to 1.0 Jlm per hour (as measured by Dektak).
The gross compositions of the gallium arsenide films were established by X-ray
photoelectron spectroscopy (XPS). Within experimental error, a Ga:As ratio of 1:1 was
confirmed in each case by measurement of the intensities of the Ga3d and As3d signals at
18.8 and 40.9 eV, respectively. Carbon was not detected in the XPS experiments which
implies that the carbon impurity levels are less than 1000 ppm. Secondary ion mass
spectroscopy (SIMS) is a more sensitive technique for the detection of impurities. SIMS
201
Table II. OMCVD Reaction Conditions
202
studies of our films revealed the carbon levels to be comparable to those found in electronic
grade gallium arsenide. Apart from carbon, traces of Na, Al, Si, K, Cr, Fe, and In were
also detected. Undoubtedly, these impurities were present in the starting materials or were
introduced during the synthetic procedures. X-ray diffraction (XRD) measurements indicate
that polycrystalline GaAs growth occurred on all three substrates. Interestingly, the GaAs
films grown on semi-insulating GaAs(lOO) and Si-doped GaAs(lOO) exhibit
low-temperature photoluminescence (PL) signals, thus implying the presence of crystalline
domains.
Preliminary thermal decomposition studies on 5 show that methane and isobutylene are
the main hydrocarbons produced during layer growth. This, plus the solid state crystal
structure of the compound which indicates a relatively short H ...Ga interaction of a
tert-butyl-As hydrogen atom4 lead us to propose tentatively a decomposition pathway
involving transfer of this H atom to the Ga and expulsion of CH4. This process,
accompanied by a cleavage of the As-C bond producing isobutylene would then result in the
clean removal of the hydrocarbon groups from the inner Ga2As2 core.
Finally, attention is drawn to the very recent work of Bradley et al.18 on the use of the
indium-phosphorus analogue of 5 for the growth of epitaxial indium phosphide using metal
organic molecular beam epitaxy (MOMBE).
Ac1mowled&ment
I am most grateful to the organizers of the IUCCP Symposium for the opportunity to
help recognize Professor F. A. Cotton's seminal work on the quadruple metal-metal bond.
Gratitude is also expressed to the National Science Foundation, the Robert A. Welch
Foundation, and the Texas Advanced Technology Research Program for fmancial support.
It is also a pleasure to thank Professors Richard A. Jones and John G. Ekerdt for their
collaboration and the following co-workers for their dedication and hard work: B. Benac, R.
Geerts, D. Giolando, P. Harris, D. Heaton, K. Kidd, M. Mardones, C. Nunn, J. Power, S.
Schwab, and D. Westmoreland.
References
203
5. C. G. Pitt, A. P. Purdy, K. T. Higa, and R. L. Wells, Or~anornetallics 5:1266 (1986);
C. G. Pitt, K. T. Higa, A. T. McPhail, and R. L. Wells, Inorg. Chern. 25:2484 (1986).
6. A.M. Arif, B. L. Benac, A. H. Cowley, R. A. Jones, K. B. Kidd, and C. M. Nunn,
New J. Chern. 12:553 (1988).
7. 0. T. Beachley, J.P. Kopasz, H. Zhang, W. E. Hunter, and J. L. Atwood, L.
Organornet. Chern. 325:69 (1987).
8. A. H. Cowley, R. A. Jones, M.A. Mardones, and C. M. Nunn, to be published.
9. A. P. Purdy, R. L. Wells, A. T. McPhail, and C. G. Pitt, Organornetallics 6:2099
(1987).
10. A. R. Barron, A. H. Cowley, R. A. Jones, C. M. Nunn, and D. L. Westmoreland,
Polyhedron 7:77 (1988).
11. C. J. Carrano, A. H. Cowley, D. M. Giolando, R. A. Jones, C. M. Nunn, J. M.
Power,Inor~. Chern. 27:2709 (1988).
12. R. L. Wells, S. Shafieezed, A. T. McPhail, and C. G. Pitt, J. Chern. Soc., Chern.
Cornrnun. 1823 (1987).
13. R. L. Wells, A. P. Purdy, A. T. McPhail, and C. G. Pitt, J. Or~anornet. Chern.
308:281 (1986).
14. R. L. Wells, A. P. Purdy, K. T. Higa, A. T. McPhail, and C. G. Pitt, J. Organornet.
Chern. 325:C7 (1987).
15. A. H. Cowley, R. A. Jones, K. B. Kidd, C. M. Nunn, and D. L. Westmoreland,
J. Organomet. Chern. 341:C1 (1988).
16. M. H. Chisholm, F. A. Cotton, andM. W. Extine, Inorg. Chem 17:1329 (1978);
R. T. Baker, P. J. Krusic, T. H. Tulip, J. C. Calabrese, and S. S. Wreford, J. Am.
Chern.Soc. 105:6763 (1983).
17. A. H. Cowley, B. L. Benac, J. G. Ekerdt, R. A. Jones, K. B. Kidd, J. Y. Lee, and J. E.
Miller, J. Am. Chern. Soc. 110:6248 (1988).
18. D. A. Andrews, G. J. Davies, D. C. Bradley, M. M. Fak:tor, D. M. Frigo, and E.A.D.
White, Semicond. Sci. Techno!. 3:1053 (1988).
204
SURFACE CHEMISTRY OF MIXED-METAL SYSTEMS
D. W. Goodman
Department of Chemistry
Texas A&M University
College Station, TX 77843-3255
INTRODUCTION
Interest in bimetallic catalysts has risen steadily over
the years because of the commercial success of these systems
[1]. This success results from an enhanced ability to control
the catalytic activity and selectivity by tailoring the
catalyst composition [2,3]. A long-standing question regarding
such bimetallic systems is the nature of the properties of the
mixed metal system which give rise to its enhanced catalytic
performance relative to either of its individual metal
components. These enhanced properties (improved stability,
selectivity andjor activity) can be accounted for by one or
more of several possibilities.
First, the addition of one metal to a second may lead to
an electronic modification of either or both of the metal
constituents. This electronic perturbation can result from
direct bonding (charge transfer) or from a structural
modification induced by one metal upon the other.
Secondly, a metal additive can promote a particular step
in the reaction sequence and, thus, act in parallel with the
host metal. Thirdly, the additive metal can serve to block the
availability of certain active sites, or ensembles,
prerequisite for a particular reaction step. If this
"poisoned" reaction step involves an undesirable reaction
product, then the net effect is an enhanced overall
selectivity. Further, the attenuation by this mechanism of a
reaction step leading to undesirable surface contamination will
promote catalyst activity and durability.
The studies reviewed here are part of a continuing effort
[7-10] to identify those properties of bimetallic systems which
can be related to their superior catalytic properties.
EXPERIMENTAL
These studies were carried out utilizing the specialized
206
co-cu
0
A.u 1 ML Cu/Au
<l
Temperature (K)
0
o..u
<I
TEMPERA TURE-K
207
Most apparent in Figure 2 is a monotonic decrease upon addition
of Cu of the CO structure identified with Ru (peaks at 400 and
480K) and an increase of the CO structure corresponding to Cu
(peaks at 160 and 210K). The buildup of a third feature at
-300K (indicated by the dashed line) is assigned to correspond
to co desorbing from the edges of Cu islands. Integration of
the 200, 275, and 300K peaks provides information regarding
island sizes, that is, perimeter-to-island area ratios, at
various cu coverages. For example, at ecu = 0.66 the average
island size is estimated to be approximately 50A in diameter.
This island size is consistent with an estimate of the 2-d
island size corresponding to this coverage of 40-60A derived
from the width of the LEED beam profiles [7].
Model studies of the CujRu(0001) catalyst have been
carried out [8] for methanation and hydrogenolysis reactions.
These data suggest that copper merely serves as an inactive
diluent, blocking sites on a one-to-one basis. Similar results
have been found in analogous studies [21] introducing silver
onto a Rh(111) methanation catalyst.
Sinfelt [22] has shown that copper in a CujRu catalyst is
confined to the surface of ruthenium. Results from the model
catalysts discussed here then should be relevant to those on
the corresponding supported, bimetallic catalysts. Several
such studies have been carried out investigating the addition
of copper or other Group 18 metals on the rates of CO
hydrogenation [23-25] and ethane hydrogenolysis [25] catalyzed
by ruthenium. In general, these studies show a marked
reduction in activity with addition of the Group 18 metal
suggesting a more profound effect of the Group 18 metal on
ruthenium than implied from the model studies. A critical
parameter in the supported studies is the measurement of the
active ruthenium surface using hydrogen chemisorption
techniques. Haller and coworkers [26,27] have recently
suggested that hydrogen spillover during chemisorption may
occur from ruthenium to copper complicating the assessment of
surface Ru atoms. Recent studies in our laboratory [5,6] have
shown directly that spillover from Ru to Cu can take place and
must be considered in the hydrogen chemisorption measurements.
H2 spillover would lead to a significant overestimation of the
number of active ruthenium metal sites and thus to significant
error in calculating ruthenium specific activity. If this is
indeed the case, the results obtained on the supported
catalysts, corrected for the overestimation of surface
ruthenium, could become more comparable with the model data
reported here. Finally, the activation energies observed on
supported catalysts in various laboratories are generally
unchanged by the addition of Group 18 metal [26-28] in
agreement with the model studies.
These arguments suggest that Ru specific rates for
methanation and ethane hydrogenolysis on supported CujRu
catalysts approximate those values found for pure Ru. As a
consequence, the rates for cyclohexane dehydrogenation reaction
on supported CujRu, similarly corrected, must exceed those
specific rates found for pure Ru. The uncorrected specific
rates for cyclohexane dehydrogenation on the supported CujRu
system remain essentially unchanged upon addition of cu to Ru
[28]. An activity enhancement for cyclohexane dehydrogenation
in the mixed CujRu system relative to pure Ru is most
208
surprising given that Cu is less active for this reaction than
Ru.
209
studied is nickel adsorbed onto tungsten [32,33]. On both
W(llO) and W(lOO), Ni is adsorbed layer-by-layer. Annealing Ni
layers with coverages less than 1.3 ML to 1200K produces little
change in the Ni(848eV)/W(179eV) AES ratio. However, for Ni
coverages above 1.3ML, a 1200K anneal results in a very slow
increase in this AES ratio with coverage, indicating either
alloy or 3-dimensional island formation.
210
10.0
8.0
•.
!
c
-.•
0
u
8.0
.
Gl
~ 4.0
c-;
a:
2.0
0.0
o.o 1.2
>
()
~ 10_,
::I
a
Ill
...a:
a::
Ill
>
0
z
a:
::I
..... 10 2
Ill
z
c
:z:
.....
Ill • .:I IlL NI/W ( 1 101
:1! 0 .I IlL NI/W (1101 n
A 1.0 IlL NI/W 11101
0 .4 IlL NI/W ( 1001
• . 7 IlL NIIW ( 1001
+ .!I IlL NI/W ( 1001
211
-•
-
'
>-
CJ 10 1
z
Ill
:I
0
Ill
a::
~
a::
Ill
> Hi 2
0
z
a::
:I
1-
Ill
z
c
...
:z: Hi 3
Ill
::1
NIIW(1101
e 0.1 ML Nl
o 0.4 ML Nl
• 1.2 ML Nl
10- 4
1.5 1.7 1.9 2.1
212
activity found for bulk Ni. The activation energy for this
reaction is lower on the strained metal overlayer, however,
very likely reflecting the lower binding strength of CO on the
bimetallic system.
In contrast, ethane hydrogenolysis, which is a structure
sensitive reaction over bulk Ni, displayed marked structural
effects on the Ni/W system [41]. We have observed, as shown in
Figure 5, that the specific rate, or rate per surface metal
atom, but not the activation energy, is a strong function of
metal coverage on the Ni/W(110) surface, suggesting that the
critical reaction step involves the need for a single,
sterically unhindered Ni atom. On the Ni/W(100) surface the
specific reaction rate was independent of Ni coverage. In
addition, the rate on bulk Ni(100), Ni/W(110) in the limit of
zero coverage, and Ni/W(100) were all equal, as were the
activation energies. This implies that on Ni/W(100) the Ni
atom geometry is sufficiently open to allow unhindered access
to each Ni atom. Apparently on the Ni/W(110) surface only
island edges and individual atoms display activity similar to
the Ni(100) surface: the island interiors, in contrast, exhibit
behavior similar to Ni(111) which has a much lower specific
rate and higher activation energy. As the Ni coverage is
reduced, the number of active, Ni(100)-like atoms increases,
leading to an increase in the specific rate. The activation
energy, however, remains unchanged.
We have studied several other metal overlayers on W(110),
W(100), and Ru(0001) substrates [42]. Table 1 lists properties
of the metal overlayers, and the effect of the substrate on co
chemisorption. In general only the first monolayer grows
pseudomorphically, though more than one monolayer may be stable
before three dimensional islands are formed (e. g. Cu/Ru grows
two stable layers). The binding strength of co is always
altered from the bulk metal, though the magnitude of the effect
is seemingly more dependent on the metal overlayer, than on the
degree of strain induced by the substrate (represented as the
atom density mismatch). As with Ni/W and Cu/Ru, the effect on
co binding energy extends primarily to only the first
monolayer: subsequent layers exhibit behavior close to the bulk
metal.
ACKNOWLEDGEMENT
We acknowledge with pleasure the support of this work by
the Department of Energy, Office of Basic Energy Sciences,
Division of Chemical Sciences.
REFERENCES
1. T. B. Grimley and M. Torrini, J. Phys. Chern., 87, 4378
(1973).
2. J. H. Sinfelt, "Bimetallic Catalysts: Discoveries, Concepts,
and Applications", John Wiley & Sons, NY, NY, 1983.
3. G. M. Schwab, Disc. Faraday Soc. 8, 166 (1950).
4. J. T. Yates, Jr., c. H. F. Peden, and D. W. Goodman, J.
catal., 94, 576 (1985).
5. D. w. Goodman, J. T. Yates, Jr., and c. H. F. Peden, Surf.
sci., 164, 417 (1985).
6. D. w. Goodman and c. H. F. Peden, J. Catal., 95, 321 (1985).
213
7. J. E. Houston, c. H. F. Peden, D. s. Blair, and D. W.
Goodman, Surf. sci., 167, 427 (1986).
8. J. E. Houston, c. H. F. Peden, P. J. Feibelman, and D. W.
Goodman, I&EC Fundamentals, 25, 58 (1986).
9. J. E. Houston, C. H. F. Peden, P. J. Feibelman, and D. R.
Hamann, Phys. Rev. Lett., 56, 375 (1986).
1o.c. H. F. Peden and D. w. Goodman, J. catal., 104, 347 (1987).
11.D. w. Goodman, R. D. Kelley, T. E. Madey, and J. T. Yates,
Jr., J. Catal. 64, 479 (1980).
12.D. w. Goodman, Ann. Rev. Phys. Chern., 37 (1986) 425; D. W.
Goodman, Accts. Chern. Res., 17, 194 (1984); D. W.
Goodman, J. Vac. Sci. Tech., 20, 522 (1982).
13.H. Shimizu, K. Christmann and G. Ertl, J. Catal., 61, 412
(1980).
14.J. c. Vickerman, K. Christmann and G. Ertl, J. Catal., 71,
175 (1981).
15.S. K. Shi, H. I. Lee and J. M. White, Surf. Sci., 102, 56
(1981).
16.L. Richter, s. D. Bader and M. B. Brodsky, J. Vac. Sci.
Techn., 18, 578 (1981).
17.J. c. Vickerman and K. Christmann, surf. sci., 120, 1 (1982).
18.J. c. Vickerman, K. Christmann, G. Ertl, P. Heiman, F. J.
Hirnpsel, and D. E. Eastman, Surf. Sci., 134, 367 (1983).
19.S. D. Bader and L. Richter, J. Vac. Sci. Techno!., A1, 1185
(1983).
20.C. Park, E. Bauer, And H. Poppa, Surf. Sci., submitted for
publication.
21. D. w. Goodman in "Heterogeneous Catalysis" (Proceedings of
IUCCP Conference), Texas A&M University, 1984.
22.J. H. Sinfelt, G. H. Via, and F. W. Lytle, Catal. Rev.-sci.
Eng., 26, 81 (1984).
23.A. K. Datye and J. Schwank, J. Catal., 93, 256 (1985).
24.G. c. Bond and B. D. Turnham, J. Catal., 45, 128 (1976).
25.L. J. M. Luyten, M. v. Eck, J. v. Grondelle, and J. H. c. v.
Hooff, J. Phys. Chern., 82, 2000 (1978).
26.A. J. Rouco, G. L. Haller, J. A. Oliver, and c. Kemball, J.
catal., 84, 297 (1983).
