Synthetic and spectroscopic studies of metal carboxylate dimers

Material Information

Synthetic and spectroscopic studies of metal carboxylate dimers
Telser, Joshua A., 1958- ( Dissertant )
Drago, Russell S. ( Thesis advisor )
Ryschkewitsch, George ( Reviewer )
Weltner, William ( Reviewer )
Stoufer, R. Carl ( Reviewer )
Maddala, G. S. ( Reviewer )
Place of Publication:
Gainesville, Fla.
University of Florida
Publication Date:
Copyright Date:
Physical Description:
vi, 308 leaves : ill. ; 28 cm.


Subjects / Keywords:
Adducts ( jstor )
Carboxylates ( jstor )
Dimers ( jstor )
Ligands ( jstor )
Pyridines ( jstor )
Rhodium ( jstor )
Solvents ( jstor )
Spectral bands ( jstor )
Spectral methods ( jstor )
Subroutines ( jstor )
Chemistry thesis Ph. D
Dissertations, Academic -- Chemistry -- UF
Molybdenum ( lcsh )
Rhodium ( lcsh )
Ruthenium ( lcsh )
Transition metal compounds -- Spectra ( lcsh )
bibliography ( marcgt )
non-fiction ( marcgt )


Synthetic and spectroscopic studies on several complexes in the metal carboxylate series are described. These complexes are of general formula M 2 (02CR) 4 where M is a transition metal and "0 2 CR is a bridging carboxylate ligand. The metals used in this study are molybdenum, rhodium, and ruthenium. The studies were undertaken to help understand the nature of the metal -metal interaction in these complexes and to see what effect this interaction has on the reactivity of these complexes. To effect removal of the bridging carboxylate ligands, acetonitrile solutions of Rh2(0 2 CCH 2 CH 2 CH3) 4 and Mo 2 (0 2 CCH 3 ) 4 were reacted with stoichiometric amounts of the strong non-complexing acids CF3SO3H and (CH 3 CH 2 ) 20.HBF4 . This generated Rh2 (02 CH 2 CH 2 CH 3 ) 2 2+ and Mo 2 (02CCH 3 )? 2+ species in solution. The former was not isolated, but characterized in solution by NMR and UV-visible spectroscopy. Two derivatives of Mo 2 (02 CCH 3 ) 2 2+ were isolated: CMo 2 (0 2 CCH3) 2 (CH 3 CN) 4 ](CF 3 S03) 2 and [Mo 2 (0 2 CCH3) 2 (CH3CN) 5 ](BF 3 0H) 2 . The reactivity of the former complex towards oxidative addition was investigated. The complex was found to be quite stable towards oxidation in contrast to other Mo(II) complexes. Rhodium trifluoroacetate was reacted with various Lewis bases to give adducts of general formula Rh2 ( 02^3)482 as had been previously reported for Rh 2 (0 2 CR)4. However, with pyridine and £-butyl Isonitrile, complexes of general formula Rh2(0 2CCF3)4B4 were isolated constituting a new class of adduct. With phosphorus donors, Rh-Rh bond cleavage occurred to give monomeric Rh(I) and Rh(III) complexes. This demonstrates enhanced reactivity for Rh2(02CCF3)4 compared to rhodium alkyl carboxylate dimers. The chemical and electrochemical generation of Rh2(02CCH 2 CH 2CH3)4B2 + is described. These results and EPR spectra of these species are explained using a molecular orbital model. The strength of the rhodium Lewis base interaction determines the chemical and spectroscopic properties of these species. The formally mixed oxidation state complex Ru2(02CCH2CH2CH3)4Cl was studied by powder magnetic susceptibility measurements over the temperature range 5-300 K, by EPR spectroscopy In various glasses at 4 K and by Far IR spectroscopy at room temperature. In agreement with previous reports, the complex has a quartet ground state with unpaired electron spin density delocalized over both Ru atoms. Reactivity studies of this compound with Lewis bases are described. A bispyridine adduct of ruthenium butyrate chloride is reported.
Thesis (Ph. D.)--University of Florida, 1984.
Includes bibliographic references (leaves 298-307).
General Note:
General Note:
Statement of Responsibility:
by Joshua A. Telser.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
000496904 ( alephbibnum )
12041966 ( oclc )
ACR6131 ( notis )


This item has the following downloads:

Full Text








There are many people and things responsible for a successful

graduate career and I would like to take this opportunity to mention


First of all, I would like to thank my research director, Professor

Russell S. Drago, for his continuous help from near and from afar. It

was a privilege to be part of a research group that has accomplished so

much in so many areas over so long.

Since a group is not one man, I would also like to thank the many

members of the Drago group who have helped me out: Charlotte Owens,

Rich Cosmano, Barry Corden, Pete Doan, Dave Pribich, Carl Bilgrien, Andy

Griffis, and Ernie Stine.

Since a group is not alone, I would also like to thank the faculty

and students of other groups at both Illinois and Florida. In

particular, thanks are due to Professor R. Linn Belford and Jeff

Cornelius and to Professor William Weltner, Jr., and Richard Van Zee.

Since faculty and students cannot do everything themselves, I would

also like to thank the many support personnel who made my work a lot

easier. In particular, I greatly appreciate the help of the glass shop

and NMR and Elemental Analysis Labs at both Illinois and Florida.



ACKNOWLOGEMENTS....................... ........................ ii

ABSTRACT...................... ................................ v

CHAPTER I. GENERAL INFORMATION................................ 1

Introduction ............................................. 15
Results and Discussion.................................... 19
Conclusion................................................ 57
Experimental Section...................................... 58

VARIOUS LEWIS BASES............................... 65
Introduction................. ............................ 65
Results and Discussion.................................... 68
Conclusion............................................... 113
Experimental Section..................................... 114

Introduction............................................. 123
Results and Discussion................................... 128
Conclusion............................................... 144
Experimental Section..................................... 145

BUTYRATE CHLORIDE.................................. 147
Introduction............................................. 147
Results and Discussion................................... 153
Conclusion............................................... 190
Experimental Section..................................... 191

CHAPTER VI. GENERAL CONCLUSIONS............................... 197

SUSCEPTIBILITY DATA ............................... 199

SPIN HAMILTONIAN USED IN METHOD 1.................. 200


S = 2 = 3/2, USED IN METHOD 5.................... 201

SPIN HAMILTONIAN USED IN METHOD 5.................. 202

SIMULATIONS...................................... 203


SPECTRAL SIMULATIONS.............................. 273

REFERENCES.................................................... 298

BIOGRAPHICAL SKETCH........................................... 308

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



Joshua A. Telser

December 1984
Chairman: Professor Russell S. Drago
Major Department: Chemistry
Synthetic and spectroscopic studies on several complexes in the

metal carboxylate series are described. These complexes are of general

formula M2(O2CR)4 where M is a transition metal and -02CR is a bridging

carboxylate ligand. The metals used in this study are molybdenum,

rhodium, and ruthenium. The studies were undertaken to help understand

the nature of the metal-metal interaction in these complexes and to see

what effect this interaction has on the reactivity of these complexes.