27.G. L. Haller, D. E. Resasco, and J. Wang, J. Catal., 84, 477
(1983).
28.J. H. Sinfelt, J. Catal., 29, 308 (1973).
29.Sachtler, J.W.A. and Somorjai, G.A., J. Catal. 89, 35 (1984).
30.Balooch, M., Cardillo, M.J., Miller, D.R., and Stickney,
R.E., Surf. Sci. 50, 263 (1975).
31.P. J. Berlowitz and D.W. Goodman, Surf. Sci. 187, 463 (1987).
32.M. Balooch, M.J. Cardillo, D.R. Miller, R.E. Stickney, Surf.
Sci. 50, 263 (1975).
33.J. Kolaczkiewicz and E. Bauer, Surf. Sci. 144, 495 (1984).
34.K. Christmann, 0. Schober, G. Ertl, J. Chern. Phys. 60,
(1974).
35.D. w. Goodman, J.T. Yates, Jr., T.E. Madey, Surf. Sci. 93,
L135 (1980).
36.I. Harnedeh and R. Gomer, Surf. Sci. 154, 168 (1985).
37.J. E. Houston, J.M. White, P.J. Berlowitz, D. w. Goodman,
surface sci., in press.
38.M. w. Ruckman, M. Strongin, X. Pan, ibid.
39.D. Prigge, W. Schlenk, E. Bauer, Surf. Sci. 123, L698 (1982).
40.R. w. Judd, M.A. Reichelt, R.M. Lambert (to be submitted for
publication).
41.C. M. Greenlief, P. J. Berkowitz, D. W. Goodman, J. Phys.
Chern. 91, 6669, (1987).
42.P. J. Berlowitz, c. H. F. Peden, and D. w. Goodman, Mat. Res.
Soc. Syrnp. Proc., 83, 161 (1987).
214
ORGANOMETALLIC CHEMICAL VAPOR DEPOSITION OF
ALUMINUM NITRIDE AND ALUMINUM METAL
INTRODUCTION
Research Goals and Approach. Thin films of solids are important because
of a particular physical property, such as electrical or thermal conductivity,
mechanical strength, or refractive index. These properties are determined by
both the atomic (crystal) structure of the material as well as the microstructure
(i. e. grain size, pore volume, etc.). The fundamental question which forms the
foundation of our research program in materials is: What is the relationship
between the chemical reaction mechanism of the film forming step and the
atomic and microstructural features of the film? Because we can systematically
change the mechanism of a reaction by modifying the molecular structure of
the precursor, we may develop an additional parameter upon which the
physical properties of the film would depend.
Metallic aluminum films are used primarily for their high conductivity
and reflectivity. Organoaluminum precursors, especially triisobutylaluminum
(TIBA), have received a great deal of attention for their ability to generate high
quality AI films.ll-1 7 Typical deposition conditions for the low pressure CVD of
AI on Si or other substrates using TIBA include substrate temperatures of 260°C,
TIBA temperature of 4S°C, with argon as the carrier gas at pressures up to 1
torr. These conditions result in deposition rates of up to 0.15 JJ.m/min.l4 The
films produced by such methods have low resistivities (2.8 - 3.5 JJ.C-cm) and
other properties which compare well with AI films prepared by evaporation.l6
A disadvantage observed in the use of TIBA is the rough surface morphology
which leads to poor reflective properties.ll We have developed a new process
for the CVD of Al that gives highly reflective films at temperatures less than
100°C at rates that are two orders of magnitude faster than TIBA 18.
EXPERIMENTAL
Two reactors were used to study the CVD of AlN and Al. The "survey"
reactor had a hot-wall quartz tube which was usually operated without carrier
216
gas and had a base pressure of 5 x w-5 torr. A removable liquid Nz trap placed
between the quartz tube and the pump allowed the trapping and quantitation of
all volatile products except Nz, Hz, and CH4. A schematic of the second reactor
in which all growth rate measurements for AlN depositions were performed is
shown in Figure 1.
Heating Unit
Figure 1
The analytical data for the AlN films, obtained using X-ray photoelectron
spectroscopy (XPS), are summarized in Table 1, and they emphasize that the
films produced in the two different reactors have similar compositions.
217
RESULTS AND DISCUSSION
Aluminum Nitride
Summarized below are the results of several studies on AIN film growth.
Most of the work involves the use of dialkylaluminum azides as the precursor.
These, along with the dialkylaluminum amides, have been found to be the most
successful of the organoaluminum precursors to AIN. At the end of this section
we have included a comparison of films preduced with precursors containing a
N-C bond.
Gas Phase Products. The analysis of the gases produced from the reaction
of [Et2AIN3l3 ( T(precursor) = 40°C, T(reactor) = 500°C) was obtained using gas
chromatography and reproducibly showed 70% C2H4, 30% C2H6, and a very
small amount of butane and butenes (C4/C2 = 0.03). Based on the mass of the
precursor consumed during the deposition and the total pressure of the C2H4,
C2H6, and C4 products trapped, 89% of the precursor ethyl groups were
accounted for in this analysis. Similar results were obtained when the furnace
temperature was increased to 650°C (C2H4 = 76%, C2H6 = 24%, C4/C2 = 0.02, 95%
recovery); or when the deposition was conducted under 1 torr of H2 carrier gas
(C2H4 = 65%, C2H6 = 35%, C4/C2 = 0.01). For comparison, analysis of the gaseous
products from CVD of [Et2AINH2h20 under similar conditions, indicates the
formation of C2H6 was significantly favored over C2H4 (C2H4 = 25%, C2H6 = 75%,
C4/C2 = 0.02, 81% recovery).
Table 1. Atomic Compositions of Thin Films Using XPS after Ar+ sputtering.
218
These results are consistent with a P-hydrogen elimination mechanism
operative in cleaving the Al-Et bond. An alternative mechanism, homolytic Al-
Et bond cleavage, would form ethyl radicals which would couple or
disproportionate to yield C4H 10 or C2H4/C2H6. respectively. The gas phase
reactivity of ethyl radicals has been studied thoroughly by Lalonde and
Price21, from which they concluded that radical coupling is favored over
disproportionation by a factor of 10. The analogous ratio under the
experimental conditions employed in the current CVD studies is 0.03 indicating
that liberation of gaseous ethyl radicals is not the predominant mechanism for
Al-Et bond cleavage. The dramatic change in the C2H4 to CzH6 ratio that occurs
upon changing the N-source to an NH2 group is consistent with a change in the
Al-Et bond cleavage path. Presumably, the availability of the weakly acidic
hydrogens of the NH2 group allows the facile protonolysis of the Al-C bond20.
Nitrogen Sources Containing N-C Bonds. Both azide (N3) and amide (NHz)
are comparably effective as sources for nitrogen in the deposition of AlN. It
was of interest to explore how much the variation in structure of the N-source
would influence the AlN deposition, and we were particularly interested in
determining the effect of incorporating a N-C bond into the precursor. The
219
I 'H
~
-N2 1-l:!C
Et 2 AIN 3 (g) _____. Et 2 AJN 3 (s) _____. 'AIN (s)
I
Et
Scheme 1.
.d
......
-<
~
"' 8
Azide!H 2 50 seem
H 2 Makeup 7.5 seem 0.2
Precursor Temp 50°C
Reactor Pressure 3 torr
7+---~--~--T---~--~--T---T-~T-~
0.9 1.1 1.3 1.5
1000/T(K-1)
Figure 2
220
precursors studied included [Me2Al(NMe2)l2 23, [Me2Al(Azir)]3 (Azir =
aziridine)24, Et3AlNH2(t-Bu)25, and {Et2Al[NH(t-Bu)] )225. All of these
structures contain direct Al-N bonds, and with the exception of Et3AlNH2(t-Bu),
all exist as cyclic four- or six-membered rings. The structures differ from the
azide and amide precursors primarily due to the presence of the N-C bond. The
substantial impact made by the presence of the N-C bond is highlighted in Table
1. All of these films contained large concentrations of carbon, and one
contained only small amounts of nitrogen. In the case where the nitrogen
source is the t-butylamine-triethylaluminum donor acceptor complex, the small
concentrations of N indicate that a facile Al-N cleavage is operative, e. g.
simple dissociation of the donor-acceptor bond. It should be noted that
[Me2Al(NMe2)]2, which also gave films with low N values, was the most stable of
all the precursors examined. Even at oven temperatures of 450°C, this
precursor would pass through the hot tube unchanged. The lower N content of
the film may be the result of an Al-N bond cleavage process involving fi-
hydrogcn elimination of the NMe2 ligand. The intermediate case of [Et2Al(NH-
t-Bu)]2 is interesting because it contains no P-hydrogens on the amido ligand.
It is also related to [Mc2Ga(As-t-Bu2)]2 which was recently reported to give
GaAs films26. The results show that most of the nitrogen is incorporated into
the film; unfortunately, the carbon content is also high. The compound
containing the 3-mcmber, highly strained aziridine ring, [Me2Al(Azir)]3, gave
results similar to those found for [Me2Al(NMe2)]2.
Aluminum
NMe 3
H.. I
~'''Al--H (1)
H.;' I
NMe 3
221
As these complexes contain no AI-C bonds, the elegant studies29 on the surface
reactivity of TIBA would suggest they might eliminate H2 at lower temperatures
to form AI films (assuming the relatively weak donor-acceptor bond is also
readily cleaved). In fact, from the early studies describing these compounds, it
was known that they decomposed to form AI at temperatures >100°C.27 Patents
describing the use of amine-alane complexes for the vapor phase30,31 and
electroless solution AI plating32 have appeared, as has a recent report of the
laser-induced deposition of AI using these compounds.3 3
222
also adsorbed from the atmosphere. All of these elements decrease rapidly to
within the detection limits of the method as the sputtering proceeds. In the
films where TiCl4 was used to pretreat the surface, no Ti or Cl was detected in
the AI films or at the interface with the silicon wafer. This is consistent with
the suggestion made regarding the TiCl4 promoted deposition of aluminum from
TIBA that less than a monolayer of TiCl4 actually adsorbs during the
pretreatment. 1 4
The resistivities of the films were evaluated using a four-point probe and
were in the range of 2.5 - 4.5 JJ.n-cm. The low resistivities observed are
223
40.0
35.0
];o
'ii 30.0
r::
.!
.5 25.0
.>1.
ilD..
20.0
,g
.>1.
c 15.0
II
D..
10.0
5.0
0.0 -,E:;::::~~:;:::::;:=;==::=;=;=:=:~~~;:::::::;~:::::;::::;:=:,.~
0.0 2.0 4.0 6.0 8.0 1 0.0 12.0 14.0 1 6.0 18.0 20.0
sputter time (min.)
Figure 3. Auger sputter profile of an AI film on Si(lOO).
virtually identical to those reported for AI films grown using TIBA, and they
are near that of bulk AI.ll - 11
The surface roughness of many of the films meant that their specular
reflectivity was very low. Typically, the total reflectivity was greater than 90%
at a wavelength of 550 nm, but the specular contribution to this value was less
than 5%. For the films grown at 100°C on TiCl4 treated surfaces, the relative
contributions of the diffuse and specular components to the total reflectivity
were reversed. Similar reflectivities are also obtained from AI deposited on
treated polyimide films.
224
5.00
a
4.00
,.....
.,
"0
.,
c:
0
:I 3.00
0
.r::.
~
.,"u
'!l
c:
:I
2.00
0
u
1.00
0.00
Ll \_ ~ k
30.0 50.0 70.0 90.0
two theto
5.00,-------,-------------------------~----------------~
4.00
.,
,.....
"0
c:
.,0 3.00
:I
0
.r::.
~
u
.,"
'
!l
c:
:I
2.00
0
u
, .00
) ~
0.004-------~--~~--r--------.--~~~~~----~n-----~ h
30.0 50.0 70.0 90.0
two theta
225
4.50
4.00
3.50
.,
..._
"C
3.00
.,c
I:
"0 2.50
2,
u
.," .L.OO
.......
!l
""u0 1.50
, .00
0.50
...._....._
II.
0.00
30.0 50.0 70.0 90.0
two theta
226
(Me3N)2A1H3(g) (Me3N)AIH3(g) + Me3N(g) (2)
Eqs. 2 and 3 are based on the previous studies, whereas eq. 4 is based on the
observed surface chemistry of TIBA29. It was found that upon adsorption of
TIBA on AI surfaces, the three isobutyl groups behaved identically. That
coupled with the fact that molecular TIBA desorption was never observed
suggested that upon adsorption the alkyl groups became equivalent by
migration to adjacent surface aluminum atoms. The similarity between alkyl
and hydride ligands is the basis for proposing the step shown in eq. 4. Finally,
it is known that aluminum surfaces do not dissociatively adsorb H2 (the reverse
of eq. 5) at low pressures. Studies of eq. 5 have been accomplished by reacting
atomic hydrogen with aluminum surfaces at low temperatures. Temperature
programmed desorption studies showed that H2 desorbs from AI surfaces around
60 oc29,37. We note that this is between the lowest temperature where AI films
were observed (100 °C; lower temperatures were not examined) and room
temperature where no AI films formed. While it is tempting to suggest that eq.
5 is the rate limiting step, we point out that the temperature range required for
deposition is also the range where eqs. 2 and 3 become significant. It is
important to note that our data does not rule out other possible steps, such as
direct adsorption of (Me3N)2AIH3 or (Me3N)AlH3. Kinetic studies of the
deposition are planned which should help delineate the mechanism.
227
Si-OH + TiCl4 ~ HCI + Si-O-TiCl3 (6)
COMMENTS
228
involves developing precursors for the deposition of other main group and
transition metal nitrides and selected metals.
ACKNOWLEDGEMENTS
REFERENCES
1. Yim, W. M.; Stofko, E. J.; Zanzucchi, P.J.; Pankove, J. I.; Ettenberg, M; Gilbert,
S. L. J. Appl. Phys. 1973, 44, 292.
2. Weast, R. C. (ed) "Handbook of Chemistry and Physics, 51st Ed." 1970, The
Chemical Rubber Company: Cleveland.
3. Baker, R. T.; Bolt, J. D.; Chowdhry, U.; Klabunde, U; Reddy, G. S.; Roe, D. C.;
Staley, R. H.; Tebbe, F. N.; Vega, A. J. Mat. Res. Soc. Symp. Proc. 1988, 121,
471.
4. Manesevit, H. M.; Erdmann, F. M.; Simpson, W. I. J. Electrochem. Soc. 1971,
118, 1864.
5. Sayyah, K.; Chung, B. C.; Gershenzon, M. J. Crystal Growth 1986, 77, 424.
6. Gaskill, D. K.; Bottka, N.; and Lin, M. C. J. Crystal Growth 1986, 77, 418.
7. Interrante, L. V.; Carpenter,Jr., L.; Whitmarsh, C.; Lee, W.; Garbaukas, M.;
Slack, G. A. Mat. Res. Soc. Symp. Proc. 1986, 73, 359.
8. Interrante, L.V.; Lee,W.; McConnell, M.; Lewis, N.; Hall, E. J. Electrochem.
Soc., 1989, 136, 472.
9. Schulze, R. K.; Mantell, D. R.; Gladfelter, W. L.; Evans, J. F. J. Vac. Sci.
Technol., 1988, A6, 2162.
10. Boyd, D. C.; Raasch, R. T.; Mantell, D. R.; Schulze, R. K.; Evans, J. F.; Gladfelter,
W. L. Chemistry of Materials, 1989,1, 119.
11. Levy, R. A.; Green, M. L. J. Electrochem. Soc. 1987,134, 37C.
12. Pierson, H. 0. Thin Solid Films 1977,45, 257.
13. Malazgirt, A.; Evans, J. W. Me tall. Trans. 1980, 11 B, 225.
14. Cooke, M. J.; Heinecke, R. A.; Stem, R. C.; Maes, J. W. Solid State Technol.
1982, 25(12), 62.
15. Green, M. L.; Levy, R. A.; Nuzzo, R. G.; Coleman, E. Thin Solid Films 1984, 114,
367.
16. Levy, R. A.; Green, M. L.; Gallagher, P. K.; Ali, Y. S. J. Electrochem. Soc. 1984,
131, 2175.
229
17. Kobayashi, T.; Sekiguchi, A.; Hosokawa, N.; Asamaki, T. lap. J. Appl. Phys.
1988, 27, L1775.
18. Gladfelter, W. L.; Boyd, D. C.; Jensen, K. F. Chemistry of Materials, in press.
19. Sugawara, K.; Yoshimi, T.; Okuyama, H.; Shirasu, T. J. Electrochem. Soc. 1974,
121' 1233.