To effect removal of the bridging carboxylate ligands, acetonitrile

solutions of Rh2(02CCHCH22CH3)4 and Mo2(02CCH3)4 were reacted with

stoichiometric amounts of the strong non-complexing acids CF3SO3H and

(CH3CH2)20.HBF4. This generated Rh2(02CH2CH2CH3)22+ and Mo2(02CCH3)22+
species in solution. The former was not isolated, but characterized in

solution by NMR and UV-visible spectroscopy. Two derivatives of

Mo2(02CCH3)22+ were isolated: [Mo2(O2CCH3)2(CH3CN)4](CF3SO3)2 and

[Mo2(02CCH3)2(CH3CN)5](BF3OH)2. The reactivity of the former complex
towards oxidative addition was investigated. The complex was found to

be quite stable towards oxidation in contrast to other Mo(II)

complexes. Rhodium trifluoroacetate was reacted with various Lewis

bases to give adducts of general formula Rh2(02CCF3)4B2 as had been

previously reported for Rh2(02CR)4. However, with pyridine and t-butyl

isonitrile, complexes of general formula Rh2(02CCF3)4B4 were isolated

constituting a new class of adduct. With phosphorus donors, Rh-Rh bond

cleavage occurred to give monomeric Rh(I) and Rh(III) complexes. This

demonstrates enhanced reactivity for Rh2(O2CCF3)4 compared to rhodium

alkyl carboxylate dimers. The chemical and electrochemical generation

of Rh2(02CCH2CH2CH3)4B2+ is described. These results and EPR spectra of

these species are explained using a molecular orbital model. The

strength of the rhodium Lewis base interaction determines the chemical

and spectroscopic properties of these species. The formally mixed-

oxidation state complex Ru2(O2CCH2CH2CH3)4C1 was studied by powder

magnetic susceptibility measurements over the temperature range 5-300 K,

by EPR spectroscopy in various glasses at 4 K and by Far IR spectroscopy

at room temperature. In agreement with previous reports, the complex

has a quartet ground state with unpaired electron spin density

delocalized over both Ru atoms. Reactivity studies of this compound

with Lewis bases are described. A bispyridine adduct of ruthenium

butyrate chloride is reported.


The discovery of transition metal complexes containing metal-metal

bonds is a relatively recent one. This discovery and much of the

subsequent progress towards understanding metal-metal bonded complexes

have been discussed in detail by Cotton and Walton in their book

"Multiple Bonds Between Metal Atoms"' as well as in various review

articles by others.2-6 Nevertheless, a brief summary of the historical

background of this class of complexes is in order.

For the first half of the 20th century, transition metal chemistry

was dominated by the concepts developed by Alfred Werner.7 That is,

most complexes were thought of as what are now referred to as classical

coordination compounds, a central transition metal ion surrounded by

electron donating ligands usually in an octahedral orientation. Square

planar, tetrahedral and other geometries were known, but the concept

that a compound could exist in which there were several metals

interacting in various ways had not been suggested. Metal-metal

interactions were something thought to occur in bulk metals and not in

complexes with oxidized metals. This idea was so firmly held that

compounds that were synthesized during that time that contained metal-

metal bonds were not investigated further. Notable examples are

chromium acetate, first prepared in 1844,8 and various tantalum9 and

molybdenuml0 halides synthesized during the early part of the 20th


With the advent of improved methods of crystal structure

determination by x-ray diffraction, the discovery of metal-metal bonds

in dimeric and cluster compounds became inevitable. It was in the metal

carbonyl complexes that metal-metal bonding was first demonstrated. The

reason for this may be a practical one in that these compounds were

relatively easy to study, but it may also be a philosophical one. Metal

carbonyls and other organometallic compounds are newer and quite

different from classical coordination compounds and thus understanding

them was not hampered by older ideas that would be invariably applied to

complexes such as metal halides and carboxylates. In Fe2(CO)g in 1938

and with greater certainty in Mn2(CO)10 in 195711 metal-metal single

bonds were first proposed. The existence of metal-metal single bonds in

carbonyl complexes containing two to twelve or more metal atoms is

widely accepted.

The existence of multiple bonds between metal atoms was not

originally found in carbonyl complexes, but in some rhenium halides.

The stoichiometry and the structure, when it was eventually

determined,12 of [Re2Clg]2- could not be explained by classical

theories. It was necessary to invoke a multiple metal-metal bonding

scheme.13 A qualitative diagram of this molecular orbital (MO) scheme

is shown in Figure 1-1. In a compound such as Mn2(CO)10 it is not

surprising that the odd d electron on each manganese pairs up to give a

single a bond. However, it is surprising that in a Re(III) compound all

four d electrons pair up to give a quadruple bond. The evidence for

this quadruple bond comes primarily from the crystal structure. The Re-

Re distance is extremely short, 2.222 A in (n-Bu)4NRe2Cl8, and there is

no twist angle between the ReCI4 subunits so that the chlorides are

Figure 1-1. Formation of metal-metal bond molecular orbitals
from individual metal d atomic orbitals.



+ a !

2 2
dx -y







= 77


dxy 2






__ c


fully eclipsed.14 This sterically unfavorable structure is the result

of the fourth bond, the 6 bond between the Re atoms which arises

primarily from the in-phase addition of the dxy orbitals. The in-phase

addition of the two dx2_y2 orbitals could give another 6 bond; however

these orbitals are usually primarily involved in forming metal-ligand

bonds and are thus rarely considered in MO schemes for metal-metal

bonded complexes. The other three Re-Re bonds are derived in a more

conventional manner, analogous to the triple bond well known for

alkynes. The in-phase addition of the two dz2 orbitals gives the a bond

and the in-phase addition of the four dxz and dyz orbitals gives a

degenerate pair of r orbitals. The out-of-phase addition of these Re d

orbitals gives a corresponding set of higher energy anti-bonding 6*, a*,

and r* orbitals. In the Re(III) dimer the eight d electrons just fill

all the bonding orbitals giving a diamagnetic compound with a total bond

order of four. Subsequent to this work on the rhenium dimer, other

dimeric metal-metal bonded complexes were studied and their properties

explained using this MO scheme. The previously known complex

Cr2(O2CCH3)4(H20)2 was reinvestigated and proposed15 to have a quadruple

bond resulting from the pairing up of the four d electrons on each

Cr(II) in the same manner as described above for Re(III). Chromium is

the only first row transition metal proposed to form a multiply bonded

dimer. The small size of first as opposed to second and third row

transition metals makes it less convincing that enough orbital overlap

occurs to give a quadruply bonded complex. It has even been suggested

that no Cr-Cr bond exists at all in the chromium dimers.16 Copper forms

the well known complex Cu2(O2CCH3)4(H20)2 which is isostructural with

chromium acetate. This general structure is shown in Figure 1-2. This

Figure 1-2. General structure for metal carboxylate dimer with
idealized D4h symmetry. Axial ligands, L, need not
be present.


o /R




~ --~-0


copper complex was studied by Bleaney and Bowers17 a number of years

ago, before the multiple metal-metal bond theory was proposed. They

found that the two copper atoms were strongly antiferromagnetically

coupled, but there was no true Cu-Cu bond. Thus it is possible, by

analogy between the two first row transition metal dimers, that chromium

acetate has a bond order of less than four with the remaining d

electrons antiferromagnetically coupled.