20. Sauls, F. C.; Interrante, L.V., private communication.
21. Lalonde A. C.; Price, S. J. W. Can. J. Chern .. 1971, 49, 3367.
22. Jasinski, J. M; Meyerson, B. S.; Scott, B. A. Ann. Rev. Phys. Chern .. 1987, 38,
109.
23. Davidson, N.; Brown, H. C. J. Am. Chern. Soc. 1942, 64, 316.
24. Atwood J. L.; Stucky, G. D. J. Am. Chern. Soc. 1970, 92, 285.
25. Gosling, K.; Smith, J. D.; Wharmby, D. H. W. J. Chern. Soc. 1969, (A), 1738.
26. Cowley, A. H.; Benac, B. L.; Ekerdt, J. G.; Jones, R. A.; Kidd, K. B.; Lee, J. Y.;
Miller, J. E. J. Am. Chern. Soc. 1988, 110, 6248.
27. Wiberg, E.; Amberger, E. Hydrides of the Elements of Main Groups 1-IV;
Elsevier: Amsterdam, 1971; Chapter 5.
28. Kovar, R. A.; Callaway, J. 0. Inorg. Syn. 1977, 17, 36.
29. Bent, B. E.; Nuzzo, R. G.; Dubois, L. H. J. Am. Chern. Soc. 1989, 1 I I, 1634.
30. Whaley, T. P.; Norman, V. U. S. Patent No. 3 206 326, (1965).
31. Carley, D. R.; Dunn, J. H. U. S. Patent No. 3 375 129, (1968).
32. Schmidt, D. L.; Hellmann, R. U. S. Patent No. 3 462 288, (1969).
33. Baum, T. H.; Larson, C. E.; Jackson, R. L. Abstracts of Papers, Fall Meeting,
Boston, MA; Materials Research Society: Pittsburgh, PA, 1988; 84.12.
34. Powder Diffraction File, International Center for Diffraction Data,
Swarthmore, PA, Card #4-0787.
35. Heitsch, C. W. Nature, 1962, 195, 995.
36. Nechiporenko, G. N.; Petukhova, L. B.; Rozenberg, A. S. Izv. Akad. Nauk SSSR,
Ser. Khimicheskaya 1975, 1697.
37. Paul, J.; Hoffmann, F. M. Surf. Sci. 1988,194, 419
230
SOLID STATE CARBON-13 NMR OF METAL CARBONYLS
Brian E. Hanson
Department of Chemistry
Virginia Polytechnic Institute and State University
Blacksburg, Virginia 24061-0212
INTRODUCTION
From the very first carbon-13 NMR study of metal carbonyls then it
was apparent that not only are metal carbonyls likely to be dynamic in
solution but also that NMR spectroscopy would be an invaluable tool for
the investigation of mechanisms which describe the dynamic behavior of
metal carbonyls. The phenomenon described by the formal breaking and
reforming of bonds while no net chemical change occurs, is referred to
as fluxionality. [5] Fluxionality has proven to be common for metal
carbonyls, and particularly for metal carbonyl clusters, in solution. [6]
For binary metal carbonyls long carbon T1's in the solid state are
overcome by enrichment in 13co to enhance the signal and by long pulse
delays to prevent saturation. In substituted metal carbonyls the cross
polarization experiment[22] can be employed to overcome the long T1
times.
232
and a 33 = ~1- Most of the tensors of carbonyl groups in metal carbonyls
have been shown to have either axial or near axial symmetry. [23]
(1)
G=<CJ> + Ga (2)
m""~~
<2\
54.7°
~ sample
0
Fig. 1. Experimental geometry for the removal of chemical shift
anisotropy through magic angle spinning. B0 represents
the applied magnetic field. Rapid spinning at the angle
54.7° causes the time average of the CSA to go to 0.
The isotropic shift is, of course, the shift observed in solution and
there is good agreement between chemical shifts obtained on powders in
the magic angle spinning experiment with solution data. However,
crystal effects sometimes cause splitting of lines which are degenerate
in solution. This has been demonstrated for several metal carbonyls
including (1] 5 -c5H5)Fe2 (CO) 4, [24] Ru3 (CO) 12, [25,26] and
Os3(COl12· [26,27]
233
extract the chemical shift parameters [28] . This has been accomplished
for many metal carbonyls. [26] Figure 2 illustrates the situation when a
MAS spectrum is recorded in the slow spinning regime. The spectrum
shown is of Rh6(COl16 and was reported by Oldfield et al. [26] For
comparison the breadline NMR of the same molecule as reported by Gleeson
and Vaughan [22] is shown also.
(b)
(a)
Fig.2 (a) The breadline NMR for RhG(CO)l6· [22] (b) Magic angle
spinning NMR for Rh6(C0)16. [26] The centerbands in the MAS
spectrum are labeled 3i and the sidebands are seen to map out
the chemical shift anisotropy of the powder pattern. The
breadline spectrum is plotted with positive chemical shifts
(downfield) to the right of 0 ppm on the x axis.
234
.
'"'-k -- -- -A
(b)
235
(a)
(b)
..........,,...,~/YO"-..~~
~~Aw~
-so•c
Further examination of the MAS NMR results for Fe(C0) 5 shows that
the linewidths of both peaks diminish as the temperature is lowered.
Although in principle a plot of the log of the linewidth vs 1/T will
yield an activation energy for the process responsible for the line
narrowing, in the present case the data is not sufficiently good to put
a great deal of faith in the result . (From 5 data points an activation
energy of 0.6 kcal mol-l is calculated however the correlation
coefficient from the least squares fit is only 0.7.)
JNEL~l fHFe!COl~l- The room temperature MAS NMR spectrum for the
anion in [NEt 4 J [HFe(C0) 4 J shows a single line in the carbonyl
region. [34] This is contrary to what is expected from the structure of
the anion and suggests a dynamic process in the solid state. We have
recently recorded the MAS NMR spectra for this compound as a function of
temperature and demonstrate that it is indeed fluxional in the solid
state. [35] Some of these spectra are shown in Figure 5.
236
either a distorted trigonal bipyramid or as a face capped tetrahedron.
Although the structure of the anion as the tetraethylamonium salt has
not yet been reported it is reasonable to expect that its geometry is
similar to that observed with PPN+ as the cation.
10"
237
DINUCLEAR METAL CARBONYLS
(a)
(b)
Cel;!(COla
----~~------T-~C_)_
TOSS
-40
238
In brief, a homolytic cleavage of the Co-Co bond in solid
dicobaltoctacarbonyl would lead to 2 Co(C0)4 radicals in the solid
state. These would exist as caged radical pairs which could recombine
to give the bridged structure. Prior to recombination the tetrahedral
radicals could rotate about one of four, 3-fold axes to exchange
carbonyl positions. Such a rotation is well established for adamantane
which also has a tetrahedral structure. [37] Rotation in solid Ni(C0) 4
is relatively slow as determined by NMR experiments. [12] However
experiments on solid nickel carbonyl must be performed below its
melting point, ie -25 °C; [12] the rate of rotation may be expected to
be faster at room temperature in Co(C0) 4 . Recombination of the radical
pair would lead to complete interchange of carbonyl positions.
2 Co(C0) 4 (3)
The fact that the EPR signal is quenched upon warming the sample
suggests that there is sufficient mobility to allow radical centers to
recombine to form Co2(CO)s. A plausible mechanism for the movement of
radical centers is the disruption of a radical pair by an adjacent
isolated Co(C0) 4 radical to generate a new Co-Co bond and a new radical.
Ultimately the cobalt radicals would be able to combine as they come
into contact with one another.
239
about the 2-fold rotation axis. [45] Evidence for motion of the metal
core in triirondodecacarbonyl is now available from MAS 13c NMR
spectroscopy. There is additional evidence that suggests dynamic
behavior in the solid state for Co4-xRhx(CO)l2• x = 0,1 or 2.[17]
These are best described by a motion of the metal core with respect to
the carbonyl polyhedron.
Figure 9 shows the 13c MAS NMR spectra of Fe3(COl 1 2 at 27 and -80
°C .[48] These were recorded at 75 MHz. Figure 10 shows the correlation
of the resonances at -80 with the resonances at 27 °C . The high
temperature limiting spectrum at low field shows 6 sharp resonances of
equal intensity. [7, 49] At high field a higher temperature would be
required to obtain the same degree of line narrowing.
240
240 220 200
Fig. 9. The high field (7 tesla, 75 MHz) MAS 13 c NMR spectra for
solid Fe 3 (COJ 12 at 27 and -80 •c.
238.0
236.1
--··.
··.:::~.._-:.,_
.. ··~
obs.
225.2
!calc I
(225.8)
·~:::~J== 224.1 (224.1)
215.6 ----::.::::::=···
••••
----··_::_.-_.._··-~:::::::::=··---- 213.2
210.4 _ _ _ _ _ _ 1~----······-------~----
(213 0)
209.6 (210.4)
203.9
201.3
199.2
__ ..............
==~~=-~~:l-:.::::~::::t====
.
202.2
201.4
(202.1)
(201.5)
-so •c 27 •c
Fig. 10. Proposed correlation of NMR signals from the spectra shown
in Figure 9 at -80 and 27 •c.
241
intensity. As mentioned above a rotation of the iron triangle by 60°
within the plane of the triangle satisfies the symmetry requirement of
generating a time averaged inversion center and making 6 pairs of 2
carbonyl groups equivalent on the NMR timescale. Furthermore the 2
bridging carbonyls labeled 1 and 2 in Figure 11 are exchanged with the
terminal carbony1s labeled 3 and 4. The correlation of the bridging
carbonyls with terminal carbonyls is illustrated in Figure 10. Thus the
peaks at 238.0 and 236.1 ppm have a relative integrated intensity of 2
compared to 10 for the terminal carbonyls at 199 to 215.6 ppm in the
spectrum at -80 °C. At room temperature the peaks at 225.2 and 224.1
ppm have a relative intensity of 4 compared to 8 for the terminal peaks
at 210 to 213 ppm.
242
215 168 ppm
Fig. 12. Variable temperature MAS NMR for co 4 (COl 12 in the solid
state. The asterisked peak is due to delrin.
s5•c
69•c
61•c
Js•c
25•c
243
crystal are staggered such that there is space for some expansion of the
polyhedron in the plane of the irons. It should be noted that although
the polyhedron defined by the carbonyl groups in solid Fe3(C0) 1 2 is
approximated by an icosahedron, the icosahedron is flattened in such a
way that the short dimension is perpendicular to the iron triangle.
The variable temperature spectra were obtained at 22.6 MHz and are
clearly complicated by 59co-13c dipolar and quadrupolar interactions.
This is a serious problem for co 4 (C0) 12 . As noted previously this leads
244
to apparently anomalous chemical shifts for the carbonyl carbons. [15]
Recently a room temperature spectrum for Co4(COl12 at 68 MHz has been
obtained and shown to be consistent with the crystal structure. [51] The
fact that Co 4 (C0) 12 is dynamic in the solid state is inferred from the
behavior of the spinning sidebands in Figure 12. The spinning rate is
constant at 2600 Hz at each temperature. However while at room
temperature the third order sidebands are observed at 75 °C only the
first order sidebands are seen. This suggest a dynamic process which
reduces the chemical shift anisotropy takes place at temperatures above
24 °C. The line shape changes over this temperature range are
completely reversible which suggests that the sample did not decompose
at the high temperatures. The resolution is not sufficiently good to
say that all 12 carbonyl ligands become truly equivalent.
When the data for Co4(COl12 and Co2Rh2(COl12 are taken together it
is apparent that some exchange of carbonyl environments takes place.
This is most likely due, in part, to rotation of the metal polyhedron
with respect to the ligand polyhedron.
CONCLUSIONS
245
400 200 0 ppm
Fig. 15. Room temperature MAS NMR spectra for the clusters
Co4(COl12• Co3Rh(COl12• Co2Rh2ICOl12• and Rh4(CO)l2·
REFERENCES
246
14. Hanson, B. E., Sullivan, M. J., Davis, R. J., J. Am. Chern. Soc.,
1984, 106, 251.
15. Hanson, B. E., Lisic, E. C., Inorg. Chern., 1986, 27, 716.
16. Hanson, B. E., Lisic, E. C., Petty, J. T., Iannaconne, G.; Inorg.
Chern., 1986, 25, 4082.
17. Lisic, E. C., Ph.D. Dissertation, VPI&SU, 1986.
18. (a) Fyfe, C. A.; Lyerla, J. R.; Yannoni, C. S.; J. Am. Chern.
Soc.; 1979, 101, 1351; (b) Lyerla, J. R.; Yannoni, C. S.; Fyfe,
C. A.; Ace. Chern. Res., 1982, 15, 208
19. Campbell, A. J.; Cottrell, C. E.; Fyfe, C. A.; Jeffrey, K. R.;
Inorg. Chern., 1976, 15, 2690.
20. Wagner, G. w., Hanson, B. E., Inorg. Chern.; 1987, 26, 2019.
21. Benn, R.; Grundey, H.; Nolte, R.; Erker, G.; Organornetallics, 1988,
7, 777.
22. Gleeson, J. W., Vaughan, R. W., J. Chern. Phys, 1983, 78, 5384.
23. Yannoni, C. S., Ace. Chern. Res., 1982, 15, 208
24. Dorn, H. C., Hanson, B. E., Motell, E., J. Organornet. Chern, 1982,
224, 181.
25. Airne, S.; Botta, M.; Gobetto, R.; Osella, D.; Inorg. Chirn. Acta,
1988, 146, 151.
26. Walter, T. H., Reven, L., Oldfield, E., J. Phys. Chern., in press.
27. Hasselbring, L., Lamb, H., Dybowski, C.; Gates, B.C.; Rheingold,A.;
Inorg.Chirn. Acta, 1987, 127, 149;
28. Herzfeld, J., Berger, A. E., J. Chern. Phys., 1980, 73, 6021.
29. Meakin, P.; Muetterties, E. L.; Jesson, J. P.; J. Am. Chern. Soc.;
1972, 94,
30 Meakin, P.; Jesson, J. P.; Tebbe, F. N.; Muetterties, E. L.; J. Am.
Chern. Soc.; 1971, 93, 1797.
31. Shapley, J. R.; Osborn, J. A.; Ace. Chern. Res., 1972, 6,
32. Wilkinson, J. R.; Todd, L. J.; J. Organornet. Chern.; 1974,-77, 1
33. Hanson, B. E.; submitted J. Am. Chern. Soc.
34. Whitmire, K. H.; personal communication.
35. Whitmire, K. H.; Hanson, B. E.; submitted J. Am. Chern. Soc.
36 Smith, M. B.; Bau, R.; J. Am. Chern. Soc.; 1973, 95
37. Rothwell, W. P.; Waugh, J. S.; J. Chern. Phys.; 1981, 74, 2721.
38. G. Wagner, personal communication.
39. (a) Leung, P. C.; Coppens, P.; Acta Cryst.; 1983, B39, 535; (b)
Sumner, G. G.; Klug, H. P.; Alexander, L. E.; Acta Cryst.; 1964,
17, 732.
40. Hanson, B. E.; chapter 2 in "Advances in Dynamic Stereochemistry,"
M. Gielen, ed. Freund, London, 1985.
41. Johnson, B. F. G.; J. C. S. Chern. Commun.; 1976, 211.
42 Dixon, W. T.; J. Chern. Phys.; 1982, 77, 1800.
43. Keller, H. J.; Wawersik, H.; z Naturforsch. B., 1965, 20, 938.
44. Connor, J. A.; Topics in Current Chemistry; 1977, 71, 71.
45. Johnson, B. F. G.; J. c. s. Chern. Commun., 1976, 703.
46. Wei, C. H.; Dahl; L. F.; J. Am. Chern. Soc.; 1969, 91, 1351.
47. Cotton, F. A.; Troup, J.; J. Am. Chern. Soc., 1974, 96, 4155.
48. Hanson, B. E.; submitted Inorg. Chern.
49. Hanson, B. E.; Lisic; E. C.; Petty, J. T.; Iannaconne, G. A.;
Inorg. Chern.; 1986, 25, 4062
50. Anson, C. E.; Benfield, R. E. Bott, A. W.; Johnson, B. F. G.;
Braga, D.; Marseglia, E. A.; J. C. S. Chern. Cornrnun., 1988, 889.
51. Airne, S.; Botta, M.; Gobetto, R.; Hanson, B. E.; Inorg. Chern.; in
press.
52. Wei, C. H.; Dahl, L.; J. Am. Chern. Soc., 1966, 88, 1821.
53. Wei, C. H.; Inorg. Chern., 1969, 8, 2384.
54. Carre, F. H. Cotton, F. A.; Frenz, B. A.; Inorg. Chern., 1976, 15,
381.
247
SURFACE CHEMISTRY OF METAL AND SEMICONDUCTOR CLUSTERS
R. E. Smalley
INTRODUCTION
The lOth cluster of nickel is, after all, just a microscopic piece
of the bulk metal. Using the new cluster beam techniques, the surface
chemistry of species like Ni10 is now beginning to be studied, and this
chemistry is turning out to be far more like the catalytic dissociative
chemisorption of bulk nickel surfaces than the chemistry of an isolated
nickel atom or organo-nickel complex 5 . Most small molecules studied by
conventional molecular beam techniques exist in a single well-defined
ground electronic state with the first excited electronic state several
eV higher in energy. In sharp contrast, Ni10 is estimated to have a vast
number of excited electronic states within 1 eV of the ground state.
250
necessary to understand these Ni10 molecules, and obvious place to start
is in the concepts of solid state physics and surface science. Instead
of talking about individual quantum states and wavefunctions, it is more
useful to consider bands of states, and the variations in spatial extent
and nodal properties with energy.
Along with these new conceptional worries comes a whole new set of
reasons to be doing such experiments in the first place. Since these
clusters act in many ways like a bulk surface, they should serve as
excellent "molecular models" of such surfaces 6 . In fact, to many
scientists currently involved in this cluster beam field, this molecular
model aspect is the principal appeal.
As is the case with all simple models, these cluster models should
not be expected to be perfect replicas of the real thing. They will be
useful as models only to the extent that some of the essential new
physics and chemistry that we need to learn about real surfaces will also
hold sway in the small cluster. If so, it may be that many of the new
insights will be gained first on these small clusters. The hope of
course is that many of the powerful techniques we normally associate with
small molecule science can be extended to these cluster models.
Particularly intriguing is the notion that it may soon be possible to
calculate the physical and chemical properties of these clusters by high
level quantum chemical techniques, and to test the effectiveness of the
approximations used by direct comparison with experimental measures of
exactly the same clusters in the laboratory.
251
especially when this chemistry occurs on the surface of what is primarily
a covalently bonded material. So a natural early topic of concentrated
research has been carbon, silicon, and gallium arsenide. This short
paper summarizes a few of the recent developments along these lines,
emphasizing one of the most promising new techniques: ion cyclotron
resonance (ICR) of the cluster surface chemistry while the particles are
trapped at high vacuum in the field of a superconducting magnet.
252
etched the surface in a spiral pattern, smoothly removing material from
the surface at a rate of roughly 1000 mono layers per pass. In such a
nozzle the hot atoms and small molecules originally produced in the
laser-driven plasma are entrained in a fast flow of helium carrier gas at
nearly 1 atm pressure, and the resultant mixture is allowed to freely
expand into a vacuum, producing an intense supersonic jet which is then
skimmed to form a well-collimated beam. Using a second laser (ArF
excimer at 1930 angstroms) to irradiate the clusters just as they begin
the supersonic expansion it is possible to generate a dense plasma of
positively and negatively charged clusters which then receive the full
cooling of the supersonic expansion.
Similar techniques have been in use in our group for a few years
now in a wide range of experiments with cluster ion beams in the size
range from 2 to several hundred atoms per cluster. Although the
techniques were originally developed the for
spectral and
10
photofragmentation study of the positive cluster ions , the negatively
charged species "l:lave turned out to be particularly useful since they
permit the detailed study of the UPS patterns for these clusters as a
function of cluster size 11 . Here, however, as shown in Figure 1 the
clusters were directed by means of a pulsed extraction field and a pair
of einzel lenses down along the central axis of a superconducting magnet
operating in the persistence mode at a peak field of 6 Tesla.
253
As shown in the figure, the forward momentum of the cluster ion is
slowed in two successive deceleration steps as it approaches the ICR
trap. This trap is composed of a 15 em long, 4.8 em diameter cylinder
divided lengthwise into 4 sectors. One opposing set of two of these
sectors were used to provide RF excitation in order to coherently pump
the cyclotron motion of the clusters. The other set of two sectors were
connected to a very sensitive differential amplifier so that the weak
image currents generated by the circling cluster ions could be detected,
digitized, and the resultant time-dependent waveform submitted to a fast
Fourier transform. Drift of the cluster ions out of the trap by motion
along the magnetic field lines was prevented by a small repelling
potential on two electrodes on either end of the cylinder.
Due to the crude initial pulsed extraction of the cluster ions from
the supersonic beam, a considerable energy spread is present in the
cluster ion packet as it enters the magnet. As a result it is
inefficient to decelerate the cluster ions to less than 5-10 eV as they
approach the ICR trap. This and the fact that often 10-100 pulses of the
supersonic cluster beam apparatus are necessary to "fill" the ICR cell
forces us to use a thermalizing gas in the cell during injection, and to
pulse the entrance electrode to the cell down momentarily as each cluster
packet arrives. The experiments discussed below used neon as the
thermalizing gas held at 1 x 10- 5 torr in the ICR trap during the
injection of (typically) 100 cluster ion pulses at 10 pulses per second.
After an additional 5 seconds to insure the clusters were effectively
thermalized, this neon gas was allowed to pump away prior to the RF
excitation pulse and subsequent detection of the cyclotron resonance.
Once injected and thermalized, cluster ions may be stored in this trap at
high vacuum for many minutes.
254
expected in accord with the 1.1% normal terrestrial isotopic abundance of
13c.
255
47
50 49
45
!)4 53 52 51 46
44
....... ~
ollro.,j ~
1500
~ ~ lw.ll ~
1400
~
_I
1lOO
~--
MASS (AMU)
A "SWIFT" TECHNIQUE
256
Ge 11
Get2
mass confusion. For example, the top panel in Figure 4 shows the FT-ICR
mass spectrum of germanium clusters in the 11-12 atom mass range.
The broad spread of multiple peaks seen here is due to the 5 major
isotopes of germanium in the 70-76 amu mass range. In order to eject all
but a single isotopic variant of these clusters, one would ideally like
to arrange the RF excitation such that a uniform RF power was given to
257
all masses shown in the blocked off area to the top panel of this figure.
The problem, of course, is that the desired single mass cluster is
exposed to the RF as well and one has to worry that off-resonant
excitation of this cluster will leave its cyclotron motion highly
excited.
In the bottom panel of figure 5 one can see the effect on this
cluster distribution of a 5 second exposure to 4 x 10- 7 torr NH3 at near
room temperature. A typical silicon cluster ion will experience roughly
70 collisions with NH3 during this exposure. Note that some silicon
clusters such as Si44+ and Si46 + have chemisorbed NH3 so efficiently
under these conditions that very little of the bare cluster remains.
258
51 48
52 49 45
47 46 44
.J....
' J . .I. j
.......-~o<lo..o.
_{ I
1$00 1400 1300
MASS (AMU)
259
by our group 16 . Now that more extensive data has been accumulated it is
clear that these variations are quite real and reproducible.
100
\ I
/1
80
J
60
...,.,.,
"0
~~
c..,
0
"" 40
20
0 10 20 30 40 so 60 70
# Si atoms
260
H2 and N2 on small transition metal clusters. But it is quite surprising
to find it here. First, sharp variations in reactivity are being seen
here for very large clusters. In all previous cases the measured
reactivity variations largely died out before the cluster size reached 30
atoms, yet here even at 64 atoms special behavior is evident. Second, it
is far easier to produce amorphous silicon by rapid quenching than it is
for typical metals. It would not have been surprising to have found
these silicon clusters to be essentially amorphous, with each cluster
size having many geometrical isomers represented in the ICR trap, each
with a distinct reaction chemistry. Yet that clearly cannot be
happening. The inertness of clusters like Si45+ indicates that there is
a special structure for this cluster size and that somehow the 45-atom
clusters have been able to anneal sufficiently to settle into this
structure.
261
conceivable that quite detailed and predictive calculations will be
available over the next few years for silicon clusters throughout the 2-
100 atom size range. As yet there is no compelling explanation for the
new reactivity data based on such fundamental calculations, although some
interesting first steps along this road have been published recently2 1 • 22
examining in some detail a structural suggestion originally put forward
by J. C. Phillips2 3 . A strikingly beautiful model with full tetrahedral
symmetry has be suggested by Kaxiras 24 as a candidate for the Si45+
cluster. Interestingly the surface of this structure is characterized by
the same restructured bonding motifs as those found on the (111) face of
bulk silicon.
CONCLUSION
262
middle, providing a number of reasonable locations for dissociative
addition of H2.
ACKNOWLEDGEMENT
REFERENCES
263
6. R. E. Smalley in Comparison of Ab Initio Quantum Chemistry with
Experiment: State of the Art edited by R. J. Bartlett (Reidel, New
York, 1985) pp 53-65.
20. D. Tomanek and M.A. Schluter, Phys. Rev. B 36, 1208 (1987).
21. D. A. Jelski, Z. C. Wu, and T. F. George, Chern. Phys. Lett. 150, 447
(1988).
264
THE ELECTRONIC STRUCI'URE OF METAL DIMERS AND METAL CLUSTERS: 1HE
Michael B. Hall
Department of Chemistry
Texas A&M University
College Station, Texas 77843-3255
INTRODUCTION
The 18-electron or effective atomic number rule (EAN)l has been extremely valuable to
organometallic chemists in directing their thinking toward new compounds and "explanations"
of the structure and bonding in known compounds. In metal dimers one often postulates the
degree of metal-to-metal bonding by enforcing the 18-electron rule. As chemists made larger
and larger clusters, there came the realization that the 18-electron rule was difficult, if not
impossible to enforce. This problem lead to the formulation of skeletal electron pair counting
(SEPC)2 and cluster valence electron counting (CVEC). 3 A strong connection between the
bonding in main group clusters such as the boranes and that in organometallic or low oxidation
state transition metal clusters can be made through the isolobal principle.4,5 Again SEPC and
the isolobal principle have lead to suggestions for new compounds and new synthetics routes, as
well as "explanations" for the structure of existing compounds.
Without any doubt these qualitative electron and orbital counting procedures have
contributed substantially, and will continue to contribute, to the development of the chemistry
of organometallic compounds, especially dimers and dusters. However, in spite of their
usefulness in organizing our thinking and providing useful relationships, it is not obvious that
any of these electron counting principles leads to an accurate picture of the bonding. They
clearly do a good job of rationalizing and predicting the molecular structure, but how well do
they do in rationalizing and predicting the electronic structure at a deeper level?
We will examine this question by discussing in some detail the bonding of the clusters
Os3(COl9X2 (X=S, Sel and the dimers (CpM}2(J.1-E0)2 (M=Co, Rh, Ir and E=C, N). Previously,
we have published the photoelectron spectra and approximate molecular orbital calculations
of these compounds.6,7
These clusters may be viewed as square based pyramids with the unique Os forming the
apex and the X's and other two Os's forming the base, 1.
One can think of 1 as an electron precise duster which satisfies the EAN rule at all atoms (Se-
at the main group atoms and 18e- at the metal atoms). To do so one can treat the X's as x2-, the
apical Os as QsO (dB), and the basal Os's as Os2+(d6). The principle resonance structure would
then appear as 2
/f\
as2• -..- as~as2•
'\J/ ~-
2
In this structure the bonds have been drawn as dative bonds (arrows) to coincide with the
original oxidation state choice. However, we do not wish to imply that these bonds are
necessarily dative or that the initial oxidation state is an accurate reflection of the final
charges. In our view an equally accurate (or equally inaccurate) representation of the electronic
structure would have all covalent bonds, 3. One should view 2 and 3 as resonance structures; the
"true" total wavefunction being a mixture of these and other related resonance forms.
266
The key feature in all these forms is the existence of eight 2-center, 2-electron (2c/2e)
bonds. To make these bonds an Os center (the apical Os in 2 and 3) has used one of its 5d t2g-like
electron pairs (one of the three pairs stabilized by backbonding to the carbonyls) in the Os-Os
bonding. One can also describe this view in terms of hybrid orbitals, again our description
should not be taken too literally as any hybrid orbital description is a gross approximation.
Here x2- can be viewed as "sp3" with three of the hybrids used in X-Os bonds and one
remaining as a lone pair. The basal Os's are "d2sp3" with three hybrids used for bonding to
CO, and two used for Os-X bonds, and one used for each Os-Os bond. The seven coordinate
apical Os is "d3sp3" with three hybrids for CO bonding, two for Os-X bonds, and two for Os-Os
bonds. On all Os's the unused Sd orbitals are occupied; thus, the basal Os's have d6 cores and
the apical Os has a d4 core. Regardless of oxidation state choice, there are exactly the correct
number of hybrids and electrons remaining to form eight 2c/2e bonds.
An alternative view of this cluster is provided by skeletal electron pair counting
(SEPC). Here, we use the isolobal principle and make an analogy to the electron deficient
boron hybrides. For our Os3(CO)gX2 clusters the Os(C0)3 fragments are isolobal to BH
fragments, where each Os(C0)3 contribute 3 orbitals and 2 electrons to cluster bonding. The
remaining 6 d electrons on each Os are not involved in the cluster bonding but are on! y involved
in bonding to the carbonyls. Each X contributes 3 orbitals and 4 electrons. Thus, the Os3(CO)gX2
clusters can be thought of as 5 vertex polyhedra with 7 skeletal electron pairs. By analogy to
the boron hydrides, they are nido clusters (n+2 pairs) and are derived from an octahedron with
one vertex missing. The SEPC approach succeeds in predicting the observed geometry.
In small, lower symmetry, clusters one can often arrive at a localized bonding picture by
following the prescription of Lipscomb.B For the boron hydride analog, BsHg, the square base
of the pyramid is bonded by 3c/2e B-H-B bonds and the capping BH group is bonded to the four
basal borons by two 2c/2e bonds and one 3c/2e bond. A number of different arrangements are
possible. In our lower symmetry system one may choose from either 4 or 5.
4 5
Thus, one could have the 3c/2e bond holding the capping Os to the other Os (4) or to the X
groups (5). A third possibility also exists in which the 3c/2e bond is a central rather than an
open bond (6). There would be four equivalent resonance forms like 6.
The principle differences between the 18-electron description and the SEPC description
is the number of electron pairs involved in cluster bonding. In enforcing the lBe- rule we must
involve one of the t2g-like pairs, which SEPC and the isolobal analogy suggest are involved
267
principally in Os-CO bonding. Now that we have set the stage, we will discuss the results of
our molecular orbital (MO) calculations in an attempt to resolve the questions: to what extent
the t2g-like pairs are involved in cluster bonding and, which, if any, of the 3c/2e models is
closest to the MO results.
Figure 1 shows the MO diagram for both Os3(C0)9S2 and Os3C0)9Se2. The relative
energies of the 5 3p and Se 4p are shown on the left, while in the high lying orbitals of the
Os(C0)3 fragment are shown on the right. Hoffmann and coworkers have shown how carbonyl
fragments such as Os(C0)3 retain much of their octahedral parentage.9 Thus, the orbitals in
Figure 1 are labelled "t2g" for the three Sd orbitals stabilized by the carbonyls 1t* (7t acceptor
orbital), "eg" for the two Sd orbitals destabilized by the carbonyls lone pair (cr donor orbital),
and "at" for the unused 6s/6p hybrid. Together the "eg" and "at'' orbitals form the three
unused hybrids on the Os(C0)3 fragment while the "t2g" orbitals are mainly spectator orbitals
according to the isolobal analogy with BH and SEPC. There are eight electrons to distribute
among these orbitals and the Os(C0)3 fragment has six in the "t2g" and two in the "eg"· Thus,
according to the SEPC formalism Os(C0)3 has two electrons and three orbitals to contribute to
cluster bonding. The 5/Se p orbitals contribute four electrons and three orbitals to cluster
bonding.
As Figure 1 shows there is strong mixing of the 5 orbitals with the "t2g", "eg", and "at"
orbitals. This mixing will only be important, i.e. produce a net bond, if it results in an occupied
bonding orbital and an unoccupied antibonding orbital. At low energy we would expect to find
six 5-0s bonding orbitals. However, Figure 1 only shows five low-lying 5-0s orbitals; the
sixth, 6a', is found in the midst of the "t2g" orbitals. It is easily identified because of its large
energy shift when 5 is replaced by Se. This orbital is plotted in Figure 2, where one can see
that it has strong X-X antibonding character which is responsible for its high energy.
268
_.-· ···... ···...
....· ..··
.....····· ··.._ ··. ··..
·····-·-· ...