A large number of second and third row transition metal dimers have

been reported and in these there is little doubt that the metal-metal

bond MO scheme is valid. In addition to crystal structure

determinations of many metal-metal bonded complexes, other spectroscopic

and theoretical studies have been performed to confirm the generality of

this MO scheme. Metal-metal bonded complexes are known for molybdenum,

tungsten, technetium, rhenium, ruthenium, osmium, rhodium, and

platinum. The compounds are far too numerous to list; however some

examples of each will be given. The first studied was the rehnium

chloride and a variety of multiple metal-metal bonded rhenium,

technetium, and molybdenum halides are known.1-4 However, carboxylate

complexes were also among the first known as with the chromium and

copper acetates mentioned above. This thesis is concerned with the

metal carboxylate dimers and thus to show the generality of this type of

ligand in forming metal-metal bonded complexes, one should note the

following compounds: Re2(02C(CH3)3)4C12, Tc2(O2C(CH3)3)4C12,

Ho2(O2CCH3)4, W2(02CCF3)4, Ru2(02CCH3)4CI and Rh2(02CCH3)4 In these

complexes the bond orders range from four in Re, Tc, Mo and W to 2.5 in

Ru to one in Rh. These bond orders can be easily determined using the

MO scheme in Figure 1-1 and adding the appropriate number of d electrons
for both metals.

Given that a great variety of metal-metal bonded complexes exist

and that a qualitative MO scheme exists which can explain their overall

properties, the question then arises as to why would one wish to study

them further. There are several reasons to do so. First, as was

indicated specifically with the chromium carboxylate, the exact nature

of the metal-metal bonding in these complexes is not yet completely

resolved. Second, the wide range in bond orders in metal-metal bonded

complexes such as the carboxylate dimers means that there is a wide

range in strength of metal-metal interaction. Thus, what exists here is

an isostructural series for which comparison of the reactivity and

spectroscopic properties of members of the series affords a direct means

of understanding the effect different metal-metal interactions have on

the chemistry of transition metal complexes. One can also compare the

reactivity and spectroscopic properties of a metal-metal bonded complex

to that of a monomeric complex of the same metal. The presence of two

or more metals in close proximity leads to the possibility of metal

synergism.18 This means that one metal can influence the chemistry and

the other metal site leading to different reactivity than would be

expected for noninteracting multi-metal or monomeric metal systems.

Synergism from metal clusters is proposed in a number of biological

systems such as the ferredoxins, nitrogenase, cytochrome oxidase, and

copper type 3 proteins.19'20 Synergism in metal carboxylate clusters

has also been used to model surface reactions.21 The variety of metals

and their oxidation states that form metal carboxylate dimers makes this

class of complexes a good model system to study synergistic effects.

The nature of the synergistic effect in metal carboxylate dimers has

been previously examined by Richman and co-workers.22-26 The enthalpies

of Lewis base binding to the vacant coordination sites along the M-M

axis in several complexes of general type M2(02CR)4 were measured.

These sites will hereafter be referred to as axial sites since they are

along the metal-metal bond axis. Enthalpies of formation of both 1:1

and 2:1 Lewis base axial adducts were obtained for metal carboxylate

dimers such as Rh2(O2CCH2CH2CH3)4, Rh2(02CCCF2CFC3)4 and

Mo2(02CCF2CF2CF3)4 These will be abbreviated as Rh2(but)4, Rh2(hfb)4
and Mo2(hfb)4, respectively. Comparison of the enthalpies for 1:1 and

2:1 axial adduct formation clearly showed significant changes in the

acidic properties of the second metal as a result of base coordination

to the first. Significant differences between rhodium and molybdenum

systems were also found. Use of the E and C equation27-29 and

modifications thereof allowed quantitative understanding of these

effects. It was found that inductive transfer of electrostatic

properties of the base was more effective through the shorter Mo-Mo

quadruple bond than through the longer Rh-Rh single bond. Inductive

transfer of the covalent properties of the base was more effective

through the more polarizable Rh-Rh bond than the Mo-Mo bond.

Differences in the type of Lewis acid-base interactions were also found

for the two metals. Using the MO scheme in Figure 1-1, disregarding the

second S bond, it can be seen that the molydenum carboxylate dimer with

the total of eight d electrons from the two Mo(II) subunits has no 7*

electron density. In contrast, the rhodium carboxylate dimers with a

total of 14 d electrons from the two Rh(II) subunits have filled v*

orbitals. This r* electron density can interact with empty n* orbitals

on bases with these orbitals of the right energy. Thus, a higher than

expected enthalpy of adduct formation was found for the rhodium

carboxylate dimers with Lewis bases such as pyridine and acetonitrile.

These bases can function as w-acceptors as well as o-donors. No such w-

backbonding stabilization was found for the molybdenum carboxylate dimer

since it has vacant n* orbitals.

A final reason for studying metal-metal bonded complexes, in

addition to understanding synergistic effects where one metal influences

the reactivity of the other, is to understand reactions where both

metals are directly involved. An example would be the reaction of M-M

with some X-Y species to give M-X and M-Y. This can be considered an

oxidative addition and would be analogous to many reactions of organic

compounds, particularly those with carbon-carbon multiple bonds. The

reactivity of metal-metal bonded complexes has been recently reviewed30

and there are many examples of this type of reaction. However, most

involve organometallic complexes such as metal carbonyl clusters. It is

not clear that this reactivity would occur to as great an extent in

carboxylate or halide complexes wherein the metals are generally in a

higher oxidation state.

One means of enhancing the reactivity of and understanding the

metal-metal interaction in metal carboxylate dimers is to achieve varied

ligand coordination to other than just the axial sites. As can be seen

in Figure 1-2, the sites perpendicular to the metal-metal axis are fully

occupied by the bridging carboxylate ligands. These sites will be

referred to hereafter as equatorial sites. If these ligands could be

wholly or partly removed, then the reactivity of the metal-metal b"nd

could be better investigated. It was reported a number of years ago by

Legdzins and co-workers31'32 that strong, non-complexing aqueous acids

could protonate the bridging carboxylates generating in solution species

with available equatorial coordination sites. The effect of strong,

non-complexing acids on metal carboxylate dimers of rhodium and

molybdenum in organic solvents, chiefly acetonitrile, is the subject of

the second chapter of this thesis.

Another approach to achieving ligand coordination to the equatorial

sites is to use a more poorly coordinating bridging carboxylate to begin

with. Recently, Girolami and co-workers33 obtained unusual products

from the reaction of Mo2(02CCF3)4 with various Lewis bases, primarily

phosphorus donors. In some of the products, equatorial rather than

axial Lewis base coordination was observed. This was found for

phosphines that were sterically small and good a-donors. Thus, use of a

fluorocarboxylate rather than an alkylcarboxylate can lead to equatorial

coordination without the need for protonation of the carboxylate by

strong acid. The reactivity of Rh2(02CCF3)4 towards Lewis bases was

systematically investigated and is the subject of the third chapter of

this thesis.

Another interesting aspect of multi-metal systems, besides their

variety of ligand coordination sites, is the variety of oxidation states

available to them. A monomeric complex might have two accessible

oxidation states, an oxidized and a reduced state. In contrast, a

dimeric complex of this metal could well have more since the two metals

could be both oxidized or both reduced or one of each to give a wider

range of electrochemical behavior. This ability is one reason why metal

clusters are proposed to play an important role in biological redox

processes such as the photochemical oxidation of water in

photosynthesis.34 Thus one would expect metal-metal bonded complexes to

exhibit a wide range of oxidation states. For metal-metal bonded dimers

this is only partly true. Some variety of oxidation state does exist.

For Mo, W, Tc and Re stable complexes with bond orders of 3, 3.5, and 4

are known1 and they can be electrochemically interconverted. In these

the metals are in the (III,III), (II,III) and (II,II) oxidation states

respectively for Mo and W with the order reversed for Re and Tc. Only

the (II,II) oxidation state for Rh and the (II,III) state for Ru give

stable metal-metal bonded complexes that have been well characterized.