The highest occupied molecular orbital (HOMO) is the 7a". A plot of this orbital,
Figure 3, shows it to be a 3c/2e bond between the apical Os and the basal Os's.6 The remaining
nine molecular orbitals are derived from the "t2g" pairs. The MO calculations indicate that
the dominant bonding is that shown in 4. However, the Os-Os bonds are shorter than one would
expect for a bond order of 1/2. A closer examination of the overlap population reveals net
bonding between the "a1" and "t2g" orbitals of the Os(C0)3 fragments. This bonding occurs
mainly in the 9a' orbital, the second highest occupied orbital, and in the 1a' orbital, the
269
lowest energy orbital shown in Figure 1. Such bonding in the ai system could help account for
the nearly typical Os-Os bond lengths that are observed. Interestingly, the high-lying 6a',
which is primarily involved with the apical Os, does not provide as much Os-X bonding as any
of the other five Os-X bonding orbitals. Thus, the Os-X bonds to the apical Os may be
somewhat weaker. This high-lying X-Os orbital also plays an important role in the chemistry
of these clusters as it seems to be involved in directing the attack of electrophiles such as PtL2
to the X-Os edge.B For these clusters the reported photoelectron spectra are consistent with our
orbital description, but with many closely spaced levels a detailed assignment was not
possible.6
In conclusion, the principle bonding features are well represented by the SEPC model
with the major form of possible resonance structures being 4. However, the EAN structures, 2
and 3, and the alternative SEPC structures, 5 and 6, appear also to contribute in some smaller
way.
In this section we will concentrate on the bonding of carbonyl and nitrosyl derivatives
of Group 9 (Co, Rh, Ir). In the solid state, both <11 5 -CsMes)2Co2(~-C0>2 and (1l 5-CsHs>2Co2(~
N0)2 appear to have planar M2(E0)2 cores as shown in 7
In solution, the Co2(N0)2 derivative apparently has a puckered geometry. In our description
of the bonding we will concentrate on the planar form. Some previous theoretical calculations
suggested the possibility of a triplet ground state,11 but the singlet nature of the ground state is
consistent with the PE spectrum7 and was recently confirmed by magnetic measurements.12
If one enforces the 18-electron rule the Co2(C0)2 dimer would have a double Co = Co
bond while the Co2(N0)2 dimer would have a single Co - Co bond. On the other hand if we
view this dimer from the isolobal analogy: CO and NO are 2 and 3 electron donors,
respectively (isolobal to BH and BH-, respectively), and the CoCp fragment supplies 2
electrons (isolobal to BH). The isolobal principle is limited here because B4H4 (isolobal to
Cp2C02(C0)2) would be tetrahedral and B4H42- (isolobal to Cp2Cp2(N0)2) would be planar,
whereas Cp2C02(C0)2 is planar and Cp2C02(N0)2 may have a folded structure. However,
the isolobal electron counts do provide us with the number of electrons available for the
principal bonds which confer stability to these molecules. Thus, the Cp2Co2(C0)2 dimer has
four pairs available for a cluster bonding and Cp2Co2CN0>2 has five pairs available. In
constrast, the 18-electron rule assigns six pairs to "cluster" bonds in Cp2Co2(C0)2,: four pairs
for C - Co bridge bonds and two pairs for Co = Co bonds; and five pairs to duster bonds in
CP2Co2(N0)2: four pairs for N- Co bonds and one pair for the Co -Co bond. For Cp2Co2CC0)2 ,
the two approaches differ widely in the number of pairs that they assign to cluster bonding.
Even for Cp2C02(N0)2,. where the number of pairs is identical, the placement of the pairs is
very different. Planar B4H42- or C4H42+ would have two electrons in the 1t system and would
270
be aromatic, but the 18-clectron rule would suggest a Co -Co cr bond in addition to the four in-
plane Co - N bonds, i.e. all five pairs in the molecular plane.
The molecular orbital (MO) energy level diagram, which is for the Rh dimers7, is
shown in Figure 4. The left side shows the upper energy levels of the CpRh fragment; the
lowest energy orbitals shown, let, are the degenerate p donor orbitals of the Cp- ring. Above
those, the e2 and lat are the remains of the "t2g" like 4d orbitals. At even higher energy are
the 2e1 and 2at which are destabilized by the donation from the Cp- ring. The 2et are mainly
4dxz and 4dyz, while the 2at is primarily an "sp" hybrid on Rh. These CpRh fragment
orbitals interact weakly with each other and strongly with the orbitals of the bridging
ligands which are shown on the right of Figure 2. Since NO is a poorer cr donor and better 1t
acceptor than CO, both its 5cr donor orbital and its 27t acceptor orbital lie lower in energy than
the corresponding orbitals of CO.
The center has two columns of levels, one for the CO dimcr, the other for the NO dimer.
At low energy are the out-of-phase and in-phase combinations of the 5cr donor orbitals, lau and
lag respectively. The lau is stabilized by the metal fragments 2e1 (dyz) while the lag is
stabilized by the 2at orbitals. The next 4 orbitals are the result of the interaction of the Cp-Rh
"1t" bonding orbitals. These are mainly Cp in character and contribute little to the net "cluster"
bonding. The next 6 orbitals are the bonding and antibonding combinations of the "t2g" orbitals.
The next two orbitals, 3bg and 4bu, are the two linear combinations of the 2e1 which are
....m..
co
CpRh EO
271
strongly stabilized by interactions with the 21t orbitals. These two orbitals are occupied in
both dimers. The Sag orbital, which is the antibonding combination of the dxz's, is occupied
only in the nitrosyl dimer.
The MO calculations suggest that the primary "cluster" bonding in Cp2Rh2(C0)2
involves the lau, lag, 3bg and 4bu. These orbitals contain the four pairs identified by the
isolobal analogy as being involved in cluster bonding. One might expect to find all four pairs
involved in Co-C a bonds, but the 4bu is an out-of-plane orbital. Figure S displays orbital
drawings for these four orbitals and the Sag which is occupied in the NO dimer. The lag is the
totally bonding combination of the CO a donor and the "sp" hybrid on the metal. The lau and
3bg are antibonding and bonding combinations of the metal dyz orbitals. Figure S shows them
as linear combinations which emphasize that the net bonding is equivalent two 3c/2e C-Rh-C
bonds. The 4bu is a totally bonding "1t" orbital similar to the lowest energy 1t orbital of C4H4.
This orbital contributes to the planarity of the carbonyl dimer. Occupation of the Sag breaks
any direct metal-metal bonding in the 4bu and would reduce the barrier to "folding" the dimer.
Thus, Cp2Co2(N0)2 appears planar in the solid but folded in solution. For the NO dimer, one
could form linear combinations of the 4bu and Sag, the new orbitals would be 3c/2e N-Rh-N "1t"
bonds similar to their a counterparts, lau ± 3bg.
Although the MO calculations suggest that the primary "cluster" bonding is confined to
these orbitals, they also indicate some important secondary bonding effects. The MO diagram,
Figure 4, shows the interaction of the la1 orbitals on each dimer to form the 3ag, M-M bonding,
and the 3bu, M-M a antibonding. The isolobal analogy suggests that the la1 orbitals are
"spectator" orbitals not involved in cluster bonding. This result would be the situation if the
3bu antibond cancelled the 3ag bond, but it doesn't. The 2a1 fragment orbital stabilizes the 3bu
converting it, at least partly, into a non-bonding orbital. Thus, there is net a M-M bonding in
this "t2g" mainfold.
A second result, also unexpected from the isolobal analogy, is that the CO and NO 1t•
orbitals contribute strongly to the mainly metal 2bu. This contribution is clearly evident in
substantial stabilization of the 2bu which occurs when the CO's are replaced by NO's. This
orbital corresponds to the "missing" fourth in-plane M-EO bonding orbital. The four M-Eb
bonds can be made from the lau, lag, 2bu, 3bg· The 3ag contributes a M-M a bond and the 4bu
contributes a M-M 1t bond. Now we are back to 6 electron pairs involved in cluster bonding as
suggested by the 18-electron rule. In the NO dimer the bonding 4bu and anti bonding Sag cancel
5aQ ~ ~
a
4b, ~ ~
0
cP ~~
~~
Iau+ 3b;
cP Ia,- 3bo
lao ®X)
~ (:)®
~
Figure S. Valence description of interfragment bonding. The
view is of the yz plane (molecular plane). Only orbitals of the
metals and proximal atoms of the bridging ligands are shown.
(Reprinted with permission from reference 7. Copyright 1988
American Chemical Society.)
272
each other forming two lone-pairs on the metal fragments. Here, the remaining 5 pairs form 4
M-NO bonds and one M-M bond.
CONCLUSIONS
We have come "full circle" on the issue of the isolobal analogy and skeletal electron
pair counting vs. the 18-electron rule. One can generalizes our observation by stating that
careful application of the isolobal principle and skeletal electron pair counting will lead to
the enumeration of the principle bonding in the cluster, but the pairs which are still missing
from those required by the 18-electron rule may provide important secondary bonding
interactions. Finally, our unpublished ab initio calculations on these dimers shows that even
the detailed analysis given above is an oversimplification of the "true" electronic structure.
REFERENCES
1. N.V. Sidgwick, "The Electronic Theory of Valency", Cornell Univ. Press, Ithaca, N.Y.
(1927).
2. K. Wade, Adv. Inorg. Radiochem., 18:1 (1976); D.M.P. Minges, Acct. Chern. Res., 17:311
(1984).
3. J.W. Lauher, J. Am. Chern. Soc., 100:5305 (1978); J.W. Lauher, J. Am. Chern. Soc.,
101:2604 (1979).
4. M. Elian and R. Hoffmann, Inorg. Chern., 14:365 (1975).
5. R. Hoffmann, Angew. Chern. Int. Ed. Engl., 21:711 (1982).
6. G.L. Griewe and M.B. Hall, Inorg. Chern., 27:2250 (1988).
7. G.L. Griewe and M.B. Hall, Organomet., 7:1923 (1988).
8. W.N. Lipscomb, "Boron Hydrides," W.A. Benjamin, New York (1963); W.N. Lipscomb,
Science, 196:1047 (1977).
9. M. Elian, M.M.-L. Chen, D.M.P. Mingos, and R. Hoffmann, J. Am. Chern. Soc., 98:7240
(1976).
10. R.D. Adams, T.S.A. Hor, and I.T. Horvath, Inorg. Chern., 23:4733 (1984).
11. J. Demuynck, P. Mougenot, and M. Benard, J. Am. Chern. Soc., 109:2265 (1987).
12. D.J. Berg and R.A. Anderson, J. Am. Chern. Soc., 110:4849 (1988).
273
EXPERIMENTAL MEASURES OF METAL-METAL SIGMA, PI, AND DELTA BONDING
INTRODUCTION
Since the discovery, in 1964 by Cotton and co-workers, that the anion
[Re 2Cl 8 ] 2- contains a quadruple metal-metal bond, 1 the field of multiple
metal-metal bonds has grown dramatically. 2 From the very beginning,
CI CI 2-
1/C I 17 I
I (I
/Re Re
c c
CI CI
interest has focused on the nature of the bonding between the metals. In
a seminal paper, Cotton proposed that the quadruple bond is composed of
one u, two w and one 6-component arising from overlap of the d orbitals on
the metal atoms, as shown in Figure 1. 10 Molecular Orbital (MO) calcula-
tions of varying degrees of sophistication have subsequently shown that
these symmetry interactions are important to the metal-metal bonding in a
-M =£:~ N .
·~:
a ~
Mz M IE M
Type II
There are several advantages to studying the ligated metal dimers over the
bare metal species. Although the presence of ligands reduces the symmetry
of the dimers from D~, the local (virtual) symmetry is high enough to
enable the assignment of a, w and 6 labels to the components of the multi-
ple bond. For theoretical calculations, the larger number of atoms is
compensated by the ligand-induced destabilization of a number of the
virtual levels so that there is less configuration interaction. 30 • 4 These
ligated metal dimer complexes are stable and can, therefore, be studied
with a variety of experimental techniques, including high resolution
spectroscopy in the gas phase. 5 In addition, ligated metal dimers offer
the opportunity to study the effects of chemical perturbations on the
metal-metal interaction.
276
photon of energy hu:
hu (2)
(3)
Figure 2 shows the gas phase ultraviolet PE spectra obtained for the
quadruple bonded acetate bridged dimers M2 (0 2 CCH 3 ) 4 of chromium, molybdenum
and tungsten. The ionizations above 10 eV are largely associated with the
acetate ligands and change very little between the spectra of the differ-
ent metal complexes. The ionizations below 10 eV are associated with the
metal valence d levels and show striking differences from Cr 2 to Mo 2 to
W2 . The first ionization of the chromium dimer is extremely broad. A
close-up examination of this band (Figure 3) shows that it is comprised of
three peaks in an approximate 1:2:1 intensity pattern. Some vibrational
structure is evident on the top of the band, but the three broad peaks
overlap too much to obtain much specific information. The low valence
region of the molybdenum complex has two clearly identified peaks. The
first has been attributed to ionization from the 5 bonding (2b 28 ) MO of
the dimer, 11 as predicted by MO calculations on the quadruple bonded
systems (see Figure 1). 10 • 3 The second is assigned to the~ ionization
state. One problem in this system has been locating the ionization
associated with the a state. The spectrum of the tungsten complex shows
three separated ionizations in the low valence region. The first two have
again been assigned to the 5 and~ ionizations, and the third has been
assigned to the a ionization (vide infra). The sharpness of the third
ionization and its close proximity to the ~ ionization are unusual for
ionization of a a bonding electron.
277
Ionization Energy (eV)
17 13 9
9 8
'
'
'
'
' \
' \
'
'
...... _:::~j-:·.:·:····················-.-·· ... '.\"'
~~ ~~
..... ~,· '......... . ........... . .," ...
278
Of the many interesting questions which have arisen out of the PES
study of quadruple bonded metal dimers, we will focus on understanding how
electron occupation of the u, w, and 6 bonds influences the metal-metal
bond length and strength. Knowledge of these influences is crucial if we
are to advance our understanding of the fundamentals of bonding in these
systems. As a first observation, the w ionization in each case shows a
broad vibrational progression indicating substantial bond distance changes
with w ionization. Thus the w bonds contribute significantly to the over-
all bond strength. The nature of the 6 and u bonds is much less apparent.
The following sections will focus on the nature of the 6 and u bonds in
multiple bonded metal-metal complexes.
THE 6 IONIZATION
The "Mo-Mo" vibrational frequency and calculated bond length for the
2 B28cation state are given in Table I, together with the experimentally
determined bond length of the neutral molecule in the ground ( 1A18 )
state. 15 For comparison, the Mo-Mo bond length resulting from exciting an
electron from the 6 orbital to the 6* orbital (thereby generating the 1A2u
279
Table I. Compilation of stretching frequencies and bond lengths for
[Mo 2(0 2CCH 3 ) 4 ]n+ (n - 0, 1).
Table adapted from reference llc. The calculated Mo-Mo bond lengths are
obtained by adding the calculated increments to the experimentally deter-
mined (gas phase) bond length from reference 15b. a Reference 14a. b
Reference 15a. c Reference 15b. d Reference 16. • References 14a and
17. f Reference llc.
75
~ 73
~ 71
~
w
z 6.9
0 2 ~ STATE
~
z
N
NEUTRAL
I""'T'"'1
-0.2 0 0 2
280
Table II. Mo-Mo bond lengths in a number of dimers; the effect of 6
ionization.
From the Table, it is apparent that the bond lengthening upon removal
of successive electrons from the 6 orbital (e.g. +0.056 A on going from
[Mo 2(S0 4) 4] 4- to the trianion 18 · 20 and +0.057 A on going from the trianion
to the closely related [Mo 2 (HP0 4) 4] 2- ion) 21 is much smaller than that
calculated from the PE spectrum of molybdenum acetate (0.18 A). 11 c It
seems, therefore, that the Franck-Condon analysis has overestimated the
magnitude of bond lengthening accompanying the removal of an electron from
the 6 MO. In the following sections we shall discuss possible reasons for
these inconsistencies between spectroscopic and crystallographic measures
of the lengthening of the Mo-Mo bond upon ionization.
1. The Hot Band Argument: The possibility that the presence of hot
bands (i.e. transitions from excited vibrational levels of the ground
electronic state) 22 are complicating the vibrational fine structure
observed for molybdenum acetate was discussed in the original presentation
of the analysis. 11 c It was concluded that the presence of hot bands does
not affect the spacing between the apparent adiabatic (v-0 ~ v'-0) and
vertical (assigned as 0~7) ionizations. This assertion was later ques-
tioned by Miskowski and Brinza. In a subsequent analysis they included
contributions from as high as the fourth vibrationally excited level (v-4)
and eventually concluded that a better upper limit on the bond lengthening
upon 6-ionization would be 0.11 A, 23 i.e. approximately the same as that
calculated for the 6 ~ 6* transition. 148 · 17 However, a simple Boltzmann
distribution calculation indicates that less than 1% of the molecules have
v ~ 4 at the temperature at which the PES data was collected (~ 200"C -
473 K). Figure 6 shows the relationship between a vibrational profile at
0 K and the effects of hot bands (v - 400 cm- 1 ) at 473 K. The figure
reveals that the addition of hot bands has little effect on the
determination of the number of vibrational quanta between the adiabatic
and vertical ionizations, even at 473 K. The Miskowski analysis also
assumes that the vibrational frequencies are the same in the neutral
molecule and the positive ion, and this assumption will over-estimate the
hot band effect. It is also not certain that the adiabatic ionization is
observed, and if it occurs at lower IP than assumed then the apparent bond
distance change is even greater. In any event, even with over-estimation
of the effects of hot bands, the analysis still results in bond distance
changes that are greater than those found in the structural studies.
281
0 K 473 K
to 0.13 A for the 2 B28 ion state. 11c In a similar analysis, Crosby et al.
obtained a smaller value (0 . 08 A) for the bond lengthening in the excited
1A2u state (o-+o*) of the neutral molecule . 24
Thus, the inclusion of hot bands and changing the effective mass of
the vibrating units does not reduce the calculated increase in metal-metal
bond length sufficiently to be consistent with the solid state structural
data that we have for the oxidized multiply bonded dimers (Table II).
282
Table III. Comparison of M-M and M-X bond length changes on oxidation of
a number of multiply bonded dimers.
a M-X bond lengths are averaged. The e.s.d.'s given are averaged over all
the M-X bonds also. b Major isomer (69.1%). c Minor isomer (30.9%). d X
= Cl for all theRe complexes.
of the second 5* electron has an even smaller effect on the Re-Re bond (it
increases by only 0.003 A, but the high e.s.d. for the dication makes this
number unreliable), with the Re-Cl bond shrinking by 0.039 A.
The structural details reveal that the removal of a 5 (or 5*) elec-
tron generally results in a shortening of the M-X bond which is equal to
or greater than the change in the M-M bond length. While part of the
change in the M-0 bonds can undoubtedly be attributed to a reduction in
the size of the metal ion upon oxidation (i.e. a tightening of the valence
orbitals), the major contribution probably comes from a reduction in the
M(d~)-X(p~) repulsion. This repulsion occurs due to the antibonding M-X
interaction which is present in both the 5 and the 5* MO's when X is a
1r-donor ligand. 3
M-X ~ Interactions
P R3 X
I /X I /p R3
ll
M===M
R3P
7
I
P R3 X
283
This eclipsed conformation is expected for the Mo and W molecules,
which possess a 6 bond, but for the Re complexes the occupation of 6*
should favor the staggered conformation, as observed for triple bonded
[Os 2X8 ] 2 " (X- Cl, Br), which has D4d symmetry. 32 The observed conformation
of the rhenium phosphine halides is favored because eclipsing the phos-
phines with halides keeps the phosphines on the two ends of the molecule
as far apart as possible and thereby minimizes steric repulsions between
them. 33
Figure 7.
6 6.44
w 7. 71
M-P(a) 8.40
6 6.53
w 7.73
M-P(a) 8.33
284
Table V. The effect of changing the halide ligands on the ionization
energies of the 6 and 6* MO's of Re 2X4 (PMe 3 ) 4 .
Re 2 Gl 4 (PMe 3 ) 4 6* 5.66
6 6.51
Re 2 Br 4 (PMe 3 ) 4 6* 5.72
6 6.56
Re 2 I 4 (PMe 3 ) 4 6* 5.80
6 6.56
One would expect from a charge potential point of view that sub-
stituting Gl by the less electronegative Br should lead to a reduced
positive charge on the metal and therefore those MO's with an appreciable
amount of metal character should be less tightly bound, as is observed for
the M-P(u) ionization. The fact that the ~ and 6 ionizations shift
instead to higher ionization energies indicates that the ~ donation effect
is dominating the charge potential or electronegativity effect. Thus, the
poorer ~ donor ability of Br, compared to Gl results in less X(p~)~M(d~)
destabilization of the metal-metal ~ and 6 bonding orbitals and conse-
quently higher ionization energies. 340
285
·o~d·
11(e)
For both molybdenum and rhenium, the change in the ionization ener-
gies of the 6 (and 6*) bands are greater than for the 11 band. This can be
understood in terms of the 6 and 6* MO's possessing more ligand character 3
and, therefore being more susceptible to a change in the 11-donor ability
of the halide. As shown in Figure 10, the 6 and 6* MO's have contribu-
tions from all four 11-donor (X) ligands, while each component of the 11 set
has contributions from only two X ligands.
286
lonizalion Energy (eV)
THE o IONIZATION
Since three peaks are observed for tungsten and only two for molyb-
denum, it was thought that the second peak in the molybdenum spectrum may
in fact contain two overlapping bands. 36 To test this hypothesis, the PE
spectrum of MoW(0 2 CC(CH 3 ) 3 ) 4 was taken and compared with those of the Mo 2
and W2 analogs. 36 The results are shown in Figure 12. The Figure clearly
shows that on going from the W2 dimer to MoW the sharp ionization band
starts to merge with the w band and that for the Mo 2 dimer only one band
is observed. By extrapolation, we can conclude that the distinct o and w
ionizations observed in W2 are completely coalesced in Mo 2 •
287
Ionization Energy (eV)
8 7 6 5
it was hoped that the a and ~ bands would separate due to their presumed
different response to inductive effects. Unfortunately, no such separa-
tion was ever observed. 36
One piece of supporting evidence for the assertion that the a ioniza-
tion lies under the ~ ( 2 Eu) band in the molybdenum case is the fine
structure in the " ionization band. The " ionization band shows vibra-
tional fine structure on the high binding edge of the band but not on the
low binding edge as shown in Figure 13. 36 There is also a slight bump on
this low binding energy edge which suggests the presence of the a band.
....Ul
c:
;:)
0
u
288
Triple-bonded metal-alkoxides. At this point, in order to understand
the position of the a ionization in molybdenum acetate and its analogs, it
is informative to consider the PES of a closely related family of dimers.
The M2X6 dimers of molybdenum and tungsten (X - OR, NR2 , R) adopt the
staggered (D 3 d) conformation shown in III. These molecules possess a
triple bond (a 2w4 ). The PE spectrum of Mo 2 (0CH 2CMe 3 ) 6 enabled Green et al.
III
Mo=Mo
WaW
80 7.0 6.0
Figure 14. PE spectra of M2 (0CMe 3 ) 6 .
289
Table VI. Peak positions, widths and relative areas of the metal ioniza-
tions of some M2 (0R) 6 dimers.
In theory, gas phase PES of molecules of the type Mo 2 (0 2 CR) 4L2 , 2 with
axially bound 2-electron donor ligands (e.g. H2 0, PR' 3 ) would be of great
interest, since the donor orbitals will overlap with the metal a-bonding
MO and may destabilize it sufficiently for it to be distinguishable from
the 1r band. Unfortunately, it is found that these molecules lose their
axial ligands before subliming. In the solid state, however, it is known
that the quadruple bonded acetate (etc.) dimers associate, with long
intermolecular M--0 contacts at the vacant axial positions (vide infra),
which make these systems worth studying.
covalent Cr-0 bond. In the case of Mo 2 (0 2CCH 3 ) 4 the Mo-Mo bond lengthens
only from 2.079 A (gas phase) 15b to 2.093 A (solid state) 15 ". The dif-
ference is smaller for molybdenum than chromium, which is consistent with
the greater sensitivity of the Cr-Cr bond length to the presence and
nature of axially coordinated ligands. 28 • 44
290
Ul 12 8
C1 11' 6
291
10 8 7 e
Although it has not proven possible to take the gas phase PES of any
axially ligated M2 (02 CR) 4L2 molecules, where L is a 2-electron-donating
ligand, the spectra of the closely related molecules Mo 2 (0 2CR) 4 (np) 2 (np -
neopentyl; R- H, CH 3 , CF3 ) have been obtained. 46 Dimers of the type
M2 (0 2 CR) 4R' 2 (M- Mo, W) are M(III) (d3 ) complexes . They are related to
the M(II) (d4 ) M2 (0 2CR) 4 dimers by the formal oxidative addition of R' 2 . 46
2R
292
CJ+"II'
\ V 11'·ac
, ~
' \_A
~ I
~ 0 0
orbital, thereby reducing the overlap. 41 In addition, there are four
oxygen atoms on each metal interacting with the d.z in the case of the
carboxylates, and only three in the case of alkoxides. The calculations
also indicate that the M-M 1r bonding MO's are destabilized more in M2 (0R) 6
by M-0 u and 11'-type interactions.
293
Table VII. Metal-metal valence d and outer core s and p overlap
integrals.
<d.zl d.z>(a) <d.zlp.> <d.zl s> <d.,.. I d.,..> C,.. > <d,.y I d,.y> <6 > <p.lp.> <sis>
---------------·---·--- ---
Mo 1 0.116 0.117 0.111 0.155 0.046 0.028 0.005
W'z 0.138 0.134 0.116 0.181 0.051 0.028 0.004
These ligand interactions are not able to account for the narrow
bandwidth of the a ionization. One important factor that may account for
the narrow bandwidth, as well as the low IP, arises from the close contact
between the metal atoms in the triple and quadruple bonded complexes.
This close contact leads to substantial overlap between the valence nd 2 z
orbital on one atom and the outer core npz (and ns) orbital of its
neighbor. Jn,p,J~b.~0.4l The overlap integralsu between the valence d
orbitals and outer core orbitals are listed in Table VII.~ 1 From the
Table, it is apparent that the overlap between dzz and the outer core Pz
is roughly of the same magnitude as the d.z-d 2 z overlap, with the d 2 z-s
overlap being somewhat smaller. Also the d.z-d 2 2 (a) overlap is smaller
than the d,..-d,.. (11") overlap.
A slightly different explanation for the role of the outer core np.
orbitals is offered by the results of ab initio calculations 48 and Fenske-
Hall calculations.~ 1 These calculations indicate that the symmetric
combination of the outer core np. orbitals mixes with the valence da
orbital as shown in Figure 19. The result is the same as for the first
valence ionization of N1 , where the mixing of the pa combination with the
2sa combination leads to a valence non-bonding a orbital that ionizes at
lower energy than the P'~~" bonding orbital. Interestingly, the binding
energy of the N2 2sa combination is almost the same as the Mo 4pz level at
38 eV.
ACKNOWLEDGMENTS
294
7
8
a
eV
I 7r \
9 I \
I \
I \
~1
I \
\
I \
I \
4p-' 4p
REFERENCES
295
(b) Norman, J.G., Jr.; Kolari, H.J. J. Am. Chem. Soc. 1975, 97, 33.
(c) Norman, J.G.; Kolari, H.J. J. Chem. Soc., Chem. Commun. 1975,
649.
(d) Norman, J.G., Jr.; Kolari, H.J.; Gray, H.B.; Trogler, W.C.;
Inorg. Chem. 1977, 16, 987.
(e) Cotton, F.A.; Kalbacher, B.J. Inorg. Chem. 1977, 16, 2368.
(f) Block, T.F.; Fenske, R.F.; Lichtenberger, D.L.; Cotton, F.A.
J. Coord. Chem. 1978, 8, 109.
(g) Hillier, I.H.; Garner, C.D.; Mitcheson, G.R.; Guest, M.F. J.
Chem. Soc., Chem. Commun. 1978, 204.
(h) Guest, M.F.; Garner, C.D.; Hillier, I.H.; Walton, I.B. J. Chem.
Soc., Faraday Trans. 2 1978, 11, 2092.
(i) Benard, M. J. Am. Chem. Soc. 1978, 100, 2354.
(j) Bursten, B.E.; Cotton, F.A.; Hall, M.B. J. Am. Chem. Soc. 1980,
102, 6348.
(k) Atha, P.M.; Hillier, I.H.; Guest, M.F. Chem. Phys. Lett. 1980,
75' 84.
(1) Goodgame, M.M.; Goddard, W.A. J. Phys. Chem. 1981, 85, 215.
(m) Arriata-Perez, R.; Case, D.A. Inorg. Chem. 1984, 23, 3271.
(n) Ziegler, T. J. Am. Chem. Soc. 1985, 107, 4453.
(o) Hall, M.B. Polyhedron 1987, 6, 679.
(p) Ziegler, T.; Tschinke, V.; Becke, A. Polyhedron 1987, 6, 685.
(q) Bursten, B.E.; Clark, D.L. Polyhedron 1987, 6, 695.
(r) Davy, R.D.; Hall, M.B. J. Am. Chem. Soc. 1989, 111, 1268.
4. (a) Atha, P.M.; Hillier, I.H. Hol. Phys. 1982, 45, 285.
(b) Cotton, F.A.; Shim, I. J. Phys. Chem. 1985, 89, 952.
5. (a) Cr 2 : Michalopolous, D.L.; Geusic, M.E.; Hansen, S.G.; Powers,
D. E.; Smalley, R.E. J. Phys. Chem. 1982, 86, 3914.
(b) Atha, P. M.,; Hillier, I. H. Hol. Phys. 1982, 46, 437.
(c) Mo 2 : Efremov, Y.M.; Samoilova, A.N.; Kozhukhovsky, V.B.;
Gurvich, L.V. J. Hol. Spectrosc. 1978, 73, 430.
6. (a) Turner, D.W.; Baker, C.; Baker, A.D.; Brundle, C.R. "Molecular
Photoelectron Spectroscopy" Wiley: London, 1970.
(b) Rabelais, J.W. "Principles of Ultraviolet Photoelectron Spectro-
scopy" Wiley: New York, 1977.
(c) Ghosh, P.K. "Introduction to Photoelectron Spectroscopy"
Wiley: New York, 1983.
(d) Lichtenberger, D.L.; Kellogg, G.E. Ace. Chem. Res. 1987, 20, 379.
7. Koopmans, T. Physica (Amsterdam) 1934, 1, 104.
8. Bohm, M.G. J. Chem. Phys. 1983, 78, 7044.
9. Hubbard, J.L.; Lichtenberger, D.L J. Am. Chem. Soc. 1982, 104, 2132.
10. (a) Lichtenberger, D.L.; Kellogg, G. E. in "Gas Phase Inorganic
Chemistry" (Ed. Russell, D.H.), Plenum: New York, 1989, Chapter
8, pp. 245-277.
(b) Lichtenberger, D.L.; Darsey, G.P.; Kellogg, G.E.; Sanner, R.D.;
Young, V.G., Jr.; Clark, J.R. J. Am. Chem. Soc. in press.
11. (a) Green, J.C.; Hayes, A.J. Chem. Phys. Lett. 1975, 31, 306.
(b) Coleman, A. W.; Green, J. G.; Hayes, A. J.; Seddon, E. A.; Lloyd,
D. R.; Niwa, Y. J. Chem. Soc., Dalton Trans., 1979, 1057.
(c) Lichtenberger, D.L.; Blevins, C.H., II J. Am. Chem. Soc. 1984,
106, 1636.
12. (a) Herzberg, G. "Molecular Spectra and Molecular Structure" Van
Nostrand Reinhold: New York, 1950; Vol.l, pp. 101-103.
(b) Eland, J.H.D.; Danby, C.J. Int. J. Hass. Spectrom. Ion Phys.
1968, 1, 111.
(c) Yersin, H.; Otto, H.; Zink, J.I.; Gliemann, G. J. Am. Chem. Soc.
1980, 102, 951.
(d) Hipps, K.W.; Merrell, G.A.; Crosby, G.A. J. Phys. Chem. 1976, 80,
2232.
(e) Mazur, U.; Hipps, K.W. J. Phys. Chem. 1980, 84, 194.
(f) Fordyce, W.A.; Brummer, J.G.; Crosby, G.A. J. Am. Chem. Soc.
1981, 103, 7061.
296
(g) Tutt, L.; Tanner, D.; Heller, E.J.; Zink, J.l. Inorg. Chem. 1982,
21, 3859.
13. Totally symmetric vibrational modes normally dominate the vibrational
progressions seen in PE spectra:
(a) Calabro, D.C.; Hubbard, J.L.; Blevins, C.H., II; Campbell, A.C.;
Lichtenberger, D.L. J. Am. Chem. Soc. 1981, 103, 6839.
(b) Stevens, A. E.; Feigerle, C.S.; Lineberger, W.C. J. Am. Chem. Soc.
1982, 104, 5026.
(c) Lee, S.-Y. J. Chem. Educ. 1985, 62, 561.
14. (a) Bratton, W.K.; Cotton, F.A.; Debeau, M.; Walton, R.A. J. Coord.