Thus, the range of oxidation states in some metal-metal bonded dimeric

systems is no greater than in monomeric complexes. Nevertheless, the

redox behavior of metal carboxylate dimers is of interest and the use of

electrochemical methods and EPR spectroscopy is helpful in understanding

this behavior. A number of studies of this nature have been made on
metal carboxylate dimers and related complexes by Cotton and Pedersen35-

39 and by other workers40,41 and show that the redox and EPR properties
of these complexes can be explained by the qualitative MO scheme

described above. In a study by Drago and co-workers23 the effect of

both axial Lewis base coordination and different caboxylate ligands was

quantitatively examined. Oxidation of the dimer is easier when the base

is strongly donating and the carboxylate is not electron with-drawing.

This oxidation converts the diamagnetic dimers to paramagnetic complexes

which can subsequently studied by EPR spectroscopy. This technique

gives information on the electronic properties of the complex that can

be directly compared to theoretical studies. In addition to the

qualitative MO scheme described earlier, a number of quantitative
studies using a variety of calculational methods have been performed.42-
47 However, these results are not always in full agreement with

experimental data. The generation of paramagnetic rhodium carboxylate

dimer species and comparison between these experimental and various

theoretical results is the subject of the fourth chapter of this thesis.

In contrast to the normally diamagnetic metal carboxylate dimers of

rhenium, molybdenum, rhodium, and others mentioned above, there exists a

normally paramagnetic dimer. Ruthenium does not form a doubly bonded

d12 (II,II) or a triply bonded d10 (III,III) dimer. Rather, a d11

(II,III) dimer of formal bond order 2.5 is formed upon reaction of

ruthenium salts with carboxylic acids.48 This complex, of general

formula Ru2(02CR)4X, is quite stable compared to the rhodium (II,III)

species described above. Thus, it can be easily studied using EPR and

magnetic susceptibility. These techniques were applied to Ru2(but)4C1 a

number of years ago by Cotton and Pedersen.36 However, due to

experimental difficulties their results on the electronic and magnetic

nature of the complex were not conclusive. Thanks to improved

technology in EPR and magnetic susceptibility instrumentation, it was

now possible to perform detailed studies on this ruthenium carboxylate

dimer. These studies are the subject of the fifth chapter of this

thesis. In addition, since in contrast to the rhodium and molybdenum

systems the reactivity of the ruthenium dimer towards Lewis bases has

not been widely investigated, some reactivity studies were performed and

are also described in this chapter.



When dissolved in a coordinating solvent, the counter anions bound

to a transition metal cation often dissociate. For example, most first

row transition metal salts dissolve in water to give M(H20)6n+

species.49 In contrast, metal carboxylate dimers do not readily give

M2n+ and RC02". The bridging carboxylate ligands remain coordinated to

give neutral species in solution of general structure as shown in Figure

1-2 where L could be solvent. To further understand the coordination

chemistry of metal-metal bonded systems, it would be desirable to

achieve ligand coordination to a variety of sites such as those

equatorial as well as axial to the metal-metal bond. Furthermore, it is

well known that generation of coordinative unsaturation about a metal

center is crucial for the generation of a catalytic cycle.50 This often

occurs by reversible ligand binding. Thus, it would be desirable to

prepare metal-metal bonded dimers with labile ligands so that they could

be used in catalytic studies and be more effective than analogous

complexes with more strongly bound ligands. These catalytic processes

could involve thermally or photochemically activated ligand

dissociation. In this way any synergistic advantages to using such a

system as opposed to one with monomeric complexes could be determined.

A metal-metal bonded complex analogous to the M(H20)6n+ species has

been reported. Maspero and Taube51 prepared Rh24+(aq) by the reduction

1 rz

of RhCl2+(aq) by Cr2+(aq). This species was identified by conversion to

Rh2(02CCH3)4(H20)2 by addition of sodium acetate. This conversion

from the solvated cationic species does not appear to be completely

reversible. Legdzins and co-workers31,32 treated various metal

carboxylate dimers with strong, non-complexing aqueous acids such as

HBF4(aq). They claim to have generated Rh24+ in this manner. This

claim was subsequently refuted by Wilson and Taube52 who proposed that

the treatment of Rh2(02CCH3)4 with, for example, hot 1 M aqueous CF3SO3

generated Rh2(02CCH3)33+(aq) and Rh2(02CCH3)22+(aq) but no Rh24+(aq).

The formulation of these species was based on UV-visible spectroscopy

and column chromatography. Neither group reported the isolation of any

stable rhodium dimer containing zero, two or three bridging acetates.

The action of strong acids does lead to carboxylate protonation which

allows generation of equatorial coordination sites on the rhodium

dimer. The use of this general method in wholly organic solvent systems

was investigated here. It was hoped that in such solvents more complete

ligand protonation of Rh2(02CR)4 would occur since there would be no

leveling effect from water. Furthermore, the species generated this way

might prove easier to isolate. Very recently, Ford and co-workers53,54

were able to prepare a series of complexes of formula Rh2(02CCH3)3L+,

Rh2(02CCH3)22+, and Rh2L44+ where L = 1,8=naphthyridine or derivatives

thereof such as 2,7-bis(2-pyridyl)-l,8-naphthyridine. These were

prepared by addition of stoichiometric amounts of the ligand and aqueous

1 M HC1 to methanol solutions of rhodium acetate. Similar results using

pyridine are discussed below and were carried out at roughly the same

time. Thus, use of a strongly donating, preferably chelating ligand

does allow isolation of cationic rhodium dimer species. However, these

are not complexes with the type of labile ligand one would desire so

that catalytic activity would result. Isolation of this type of rhodium

dimer has not yet been achieved.

The analogous Mo2(02CR)4 system has proven somewhat more amenable

to study. A very large number of complexes containing the Mo-Mo

quadruple bond are known. Most have bridging carboxylate ligands,

halides or a variety of other anionic groups. Complexes with neutral

ligands are far fewer. Species such as Mo2X4L4 are known where X =

halide or alkyl and L = phosphine such as P(CH3)354-57 Fewer still are

complexes which contain the Mo-Mo quadruple bond coordinated by neutral,

weakly donating ligands. A number of years ago, Bowen and Taube58

reported the Mo24+(aq) species. This was not isolated as solid, but

prepared in solution in the following manner. First, K4Mo2C18 was

reacted with K2S04 in 0.2 M CF3SO3H(aq) to give K4Mo2(S04)4 a stable

salt. This sulfate was reacted with Ba(CF3S03)2 to precipitate BaSO4

and give the red aquo molybdenum dimer in solution. This species was

identified by UV-visible spectroscopy and could be converted back to the

acetate. This work and the corresponding study with rhodium described

above showed that these metal-metal bonded dimers could exist in

solution without the presence of bridging or even anionic ligands. As

was found much later with rhodium,53,54 Bowen and Taube58 were able to

isolate salts of the molybdenum dimer by using strongly donating

chelating neutral ligands. Addition of ethylendiamine (en) and 2,2'-

dipyridyl (dipy) to solutions of Mo2C184- led to isolation

of Mo2(en)4Cl4 and Mo2(dipy)2Cl4 Although these complexes have not

been structurally characterized, the former species presumably has no C1

coordinated to the molybdenum atoms. The Mo24+(aq) species and related

complexes such as Mo2C184- have been used in a photochemical study by

Trogler and co-workers.59 Ultraviolet irradiation of aqueous solutions

of these complexes produced dihydrogen. This reaction was proposed to

proceed via Mo2(v-OH)24+(aq) which was generated directly from Mo24+(aq)

and via Mo2(u-Cl)2Cl4(u-H)3- with the halides. This indicates a

potential application for molydenum dimer species. Analogous complexes

that would be soluble in organic solvents might also show interesting

photochemical behavior.