Chem. 1971, 1, 121.
(b) Cowman, C.D.; Gray, H.B. J. Am. Chem. Soc. 1973, 95, 8177.
(c) Mortola, A.P.; Moskowitz, J.W.; Rosch, N.; Cowman, C.D.; Gray,
H.B. Chem. Phys. Lett. 1975, 32, 283.
(d) Cowman, C.D.; Trogler, W.C.; Gray, H.B. Isr. J. Chem. 1977, 15,
308.
(e) Trogler, W.C.; Solomon, E.l.; Trajberg, I.B.; Ballhausen, C.J.;
Gray, H.B. Inorg. Chem. 1977, 16, 828.
(f) Fanwick, P.E.; Martin, D.S.; Cotton, F.A.; Webb, T.R. Inorg.
Chem. 1977, 16, 2103.
(g) Hutchinson, B.; Morgan, J.; Cooper, C.B.; Mathey, Y.; Schriver,
D.F. Inorg. Chem. 1979, 18, 2048.
15. (a) Solid State: Cotton, F.A.; Mester, Z.C; Webb, T.R. Acta Cryst.
1974, B30, 2768.
(b) Gas Phase: Kelley, M.H.; Fink, M. J. Chem. Phys. 1982, 76, 1407.
16. Manning, M.C.; Trogler, W.C. Inorg. Chem. 1982, 21, 2797.
17. Martin, D.S.; Newman, R.A.; Fanwick, P.E. Inorg. Chem. 1979, 18, 2511.
18. (a) Angell, C.L.; Cotton, F.A.; Frenz, B.A.; Webb, T.R. J. Chem.
Soc., Chem. Commun. 1973, 399.
(b) Cotton, F.A.; Frenz, B.A.; Pederson, E.; Webb, T.R. Inorg. Chem.
1975, 14, 391.
19. Cotton, F.A.; Feng, X.; Matusz, M. Inorg. Chem. 1989, 28, 594.
20. Cotton, F.A.; Frenz, B.A.; Webb, T.R. J. Am. Chem. Soc. 1973, 95,
4431.
21. Bino, A.; Cotton, F.A. Inorg. Chem. 1979, 18, 3562.
22. Lichtenberger, D.L.; Copenhaver, A.S., J. Chem. Phys., in press.
23. Miskowski, V.M.; Brinza, D.E. J. Am. Chem. Soc. 1986, 108, 8296.
24. Fordyce, W.A.; Brummer, J.G.; Crosby, G.A. J. Phys. Chem. 1981, 103,
7061.
25. (a) Heller, E.J. J. Chem. Phys. 1975, 62, 1544.
(b) Heller, E.J. J. Chem. Phys. 1978, 68, 3891.
(c) Kulander. K.C.; Heller, E.J. J. Chem. Phys. 1978, 69, 2439.
(d) Heller, E.J. Ace. Chem. Res. 1981, 14, 368.
(e) Tutt, L.; Tanner, D.; Heller, E.J.; Zink, J.I. Inorg. Chem. 1982,
21, 3859.
26. (a) Ketteringham, A.P.; Oldham, C. J. Chem. Soc., Dalton Trans. 1973,
1067.
(b) Ketteringham, A.P.; Oldham, C.; Peacock, C.J. J. Chem. Soc.,
Dalton Trans. 1976, 1640.
27. Counter ion- Y3 +: Cotton, F.A.; Davison, A.; Day, V.W.; Friedrich,
M.F.; Orvig, C.; Swanson, R. Inorg. Chem. 1982, 21, 1211.
28. (a) Counter ion- [NH 4 ]+: Cotton, F.A.; Shive, L.S. Inorg. Chem.,
1975, 14, 2032.
(b) Counter ion- K+: Cotton, F.A.; Pedersen, E. Inorg. Chem., 1975,
14, 383.
29. Counter ion- [(n-C 4H9 ) 4N]+: Cotton, F.A.; Daniels, L.; Davison, A.;
Orvig, C. Inorg. Chem. 1981, 20, 3051.
30. Cotton, F.A.; Dunbar, K.R.; Falvello, L.R.; Tomas, M.; Walton, R.A.
J. Am. Chem. Soc. 1983, 105, 4950.
31. Cotton, F.A.; Fanwick, P.E.; Fitch, J.W.; Glicksman, H.D.; Walton,
R.A. J. Am. Chem. Soc. 1979, 101, 1752.
32. (a) Fanwick, P.E.; King, M.K.; Tetrick, S.M.; Walton, R.A. J. Am.
Chem. Soc. 1985, 107, 5009.
297
(b) Agaskar, P.A.; Cotton, F.A.; Dunbar, K.R.; Falvello, L.R.;
Tetrick, S.M.; Walton, R.A. J. Am. Chem. Soc. 1986, 108, 4850.
33. (a) Ebner, J.R.; Walton, R.A. Inorg. Chem. 1975, 14, 1987.
(b) Cotton, F.A.; Frenz, B.A.; Ebner, J.R.; Walton, R.A. J. Am. Chem.
Soc. 1976, 15, 1630.
34. (a) Cotton, F.A.; Hubbard, J.L.; Lichtenberger, D.L.; Shim, I. J. Am.
Chem. Soc. 1982, 104, 679.
(b) Root, D.R.; Blevins, C.H., II; Lichtenberger, D.L.; Sattelberger,
A.P.; Walton, R.A. J. Am. Chem. Soc. 1986, 108, 953.
(c) Hinch, G.D., Ph.D. Dissertation, University of Arizona, 1989.
35. (a) Siegbahn, K.; Nordling, C.; Johansson, G.; Hedman, J.; Heden,
P.F.; Hamrin, K.; Gelius, U.; Bergmark, T.; Werme, L.O.; Manne,
R.; Baer, Y. "ESCA Applied to Free Molecules" North-Holland:
Amsterdam, 1969.
(b) Briggs, D. "Handbook of X-ray and Ultraviolet Photoelectron
Spectroscopy" Heyden: London, 1977.
(c) Jolly, W.L. Ace. Chem. Res. 1983, 16, 370.
36. Blevins, C.H. II, Ph.D. Dissertation, University of Arizona, 1984.
37. (a) Atha, P.M.; Ford, P.C.; Garner, C.D.; MacDowell, A.A.; Hillier,
I.H.; Guest, M.F.; Saunders, V.R. Chem. Phys. Lett. 1981, 84,
172.
(b) Atha, P.M.; Berry, M.; Garner, C.D.; Hillier, I.H.; MacDowell,
A.A. J. Chem. Soc., Chem. Commun. 1981, 1027.
(c) Atha, P.M.; Hillier, I.H.; MacDowell, A.A.; Guest, M.F. J. Chem.
Phys. 1982, 77, 195.
(d) Atha, P.M.; Campbell, J.C.; Garner, C.D.; Hillier, I.H.;
MacDowell, A.A. J. Chem. Soc., Dalton Trans. 1983, 1085.
38. (a) Bancroft, G.M.; Pellach, E.; Sattelberger, A.P.; McLaughlin, K.W.
J. Chem. Soc., Chem. Commun. 1982, 752.
(b) Sattelberger, A.P. in "Inorganic Chemistry: Towards the 21st
Century"; Chisholm, M.H., Ed.; American Chemical Society:
Washington DC, 1983, p. 291.
39. (a) Cotton, F.A.; Stanley, G.G.; Kalbacher, B.J.; Green, J.C.;
Seddon, E.; Chisholm, M.H. Proc. Natl. Acad. Sci. U.S.A.
1977' 74' 3109.
(b) Bursten, B.E.; Cotton, F.A.; Green, J.C.; Seddon, E.A.;
Stanley, G.G. J. Am. Chem. Soc. 1980, 102, 4579.
40. Kober, E.M.; Lichtenberger, D.L. J. Am. Chem. Soc. 1985, 107, 7199.
41. Kober, E.M.; Lichtenberger, D.L. unpublished work.
42. Ketkar, S.N.; Fink, M. J. Am. Chem. Soc. 1985, 107, 338.
43. Cotton, F.A.; DeBoer, B.G.; LaPrade, M.D.; Pipal, J .R.; Ucko, D.A.
Acta Cryst. B 1971, 27, 1644.
44. (a) Kristofzski, J.G., Ph.D. Dissertation, University of Arizona,
1988.
(b) Cotton, F.A.; Extine, M.W.; Rice, G.W. Inorg. Chem. 1978, 17,
176.
(c) Cotton, F.A.; Ilsley, W.H.; Kaim, W. J. Am. Chem. Soc. 1980, 102,
3464.
(d) Wiest, R.; Benard, M. Chem. Phys. Lett. 1983, 98, 102.
45. (a) For experimental details regarding film deposition, thickness
calibration and spectral measurement see Reference 44a.
(b) Lichtenberger, D.L.; Kristofski, J.G. J. Am. Chem. Soc. 1987,
109. 3458.
46. (a) Chisholm, M.H.; Clark, D.L.; Huffman, J.C.; VanDer Sluys, W.G.;
Kober, E.M.; Lichtenberger, D.L.; Bursten, B.E. J. Am. Chem. Soc.
1987, 109, 6796.
(b) Braydich, M. D.; Bursten, B. E.; Chisholm, M. H.; Clark, D. L. J.
Am. Chem. Soc. 1985, 107, 4459.
47. Hall, M.B.; Fenske, R.F. Inorg. Chem. 1972, 11, 768.
48. (a) Bernholz, J.; Holzwarth, N. A. W. Phys. Rev. Lett., 1983, SO,
1451.
(b) Sundholm, D.; Pyykk6, P.; Laaksonen, L. Finn. Chem. Lett. 1985,
51.
298
FORMATION, STRUCTURE AND LUMINESCENT PROPERTIES OF METAL-META!.
BONDED COMPOUNDS OF TBE LATE TRANSITION METAL AND POST
TRANSITION METAL IONS
Alan L. Balch
Department of Chemistry
University of California
Davis, CA 95616
ABSTRACT
INTRODUCTION
1 M = Ir 2 M = Rh
300
interacting planar rhodium ions that are separated by 5.428(1) A and two
uncoordinated arsenic atoms whose lone pairs are directed away from the
center of the open cavity.3 Presumably the iridium analog is similar. A
variety of metal ions, including Rh(I), Ir(I), Pd(II), Cu(I), Ag(I),
Au(I), Au(III), Hg(II), In(I), Tl(I), Sn(II), Pb(II), Sb(III), and
Bi(III), have been found to coordinate to the binding sites of 1 and 2.
These sites include both the arsenic and iridium atoms within the
macrocycle and also the chloride and carbon monoxide ligands that are
bound to rhodium or iridium. Interestingly, at least nine different
types of structures have been identified in the metal ion complexes
formed from 1 and 2. This article will concern the case where metal-
metal bonding occurs and will discuss certain of the features of the
electronic spectra of these compounds.
Metal halogen bonds can add to low valent transition metal centers
to give new metal-metal bonds. For example, mercury(!!) chloride reacts
with Ir(CO)Cl(PPh3l2 to form six-coordinate Ir(HgCl) (CO)Cl2(PPh3)2 with
an Ir-Hg bond length of 2.570(1) A.4 Addition of mercury(!!) chloride to
1 results in the uptake of two moles of HgC12.5 The 31p NMR of the
product reveals that it is unsymmetrical with 51 = 18.1 ppm, &2 = 16.6Hz,
J(P, Hg) = 31.9 Hz. For the latter resonance, the satellites are
consistent with oxidative addition of one mercury (II) chloride to one
iridium ion within 1. The structure of the product,
[Ir2(HgC1)2(C0)2Cl3(~-dpma)2]Cl, as determined by X-ray crystallography,
is shown as A in Figure 2, where three different examples of structures
formed by oxidative-addition reactions are shown. As indicated by the
NMR work, one mercury (II) chloride molecule has undergone oxidative
addition to iridium wnile the second has formed the Hgcl+ unit that is
suspended between the two arsenic atoms, a chloride bridge to iridium and
the other iridium center. The Ir(2)-Hg(l) distance (2.788(3) A) is
somewhat longer than the Ir(l)-Hg(2) distance (2.573(3) A). Both appear
to involve significant metal-metal bonding.
p..........--...As
rrI/If +
r/"-A·---r
I (\: +
C'J
I ··--..J CI-P.d~lh-CI
Rh" I
Rh
' (1)
I I
C l - ' \ . . - d - Cl"
'c
~As-../
0 cf
/1 p P_/As............_,.p
Cl
301
+
302
The two preceding examples show single-center two-fragment
oxidative additions. Other examples of this sort of reactivity are
known7 and it has been established that bifunctional phosphine ligands
can facilitate such reactions, particularly for Pd-Cl additions to other
transition metals.B The constructions of 1 and 2 are such, however,
that oxidative addition of a single substrate to both metal ions should
be possible and indeed the reactions above were examined with a view
toward obtaining such an addition. This goal has been realized in the
next example. The reaction of AuCl4- with 1 proceeds in a three-fragment
two-centered fashion (Eq. 2).9
~
n+ n+
M M + X·Y·X
\.____.)
(2)
303
•u
t:
!
•0
J:J
c
304
+
305
2 lr Tl
85 picoseconds
:
n•'~
I\
Absorption j \ Emission
:
: \
: .
:
.::. .
\
~
~
: : 1.4 microseconds
:i \: ,#'·.,
/ \
I \\ / \
i \ ./ \\'..
..,.··
--· \,
'~ ....__ ........• ...... __
306
The absorption and emission spectra of the thallium adduct in
dichloromethane solution at 25° C are presented in Figure 6. An
excitation spectrum obtained for either emission parallels closely the
absorption spectrum presented here. The lifetimes observed for the two
emissions are also recorded in Figure 6. Based on the small Stokes shift
and the short lifetime, the emission at 580 nm is assigned to
fluorescence while the longer-lived emission at 814 nm results from
phosphorescence. The lifetime of the phosphorescence is within a factor
of three of that of the well-studied Pt2(~-P205H2l44- 10 and the
opportunity to observe binuclear reactions involving the excited state of
the thallium adduct exists.
307
d8d8d8
P_.-......._As.........-.....f •
r---TI--r'Co
/........-c' c'........._
cf=/
"'-.....--A-__. p
s2d&s2
-~
-::r·
~--· -a2u
J:
-~
[
-•2u
~·h . it~ *~
*~
~M~ ;M~
;M":.
'; M '!
:,M-:.
o:.M:
Figure 8 Qualitative molecular orbital diagram showing the interactions
of the filled dz 2 and empty Pz orbitals for monomeric ML 4 de
complexes and the result of stacking to form dimeric M2L8 and
trimeric M3 Ls in D4h symmetry.
308
that tendency to stack. Tl2Pt(CN)4 is stronqly luminescent in the solid
state. The blue emission results in a single sharp band centered at 444
nm. Again the bonding can be described by considering interaction of the
filled thallium s and platinum dz2 orbitals and the empty pz orbitals on
each.
ACKNOWLEDGEMENTS
Dr. Phil E. Reedy, Jr., Dr. Alan Fossett, Vince Catalano, Steve
Reimer, Ella Fung, Professor Jeff Nagle, Professor Mauro Ghedini, and
Francesco Neve have been involved with the synthetic and spectroscopic
work. Mark Chatfield and Professor R. Rosenfeld helped with the lifetime
studies. Dr. Marilyn Olmstead and Doug Oram were responsible for the x-
ray structures. The National Science Foundation (CHE-8519557) has been
generous with financial support.
REFERENCES
1. A.L. Balch. Pure and Applied Chemistry, 60: 555 (1988).
2. A.L. Balch, L.A. Fossett, M.M. Olmstead, D.E. Oram, P.E. Reedy, Jr.
J. Am. Chern. Soc , 107: 5272 (1985).
3. A.L. Balch, M. Ghedini, D.E. Oram, P.E. Reedy, Jr. Inorg. Chern ,
26: 1223 (1987).
4. P.D. Brotherton, C.L. Raston, A.H. White, S.B. Wild. J. Chern. Soc.
Dalton Trans , 1799 (1976).
5. P.E. Reedy, Ph.D. Thesis, University of California, Davis, 1987.
6. D.A. Bailey, A.L. Balch, L.A. Fossett, M.M. Olmstead, P.E. Reedy,
Jr. Inorg. Chern., 26: 2413 (1987).
7. A.L. Balch. Ann. New York Acad. Sci., 313: 651 (1978).
8. A.L. Balch in Homogeneous Catalysis with Metal Phosphine Complexes.
L.H. Pignolet, ed. Plenum Press, New York, P.l67.
9. A.L. Balch, D.E. Oram, P.E. Reedy, Jr. Inorg. Chern., 26: 1836
(1987).