Another approach towards generating molybdenum dimers with weakly

coordinating ligands was taken by Abbott and co-workers.60 Molybdenum

acetate was reacted with neat CF3SO3H at 100 C. Removal of solvent and

drying at 100 C under vacuum yielded a tan solid formulated as

Mo2(CF3S03)4. This product was frequently contaminated by Mo2(02CCH3)1_3

impurities which were very difficult to remove. Furthermore, the

complex was extremely prone to decomposition making it difficult to use

for subsequent reactions. Very recently, Mayer and Abbott61 achieved

greater success using Mo2(02CH)4 rather than the acetate as starting

material. Molybdenum format was reacted with CF3SO3H and (CF3SO2)20

for six days to yield CO and a tan product formulated as [Mo2(H20)4

(CF3S03)2](CF3SO3)2. This complex reacts with acetronile to yield blue

[Mo2(CH3CN)8](CF3S03)4. These very air and water sensitive complexes
were characterized by IR and UV-visible spectroscopy and elemental

analysis. This represents the first reported example of a quadruply

bonded molybdenum dimer coordinated only by monodentate, weakly

donating, natural ligands. The reaction of Mo2(O2CCH3)4 with CF3SO3H

and other strong nonaqueous acids in acetonitrile and other organic

solvents is described here. These studies were carried out at roughly

the same time as those of Mayer and Abbott61 and gave similar results.

The reactivity of the resulting complexes was also investigated.

Results and Discussion


Treatment of the purple solution of Rh2(but)4 in acetonitrile with

CF3SO3H leads to an immediate, although slight, color change to dark

red. The n-butyrate ligand was chosen for increased solubility; similar

results were obtained with acetate. Trifluoromethanesulfonic acid was

chosen since it is a very strong, poorly coordinating acid that is

somewhat soluble in organic solvents. The effect of different amounts

of CF3SO3H on the UV-visible absorption spectrum of Rh2(but)4 is shown

in Figure 2-1. Rhodium butyrate with no acid showed absorption bands at

552 (e=202) and 438 nm (E=101). These are virtually identical to the

results reported for rhodium acetate in acetronitrile, bands at 552

(e=235) and 437 nm (e=125).62 Single crystal polarized electronic

absorption spectra of rhodium acetate led to a proposal that these bands

correpond to (Rh-Rh)i* + (Rh-Rh)a* and (Rh-Rh)r* + (Rh-O)o* transitions,

respectively.63 Some controversy has recently arisen as to this

assignment and will be discussed in Chapter IV. Addition of two

equivalents of CF3SO3H leads to a very little change in the UV-visible

spectrum. However, addition to four equivalents of acid leads to a

dramatic change. The primarily metal-metal bond trnasition at 552 nm is

relatively unaffected, but the metal-liga~d transition is strongly

affected, shifting to 380 nm. This shift to lower wavelength may result

from a strengthening of the Rh-0 bonds of the remaining butyrates caused

by the higher relative charge on the metal dimer. This would lower the

-1 E

4Nm 4- S-
C -,
o u


'O a

0 I0 0
> CC
*- -o

C) W}

s- C

0. O *-
C '0

o >
4-- *a *

S1C C-


.3 0
/ C C0

.0 VO
U >

a- O- a
t/ C-
C ) 0
o (C-


I f 1




c o

energy of the Rh-O a bonding orbitals and raise the energy of the Rh-

0 a* antibonding orbitals leading to the observed shift to higher

energy. Addition of excess CF3SO3H (-10 equivalents) and allowing the

solution to sit for one hour does not greatly change the spectrum. The

main transitions are observed at roughly the same wavelengths. A large

band extending into the ultraviolet region is observed which may result

from rhodium dimer or solvent decomposition. Solvent decomposition is a

definite problem when trifluoromethanesulfonic acid is used with organic

solvents. For example, it catalyzes the decomposition of THF.64 With

acetonitrile, trimerization probably occurs to give 2,4,6-trimethyl-

1,3,5-triazine. This compound was not isolated, but when benzonitrile

was used as a solvent in the above procedure, the analogous compound

kyaphenine (2,4,6-triphenyl-1,3,5-triazine) was isolated and easily

identified by elemental analysis, melting point and mass spectroscopy.

Kyaphenine is normally synthesized by the addition of excess CC13CO2H to

benzonitrile.65 Thus, the use of excess CF3SO3H should be avoided.

Nevertheless, the UV-visible spectrum suggests that protonation of the

bridging carboxylates is occurring with four equivalents of acid. The

species generated in this manner can be identified by NMR

spectroscopy. NMR data are summarized in Table 2-I. The 13C{1H} NMR

spectra of n-butyric acid and Rh2(but)4, both in acetonitrile-d3, are

shown in Figures 2-2 and 2-3, respectively. Of particular importance is

the signal for the carboxyl carbon which has a very different chemical

shift in the two compounds. Addition to four equivalents of CF3SO3H

leads to a spectrum as shown in Figure 2-4. The four peaks at 139.47,

126.87, 114.85, and 102.46 ppm relative to internal TMS correspond to

the carbon in CF3SO3(H) split by three equivalent fluorine nuclei (19F,

4c 4- 4-'

ccoo o


U CO 10 an
Ic 1 n f 10 )~-

E 4-' 4-' 4-'

a ). C o M z

-0o -0

m) m =3
C 4- 4-

0 o*

U 0 4-c
SU) -- Nc' (U C
C'.* . *- -

.0 -N - '- 5

SC c 0 *- O
-I 0 C U

+ . 1 (

E oco C -- o 7 -
Ci 5On C- C l n
4-' L.

+ L* CO U) C L

.0 1I - C(

C= c cS -C

o U L. -
+ 'r- t. O W 0
C'. CJ '4. > I.-' U S

Sc v- a.-.. 00 i. 0

U 'I Sr E > r

C 4 o - 0
C'. 4-' +-1 cc U 0. -i- i CL 0

0 C a U. L.. .. 0. 0
I- CC cc LJ Cc (0 .0 U 0O

= -0


S.. -
a e

Q) u

E a

o a
4- 4-'

4-' C 'a
'a 1- >

00 0
- c

4- 'a
4- 0 =

u Cm

C 4 II
*-I- 5 -

CL -
0 C4 e

*5- a-
C) -0

n D 4-'

I 4 ,..

L 0-
\E ; r-
= 0)

5S + *i-

o Q.

Q. C
(A 0r

:3 -0
0 C

- 0



. ~ ~|
















s ______

0- -r

U- C)

u o

o a
un a

= CI



>r '

- S--


I a




a 01

4 C.
CL -






------- "_

___________ __^a

3 __?-

0 --------------------

a _______________^



^ _^

s -=^,

I=1/2). The other signals correspond to free and Rh coordinated butyric

acid, with the carboxyl carbon peaks occurring in the expected places,

by analogy with Figures 2-3 and 2-3. Proton NMR spectroscopy gave

similar results. Figure 2-5A shows the 1H NMR spectrum of Rh2(but)4

with the expected splitting patterns for an n-propyl group. Figure 2-5B

shows the spectrum after addition of four equivalents of CF3SO3H.

Signals are observed for coordinated and free butyric acid. The poor

resolution of some of these peaks may be due to the presence of acid,

causing proton exchange and perhaps decomposition. Based on the peak

intensities, the dominant species in solution has an average composition

of Rh2(but)22+. The area ratio of peaks corresponding to free and

coordinated butyrate was 1:1 in several separately prepared solutions.

The ratio of peak areas for the protons for both free and coordinated

butyrate was 2:2:3 for H :HB:HY as expected. This solution gave no EPR

signal at 77 K indicating that a Rh(II) monomer was not present. This

does not prove that the dimer remains intact since disproportionation to

diagmetic Rh(III) and Rh(I) may have occurred. However, the NMR data

suggest that the solution contains the dimer since the signals for

coordinated butyrate, particularly C1, occur near those for Rh2(but)4.

Also Rh(III) and Rh(I) complexes are generally orange or yellow. The

NMR data, in conjunction with the UV-visible results, indicate that four

equivalents of CF3SO3H generate Rh2(but)22+ in acetonitrile solution.

Attempts to isolate a solvated Rh2(but)22+ salt were unsuccessful.

Evaporation of the acetonitrile solution left a dark red, water soluble

oil. Previous workers32,52 reported that evaporation of the solution

obtained by the titration of Rh2(02CCH3)4 with aqueous CF3SO3H led to a

deliquescent green oil which was similarly intractable. Attempts to




0 =
E -

4 -
o =
s- m

41 CD

a re
en o

0- LL
.) C

o0 0
a) *1-

a) 0-



4- L





isolate a product using BPh4 and PFg" as counterions were

unsuccessful. Some solids were obtained, but the results were not

repeatable and the products could not be well characterized.

Another approach that was taken to isolate a Rh2(but)42+ species

involved using a bridging dianionic ligand, Y2-, to form a neutral

compound Rh2(but)2Y. This type of compound has precedent in the A-frame

series of complexes, which have been found to coordinate a wide variety

of molecules to their exposed side.66-69 The sulfide ligand was chosen

since it is used in the A-frame complexes,69 is readily available, has a

high affinity for transition metals, and bridges easily. Anhydrous

sulfide was generated directly in THF by addition of "Super-Hydride"

(LiBH(CH2CH3)2) to elemental sulfur. This solution was added to the

Rh2(but)22+ solution leading to immediate formation of a black

precipitate. The black compound was insoluble in all solvents so it

could only be characterized by elemental analysis. Elemental analysis

indicated a complex with 1-2 sulfurs and two butyrate groups. Mass

spectroscopy was used to little avail. No molecular ion peaks were

detected at m/e=412 for Rh2(but)2S or m/e=446 for Rh2(but)2(SH)2.

Intense peaks corresponding to H2S and HS fragments were observed. The

compound is possibly a sulfide or hydrosulfide bridged rhodium polymer

which contains two butyrates per rhodium dimer. A similarly intractable

compound was prepared using selenide. Rakowski Dubois and co-workers70

successfully converted the molybdenum sulfide polymer [(CSH5)MoSx]y to

the soluble binuclear complex [(C5H5)MoS(SH)232 by stirring the polymer

for 5-7 days under 1 atm of H2. This was attempted with the rhodium

sulfide polymer, but no dissolution was observed. The solid was also

treated with 1-iodoheptane in the hope of alkylating any bridging SH

groups to solubilize the complex. However, no reaction or dissolution

was observed and the solid recovered should no increase in carbon or

hydrogen content. The anion of 1,3-dithiopropane, generated in the same

manner as the sulfide, was used in the hope of obtaining more soluble

products, but gave only an oil. Other Y2- type ligands that could be

considered are cis-1,2-dicyanoethene-1,2-dithiolate (mnt) which forms

complexes with many transition metals71 and (HOPO)202- (pop) which forms

binuclear complexes with platinum.72,73 However, mnt reacts with

Rh2(O2CCH3)4 to give a monomeric Rh(II) complex,71 and would doubtless
do the same with Rh2(02CR)2+. In contrast to the results with

platinum,72,73 rhodium as both RhC13 and Rh2(02CCH3)4 does not appear to

react readily with pop or H3PO3 to give analogous P-bonded dimeric


A final attempt at isolating a cationic rhodium carboxylate dimer

involved the use of pyridine, a Lewis base far stronger than

acetonitrile. Addition of excess pyridine to the acetonitrile solution

of Rh2(but)42+ led to an immediate color change to orange. The UV-

visible absorption spectrum of this solution showed a band at 465 nm

(F=481) presumably corresponding to the (Rh-Rh)r* + (Rh-Rh) a*

transition and a very large band extending into the UV region. This

latter band may involve rhodium to pyridine w* transitions. The order

of addition of pyridine and acid is important. When pyridine (10

equivalents) is added to rhodium butyrate in acetonitrile, the purple

Rh2(but)4(CH3CN)2 solution immediately turns red, indicative of

Rh2(but)4(pyr)2. The equilibrium constants for axial coordination of

various Lewis bases to rhodium butyrate have been determined,23-25 and

Keq for pyridine binding is several orders of magnitude larger than that

for acetonitrile. However, when CF3SO3H (10 equivalents) is added, the

purple color is restored indicating that the pyridine is completely

protonated by the strong acid. This occurs even though some pyridine is

coordinated to the Lewis acid Rh2(but)4. Adding more pyridine

neutralizes the acid present and the red color of the axial pyridine

adduct eventually develops. When excess acid is added to this, as with

any acetonitrile solution of rhodium butyrate, the Rh2(but)42+ species

results and subsequent addition of pyridine leads to the orange color of

what is presumably a pyridine adduct of Rh2(but)4 with equatorial base

coordination. Addition of excess acid to the orange solution restores

the purple color indicating that the equatorially coordinated pyridines

can also be protonated. Attempts were made to isolate solids from the

orange solution by evaporation, cooling, and the use of various solvents

such as water and methanol and various counter anions such as BPh4- and

PF6-. Oils were usually obtained; however a solid was isolated which

analyzed approximately for Rh2(but)2(pyr)4(PF6)2. With Rh2(O2CCH3)4,

less oiling occurred and what is presumably [Rh2(02CCH3)2(pyr)4]

(CF3SO3)2 was isolated. Based on this result and those from Ford's

laboratory,53,54 use of butyrate, while helpful in solution studies, is

not recommended for isolation of solids.