10. A.L. Balch, J.K. Nagle, M.M. Olmstead, P.E. Reedy, Jr. JAm· Chern,
.s.o.c........ 109: 4123, (1987) .
11. A.L. Balch, M.M. Olmstead, D.E. Oram, P.E. Reedy, Jr., S.H. Reimer.
J. Am· Chern. Soc., in press.
12. D.M. Roundhill, H.B. Gray, C.-M. Che. Accounts Chern. ReSearch, 22:
55 (1989).
13. Balch, A.L., L.A. Fossett, J.K. Nagle, M.M. Olmstead. J. Am· Chern.
~. 110: 6732 (1988).
14. K.R. Mann, N.S. Lewis, R.M. Williams, H.B. Gray, J.G .. Gordon, II.
Inorg. Chern., 17: 828 (1978).
15. A.L. Balch, L.A. Fossett, M.M. Olmstead, P.E. Reedy, Jr.
Organometallics,_5: 1929 (1986).
16. A.L. Balch, J.K. Nagle, D.E. Oram, P.E. Reedy, Jr. J. Am. Chern.
~. 110: 454 (1988).
17. J.K. Nagle, A.L. Balch, M.M. Olmstead. J. Am· Chern. Soc., 110: 319
(1988).
309
THE PREPARATION AND CHARACTERIZATION OF NEW HETEROPOLYOXOFLUOROMETALATE
Sadiq H. Wasfi
The ammonium salt of the anion [Few17 o56 F6H5 ] 8 - has been isolated in
the crystalline form. Chemical analysis, FABMS, visible and FTIR
spectroscopy indicate that the anion has the Dawson structure with one of
the w+ 6 in the original 2:18 replaced by Fe+ 3 ion. The anion appears to
be isomorphous with that reported by Baker et al in the Thir~ Chemical
Congress of North America, June 5-10, 1988, [Znw17 o56 H4 NaF 6 ] -. Single
crystal structure study is now underway.
311
FACILE EXCHANGE OF TERMINAL, DOUBLY BRIDGING, AND QUADRUPLY BRIDGING
312
L(COh
/1 w-H
(CO)~~ /!M(C0) 3
C-M-C
I
0 (CO)z 0
M = Ru, L = C 5H 5; 3c (0.55)
M =Ru, L = C5Me5; 4c (0.22) 4d (0.62)
Scheme 1. Note: The numbers in parentheses are their
relative ratio in solution state.
313
THERMAL CONSTANTS AND STRUCTURE OF TIN CLUSTERS
314
REACTIVITY AND ISOMERIZATION OF Mo 2 (ALLYL) 4
315
SURFACE COORDINATION/ORGANOMETALLIC CHEMISTRY OF MONOMETAL AND
BIMETALLIC ELECTROCATALYSTS
REFERENCES
316
6. J. F. Rodriguez and M. P. Soriaga. J. Electrochem. Soc. 135 (1988)
616.
7. B. G. Bravo, J. F. Rodriguez, T. Mebrahtu and M. P. Soriaga. J.
Phys. Chem. 91 (1987) 5660.
8. B. G. Bravo, S. L. Michelhaugh and M. P. Soriaga. J. Electroanal.
Chem. 241 (1988) 199.
9. J. F. Rodriguez, J. E. Harris, M. E. Bothwell, T. Mebrahtu and M. P.
Soriaga. Inorg. Chim. Acta. 148 (1988) 123.
10. M. P. Soriaga. J. Electroanal. Chem. 240 (1988) 309.
11. J. F. Rodriguez, M. E. Bothwell, J. E. Harris and M. P. Soriaga. J.
Phys. Chem. 92 (1988) 2702.
12. T. Mebrahtu, G. M. Berry and M. P. Soriaga. J. Electroanal. Chem.
239 (1988) 375.
13. M. P. Soriaga. In M. P. Soriaga, Editor, "Electrochemical Surface
Science: Molecular Phenomena at Electrode Surfaces." ACS Books:
Washington, DC. Chapter 1 (1988).
14. T. Mebrahtu, G. M. Berry, B. G. Bravo, S. L. Michelhaugh and M. P.
Soriaga. Langmuir. 4 (1988) 1147.
15. B. G. Bravo, S. L. Michelhaugh, T. Mebrahtu and M. P. Soriaga.
Electrochim. Acta. 33 (1988) 1507.
16. Mebrahtu, G. M. Berry and M. P. Soriaga. J. Electroanal. Chem. 247
(1988) 241.
17. M. P. Soriaga, G. M. Berry, M. E. Bothwell, B. G. Bravo, G. J. Cali,
J. E. Harris, T. Mebrahtu, S. L. Michelhaugh and J. F. Rodriguez.
In M. P. Soriaga, Editor, "Electrochemical Surface Science:
Molecular Phenomena at Electrode Surfaces." ACS Books: Washington,
DC. Chapter 36 (1988).
18. M. E. Bothwell, J. F. Rodriguez and M. P. Soriaga. J. Electroanal.
Chem. 252 (1988) 453.
19. M. E. Bothwell and M. P. Soriaga. J. Electroanal. Chem. 260 (1989)
193.
20. J. E. Harris, M. E. Bothwell, J. F. Rodriguez and M. P. Soriaga. J.
Phys. Chem. 93 (1989) 2610.
21. J. E. Harris and M. P. Soriaga. Electrochimica Acta. In press
(1989).
22. G. M. Berry, B. G. Bravo, M. E. Bothwell, G. J. Cali, J. E. Harris,
T. Mebrahtu, S. L. Michelhaugh, J. F. Rodriguez and M. P. Soriaga.
Langmuir. 5 (1989) 707.
23. J. F. Rodriguez, T. Mebrahtu and M. P. Soriaga. J. Electroanal.
Chem. In press (1989).
24. B. G. Bravo, S. L. Michelhaugh and M. P. Soriaga. Langmuir. In
press (1989).
25. G. J. Cali, M. E. Bothwell and M. P. Soriaga. J. Electroanal. Chem.
In press (1989).
317
AMBIENT-TEMPERATURE CHLOROALUMINATE MOLTEN SALTS: SOLVENTS FOR CHLORO
318
THE MAXIMUM STRENGTH OF THE CHEMICAL BOND BETWEEN TWO METAL ATOMS
K. A. Gingerich
REFERENCES
319
REACTIONS OF (~ 7 -c 7 H 7 )M(~ 5 -c 5 H 5 ), M- Ti OR Zr, WITH CARBOXYLIC AND
DITHIOCARBOXYLIC ACIDS
320
MODEL HYDRODESULFURIZATION SYSTEMS: REACTIONS OF SULFUR CONTAINING
MOLECULES ON Ni(llO)
D. R. Huntley
321
ELECTROCHEMICAL STUDIES OF TRIANGULAR NIOBIUM CLUSTER,
322
REACTIVITY OF DITHIOETHERS TOWARD [Re 2X8 ] 2 -
323
THEORETICAL INVESTIGATIONS OF THE METAL-METAL INTERACTIONS WITHIN THE
324
STRUCTURAL AND THEORETICAL STUDIES ON HETERONUCLEAR TRANSITION-METAL
325
ELECTRONIC STRUCTURE AND NATURE OF BONDING IN TRANSITION METAL DIMERS
Irene Shim
326
THEORETICAL CALCULATIONS ON THE INTERACTION OF BRIDGING CARBONYLS WITH
327
BIMETALLIC HYDROFORMYLATION CATALYSIS
328
252 cF-PLASMA DESORPTION MASS SPECTRA OF VERY LARGE CLUSTERS
329
SYSTEMATIC KINETIC STUDIES pF ASSOCIATIVE AND DISSOCIATIVE REACTIONS OF
(1)
330
eq (1) but data over a wider range of 6( 13 co) and 9 are required before
this can be fully substantiated.
REFERENCES
331
MIXED Pd-Au AND Pt-Au CLUSTER COMPOUNDS
Louis H. Pignolet
332
STABILITY OF SMALL BICLUSTERS OF TRANSITION METALS WITH SEMI-CONDUCTORS
This work has been supported by the National Science Foundation, the
Robert A. Welch Foundation, and by NATO Grant No. RG 85/0448.
REFERENCES
333
THE TOPOLOGY OF THE TOTAL CHARGE DENSITY IN BINUCLEAR TRANSITION-METAL
REFERENCES
334
THE EFFECT OF CARBONYL LIGANDS ON OSMIUM AND RUTHENIUM METAL-METAL BONDS
335
CONTRIBUTORS
337
Jeffery, J.C. Texas A&M University 325
Jennings, J.G. Abilene Christian University 323
Jensen, K.F. University of Minnesota 215
Johnson, B.F.G. University of Cambridge 141
Johnston, R.L. University of Arizona 275
Katovic, V. Wright State University 322
Kingcade, Jr., J.E. Texas A&M University 314
Kingcade, Jr., J.E. Texas A&M University 333
Klemperer, W.G. University of Illinois 161
Lagow, R.J. University of Texas-Austin 171
Laneman, S.A. Louisiana State University 328
Lictenberger, D.L. University of Arizona 275
Lockledge, S.P. University of Illinois 161
MacDougall, P.J. Texas A&M University 334
Main, D.J. University of Illinois 161
Marks, T.J. Northwestern University 113
Marseglia, E.A. Cavendish Laboratory 141
Martuch, R.A. Cleveland State University 320
May, M. Wright State University 322
McCarley, R.E. Iowa Sate University 91
McNeal, C.J. Texas A&M University 329
Mebrahtu, T. Texas A&M University 316
Michelhaugh, S.L. Texas A&M University 316
Miller, A. E. California Institute of Technology 335
Nolan, S.P. Northwestern University 113
Osteryoung, R.A. State University of New York-Buffalo 318
Peng, S.-M. National Tsing Hua University 312
Pignolet, L.H. University of Minnesota 332
Poe, A.J. University of Toronto 330
Powell, G.L. Abilene Christian University 323
Rodger, A. University of Cambridge 141
Rodriguez, J.F. Texas A&M University 316
Rosenberg, F.S. University of Illinois 161
Sargent, A.L. Texas A&M University 324
Sayers, W. Wright State University 322
Schmude, Jr., R.W. Blinn College 314
Schulze, R.K. University of Minnesota 215
Seyam, A.M. Northwestern University 113
Sherwood, P. Texas A&M University 325
Shim, I. The Technical University of Denmark 326
Shim, I. The Technical University of Denmark 333
Simpson, II, C.Q. Texas A&M University 327
Sinfelt, J.H. Exxon 103
Smalley, R.E. Rice University 249
Soriaga, M.P. Texas A&M University 316
SriHari, S. Cleveland State University 320
Stanley, G.G. Louisiana State University 328
Stern, D. Northwestern University 113
Stone, F.G.A. University of Bristol 325
Tsay, R.-J. University of Texas-Arlington 315
Walton, R.A. Purdue University 7
Wang, R. -c. University of Illinois 161
Wang, S.-L. National Tsing Hua University 312
Wasfi, S.H. Delaware State College 311
Winpenny, R.E.P. Texas A&M University 329
Yaghi, O.M. University of Illinois 161
338
INDEX
339
hexalithiobenzene, 171 molybdenum ternary oxides, 92
hexalithioethane, 171 molybdenum(!!) carboxylates, 2
hydride tunnelling, 235 molybdic anhydride, 164
hydrides of Os, Ru, 312 molybednum sulfide, 98
hydroaminations, 122 MoW(0 2CC(CH 3 ) 3 ) 4 , 287
hydrodesulfurization, 321 Mo-Mo bond lengths, 281
hydroformylation catalysis, 328 multicenter transformations, 75
hydrogenation, 62
Ni 10 , 250
indium phosphide, 195 nickel on tungsten, 210
indium phosphide, epitaxial, 203 niobium, 19th cluster of, 262
interpenetrating icosahedra, 262 niobium ternary oxides, 92
Ir, [Ir 2 Tl(C0) 2 c1 2 (~·dpma) 2 ] nitriles, 62
[N0 3 ], 305 Ni-Au, 105
Ir, [Ir 2 (SnCl)(C~) 2 cl 2 Ni-Cu, 104
(~·dpma) 2 ] , 305 Ni-Cu, 105
Ir, {Ir 2Au(C0) 2Cl 4
(~·dpma) 2 }+, 302 octachlorodirehate(III) dianion-
Ir, {Ir 2 (HgC1) 2 (C0) 2 organometallicjcovrdination
(~·dpma) 2 }, 302 chemi~try, 316
Ir(HgCl)(CO)Cl 2 (PPh 3 ) 2 , 301 Os 10 c(C0) 24 ·, 129
Ir-Cu, 110 Os 3 (C0) 9 x 2 ~ 265
isocyanides, 62 Os 5 C(C0) 14 -, 129
Os 6 6(C0) 18 , 76
lanthanide-alkyl bonds, 121 osmium hydride cluster, 137, 312
late transition metal ions, 299 osmium sulfur-containing cl~ster, 87
Li, (CLi 2H2 ) 4 , 171 Os-Cu, 110
Li, (CLi 4 ) 3 , 171 oxametallacyclobutane, 167
lithium arsenide, 197 oxidations, 161
lithium phosphide, 197 oxidative addition, dinuclear, 59
oxides, 161
M2 (o 2ccH 3 ) 4 of chromium, oxo-alkylidenes, 67
molybdenum and tungsten,
277 Pd-Ag, 105
magic angle spinning, 233 Pd-Au, 104
McMurray reagent, 63 Pd-Au and Pt-Au cluster compounds, 332
metal carbonyl clusters, 149, 231 pentacoordinate complexes, 234
metal carbonyls, dinuclear, 238 photoelectron spectroscopy, 275
metallic aluminum films, 216 phthalic anhydride, 167
metallomacrocycles, 299 plasma desorption mass spectra, 329
metal-ligand bond enthalpies, 113 polycrystalline Al films, 223
metal-metal bonded arrays, 91 polylithium, 171
metal-metal multiple bonds, post transition metal ions, 299
1' 2' 7' 19 pseudorotation4 141, 144, 148
Metal-Metal triple bond, 9 Pt 2 (~-P 2 o 5 H 2 ) 4 ·, 307
methanation catalyst, 208
methyl formate, 45 Quaduruple bond, 8
mixed-metal (Au-Pt, Ag-Pt)
electrocatalysts, 316 Re 2Br 82- , 2
Mo, (H0) 4MoMo(PH 3 ) 4 , 26 Re 2c1 4 (PR 3 ) 4 , 28
MO calculations, 272 Re 3c1 9 (dppm) 3 , 11
Mo 2 , 32 reactivity of dithioethers toward
Mo 2 (ALLYL) 4 , 315 [Re 2X ] 2 -, 323
Mo 2 (o 2 ccF 3 ) 4 , 287 reductive e~imination, dinuclear, 59
Mo 2 (o 2 CR) 4 , 4 Re-Cu, 110
Mo 2 (0CH 2 CMe 3 ) 6 , 289 Rh, [Rh 2 Pd(C0) 2 c1 3 (~·dpma) 2 ]+, 301,
Mo 2 (PH 3 ) 4 cl 4 , 24 302
Mo 6 o 12 ~ 97 Rh 2 (o 2CR) 2 , 4
Mo 8o14 -, 99 Rh4 (C0) 12 , 246
molecular beam epitaxy, 195 Rh-Cu, 109
molybdenum, 55 Ru 2 (o 2CR) 4 , 4
340
Ru 3 (C0) 12 , 50 Tc 2c1 83- , 2
Ru 6C(C0) 16 2 -, 129 technetium clusters, 12
RuC1 3 , 48 tetralithioethylene, 171
ruthenium catalyst, 43 tetralithioethylene, 179
Ru-Cu, 106 tetralithiomethane, 171
Ru-Cu, 109 tetraruthenium cluster, 77
thin film PE spectrum of
semiconductor clusters, 249 Mo 2 (o 2ccH 3 ) 4 , 292
semiconductors, 95 Tl 2Pt(CN) 4 , 307
Si +, 261 triangular niobium cluster, 322
sii~con clusters, 255 triirondodecacarbonyl, 240
skeletal electron-pair trilithiomethane, 171
counting, 265 trinuclear Au2Pt(CH 2 (S)PH 2 ) 4 , 324
s~ (n- 4-7), 314 triosmium clusters, 77
soiid state carbon-13, 231 tungsten, 55
solid state clusters, 91
spin lattice relaxation times,
232
stability of small biclusters, 311
333
stored waveform inverse
Fourier transform, 258
structure sensitive reactions,
105
supersonic cluster beam ICR, 252 Xa-SW calculations, 23, 24, 26
systematic kinetic studies
of metal carbonyl ynamine (aminoacetylene), 85
clusters, 330
zeolite cages, 129
zeolites, 50
341