Some other reactivity studies were undertaken on the Rh2(but)42+

solution. This solution showed no visible change upon exposure to air

and the 1H NMR spectrum was unchanged. On the basis of kinetic data,

HRh2(02CCH3)3 has been proposed as an intermediate in the hydrogenation

of olefins catalyzed by rhodium acetate.74 Rhodium acetate itself shows

no reactivity towards H2 (1 atm) at temperatures up to 80 C. It was

hoped that such a hydride species might be observed in the reaction of

H2 with the cationic rhodium dimer solution. This solution was sealed

in an NMR tube under 1 atm of H2, but showed no visible or 1H NMR

spectral change. Furthermore, the solution shows no reaction with one

or two equivalents of 1-hexene or CH302CC=CCO2CH3, both of which might

be expected to add oxidatively to the Rh-Rh bond. Thus, reactivity with

organic molecules has not been enhanced by exposing the metal-metal



In contrast to the work with rhodium described above, it was

possible to isolate stable, cationic acetonitrile coordinated

derivatives of the molybdenum carboxylate dimer. Molybdenum acetate is

completely insoluble in organic solvents, but when suspended in

acetonitrile, addition of two equivalents of CF3SO3H leads to immediate

formation of an intensely colored purple solution. Removal of solvent

and recrystallization of the resulting solid from 1:1 acetonitrile/

toluene allows isolation in good yield of a purple crystalline

complex. Elemental analysis suggests its formulation as

[Mo2(02CCH3)2(CH3CN)4] (CF3S03)2, (1). Use of more acid, up to 10
equivalents, leads to essentially the same product with greater solvent

decomposition. The use of neat acid will be discussed below. This

compound is air sensitive and very hygroscopic but is indefinitely

stable under an inert atmosphere. Various methods were used to confirm

that I is an acetonitrile coordinated Mo-Mo quadruply bonded species as

formulated above. The oxidation state of molybdenum was found to be 2+

using Fe3+ as oxidant using a reported method.58 However, metal-metal

bond cleavage can occur without oxidation state change. Examples

include the photolysis of Re2Cl82- in acetonitrile to give

Re(CH3CN)3C1375 and the reaction of Mo2(02CCH3)4 with t-BuNC to give

Mo(t-BuNC)5(02CCH3)2.76 The conditions required were more strenuous

than those used here. Irradiation at 366 nm for 24-48 hours was needed

for photolysis and in the second case, t-BuNC is a far stronger ligand

than acetonitrile.

The UV-visible absorption spectrum of I is of interest and provides

conclusive evidence that the Mo-Mo quadruple bond remains intact. In

acetonitrile solution bands are observed at 535 (E=864), 390 (e=117) and

255 nm (E=7383). This resembles the results of Bowen and Taube58 who

found for Mo24+(aq) and Mo2(en)44+ absorption bands at 504 (E=337) and

478 nm (e=483), respectively and weaker bands at 370 (E=40) and 360

(E=36.4), respectively. A band at 235 nm (e=966) was also observed for

Mo2(en)42+. Some controversy exists as to the assignment of the
electronic transitions in the Mo-Mo quadruply bonded system. However, a

very detailed single crystal polarized electronic absorption spectrum

study by Martin and co-workers77 indicated that the band observed at 435

nm corresponds to a (Mo-Mo)6 + (Mo-Mo)6* transition. The second band at

377 nm was more tentatively assigned to a (Mo-Mo)6* + (Mo-Mo) r*

transition. A recent study by Manning and Trogler78 of the electronic

spectrum of matrix isolated Mo2(02CCH3)4 confirmed the assignment of the

6+6* transition although suggested that other, probably Mo-0 states,

contribute to the observed band. At any rate, these two transitions are

observed for 1. The UV-visible spectrum of I was also obtained in THF

solution and gave qualitatively the same results. Bands at 490 (E=321),

335 (E=461), and 277 nm (E=3066) were observed. Dissociation of

coordinated acetonitrile probably occurs which changes the absorption

bands. The shift to shorter wavelengths may result from the replacement

of the r-acceptor CH3CN by the a-only ligand THF. Thus i in THF shows

absorptions closest in wavelength and intensity to those of Mo24+(aq)

and Mo2(en)44+. In addition, I in THF is far more air sensitive than 1

in acetonitrile, changing color almost immediately upon air exposure,

perhaps indicating poorer stabilization of the Mo24+ unit.

The IR spectrum of I (Nujol mull) is shown in Figure 2-6. Most of

the absorption bands can be readily assigned. Very sharp bands

corresponding to v(CN) of coordinated acetonitrile are observed at 2300

and 2285 cm-1. This shift to higher frequency, compared to 2266 cm-1

for free acetonitrile, is indicative of end-on nitrile coordination with

little i-backbond stabilization.79 Two v(CN) bands are seen because in

addition to the v(CN) fundamental, there is a combination of the

symmetrical CH3 deformation and the C-C stretch. These two bands are

subject to Fermi resonance coupling which affects their frequencies and

intensities. Unfortunately, no assignment can be made as to Mo-N

stretches. Very few metal organonitrile complex M-N stretches have been

conclusively identified and they usually are weak and of widely varying

frequency.80 Absorption bands corresponding to the acetate ligand are

of interest. For 1 no band corresponding to vasy(CO2) was observed.

This is seen at 1578 cm"1 in Na02CCH3.81 However, a strong band at 685

cm-1 is observed which is most likely 6(C02). This occurs at 675 cm-1

in Mo2(02CCH3)460 and at 646 cm-1 in Na02CCH380 and indicates the

presence of bridging acetate in 1. Finally, a weak band is observed at

410 cm-1 which may correspond to a Mo-Mo stretch. In centrosynm=tric

metal carboxylate dimers this band is IR inactive. However, Raman

spectroscopy studies56 on a number of derivatives of the quadruply

bonded Mo dimer show v(Mo2) occurring at 383 to 404 cm-1 with weak to


r 0






U, r

-? 0



C ,

S" O

m VI

1/1 s


0; 0
l -i


33lOlHIISNyil %

medium intensities. It is possible that non-centrosymmetric isomers of

I are present allowing observation of v(Mo2) in the IR spectrum. Known

molybdenum dimer complexes with this geometry in which the acetates are

cis are Mo2(02CCH3)2((Pz)3BH)282 and Mo2(02CCH3)2(CH3COCHCOCH3)2.83

Infrared studies on these complexes were not reported; however these and

an analogous isomer of 1, all with C2v symmetry, would have an IR active
Mo-Mo stretch. This would most likely be of low intensity due to the

small dipole moment change involved. This possible structure and that

of a centrosymmetric isomer are shown in Figure 2-7. The remaining IR

bands can be assigned to the counterion, non-coordinated CF3S03". Bands

are observed at 1285, 1245, 1160, 1030, 755, 720, 635, 575, and 520

cm-1. The IR and Raman spectra of solid Na03SCF3 have been carefully

analyzed by Miles and co-workers.84 The vibrations they observed and

their assignments are as follows: 1280 (vasy(CH3)), 1232 (vs(CF3)),

1168 (vasy(S03)), 1036 (vs(S03)), 766 (6s(CF3)), 630 (6asy(S03)), and

531 and 515 cm-1 both asy(CF3)). These bands can be directly compared

to those observed for 1. When CF3S03- is coordinated, the IR bands for

v(S03) change greatly. For example, Mo2(03SCF3)4 has S-0 stretches at
1350, 1110, and 990 cm-1.60 The band observed in I at 720 cm-1 cannot

be assigned to the CF3S03- and is probably an acetate or acetonitrile

Proton NMR spectroscopy was performed on 1, but did not provide

much insight into its structure. Signals were observed at 4.3 and 2.0

ppm relative to internal TMS in nitromethane-d3. The upfield signal is

probably coordinated CH3CN since in CD3CN solution it broadened and

decreased in intensity over time, disappearing after about one hour,

indicating exchange with the solvent. The downfield signal is probably




n I
I c

o Z


\ 0
0 0


0 /

CH3C02- although it is rather far downfield for metal coordinated

acetate. Paramagnetic impurities initially present or arising from

complex decomposition would cause line broadening and unusual chemical

shifts.85 However, I does not show an EPR signal in 1:1

acetonitrile/toluene at 77K.

A different synthetic approach was used to study the

interconvertability of the Mo24+ derivatives. What is presumably the
reported60 Mo2(03SCF3)4 complex was prepared but not characterized. To

this was added acetonitrile to give an intensely colored blue

solution. Addition of toluene led to formation of a bright blue

precipitate. This complex did not give a satisfactory elemental

analysis. The IR spectrum indicated coordinated CH3CN, non-coordinated

CF3SO3- and some residual bridging acetate as well as a strong v(OH)

band. Slow evaporation of the filtrate led to formation of purple

crystals. The IR spectrum and elemental analysis of these crystals

coresponded to that of 1. The initially isolated blue complex is most

likely one with less than two acetates giving a more highly charged

species which is less soluble in organic solvents. Subsequent to this

work, Mayer and Abbott61 reported the synthesis of [Mo2(H20)4(03SCF3)2]

(CF3S03)2. This complex was synthesized from Mo2(02CH)4 and thus, in

contrast to the previously reported Mo2(O3SCF3)4, could be reproducibly
prepared free from any carboxylate contamination since the format

decomposes to CO and H20. Addition of acetonitrile to this complex led

to isolation of blue [Mo2(CH3CN)8](CF3S03)2. The blue complex reported

here is most likely impure [Mo2(CH3CN)8](CF3SO3)2, indicating that the

acetonitrile solvate of Mo24+ can also be prepared from molybdenum

acetate, but much less successfully than by the method using molybdenum

format as starting material. What is interesting is that as found with

the rhodium systems, the M2(02CR)22+ species is very stable. Even

following the extreme conditions of Abbott and co-workers,60 a

considerable amount of the Mo2(02CCH3)22+ species is isolated.

Some reactivity studies on I were performed. As stated previously,

the complex is air sensitive and quite hygroscopic as a solid. However,

in acetonitrile solution, the complex is relatively stable. Dioxygen

can be bubbled through this solution for at least 30 minutes without any

visible change. Exposure to air does eventually decompose the dimer.

This decomposition was monitored by UV-visible spectroscopy and is shown

in Figure 2-8. The characteristic Mo-Mo quadruple bond absorption bands

disappear and most likely a variety of monomeric molybdenum species

result. Prolonged exposure to air gives a blue-green solution

characteristic of high oxidation state Mo. It is likely that this

decomposition is assisted by replacement of coordinated acetonitrile by

water. Complex 1 dissolves in water to give a red solution similar to

that of Mo24+(aq). This solution is very air sensitive, as is

Mo24+(aq), in contrast to I in acetonitrile. Complex 1 reacts
immediately with Et4NO2CCH3 in acetonitrile to give a yellow solution

from which yellow crystals of Mo2(02CCH3)4 precipitate. Complex 1

reacts readily in acetonitrile solution with stronger Lewis bases.

Addition of excess (-10 equivalents) of nitrogen donors such as pyridine

or N-methylimidazole gave red solutions and phosphorus donors such as

tricyclohexylphosphine gave blue solutions. These reactions were not

investigated further; however, it is likely that a variety of derivatives

350 400 450 500 550 600

X (nm)

Figure 2-8. UV-visible absorption spectrum of 1 in acetonitrile
(6.0 x 10- M) showing changes on exposure to air.

of the Mo2(02CCH3)22+ unit with various ligands stronger than

acetonitrile could easily be prepared.
The electrochemistry of 1 was Investigated to determine if stable

Mo2(II,III) or other species could be generated. Previous
electrochemical studies by Cotton and Pedersen38 on Mo2CIS4- and
Mo2(but)4 indicated that these complexes could be quasireversibly

oxidized at -0.4 V vs. sce to short-lived Mo2(II,III) species that were

not isolated. The Mo2(but)4+ species was observed by EPR
spectroscopy. Complex 1 in acetonitrile with 0.1 M (n-Bu)4NBF4 as the

supporting electrolyte showed no reversible redox waves over the range

+2.0 to -2.0 V vs. Ag/AgC1,KC1(sat'd). A weak irreversible oxidation

occurred at +1.5 V. Acetonitrile coordination may stablize the dimer

towards oxidation, but does not facilitate isolation of oxidized


Another form of oxidation of the Mo2(II,III) unit could involve
oxidative addition to the Mo-Mo quadruple bond. Such reactions are well

known for carbon-carbon multiply bonded compounds. For example, Br2
oxidatively adds to olefins to give dibromo compounds. Of more

relevance here are the studies by Chisholm and co-workers86-88 who have
achieved oxidative addition to the Mo-Mo triple bond in Mo2(OR)6

complexes. For example, Mo2(i-PrO)6 reacts with (i-PrO)2 to give

Mo2(u-i-PrO)2(i_-PrO)6,86 with various alkynes in the presence of
pyridine to give Mo2(u-i-PrO)2(i-PrO)4(pyr)2(u-C2R)87 and with
dimethylcyanamide to give Mo2(i-PrO)6(u-NCMNe2).88 Such reactions might
be expected for Mo2(02CR)4 complexes if the ilo-Mo bond were more

exposed. Thus, complex 1 is a likely candidate. Unfortunately, no
reaction was observed between 1 and CH3I, (CH3CH2S)2 and 1-hexene, all

of which could oxidatively add across the Mo-Mo quadruple bond. Complex

1 did react readily with dimethylcyanamide (~10 equivalents) in

acetonitrile to give a blue solution. From this a bright blue solid was

isolated. The IR spectrum of this solid showed a very strong v(CN) band

at 2260 cm-1 (Nujol mull) with no bands in the 1600 to 2200 cm-1

range. This can be compared to free (CH3)2NCN, with v(CN) at 2205 cm-1

and to complexes with side-on dimethylcyanamide coordination such as

[Ni(CO)(NCN(CH3)2)]2 with v(CN) at 2008 cm-1 89 and the above Mo(III)

alkoxide dimer with u(CN) at 1582 cm-1.88 The shift to higher frequency

seen here indicates end-on nitrile coordination as seen with

acetonitrile, with no evidence for side-on coordination involving the

Mo-Mo bond. Complex I also showed no reaction with SnC12 or Vaska's

compound (Ir(CO)(PPh3)2C1). These complexes were hoped to add

reductivity to 1 replacing acetonitrile to give clusters containing 4-

coordinate Sn or 6-coordinate Ir, respectively.

A final approach towards investigating the reactivity of I was to

use anionic metal fragments to generate clusters resembling the

reactions attempted above to generate Sn or Ir containing clusters. The

reaction of monomeric metal fragments to form clusters has been widely

studied.90 Of particular use is the reaction of anionic metal complexes

with species containing a weakly bound ligand. An example is the

reaction of Fe5C(CO)142- with W(CO)3(CH3CI)3 to give WFe5C(CO)172-.91

An important point regarding the complexes synthesized in this manner is

that the reactants ar" generally both metal carbonyl complexes or at

least compounds containing metals in similar, low oxidation states with

similar ligands. If 1 would react with these anionic species to form

clusters, the result would be a cluster in which two of the metals, the

molybdenum atoms, would be in a relatively high oxidation state with

relatively electron-withdrawing ligands, carboxylates, while the other

metal would be in a relatively low oxidation state with electron-

donating ligands such as carbonyls.

Complex I was reacted with Mn(CO)5- (C5H5)Mo(CO)3- and Fe(CO)42-

The first two can be easily prepared by reduction of the dimeric species

Mn2(CO)10 and (C5H5)2Mo2(CO)6 by Na/K alloy.92 The iron complex is

available commercially and is often referred to as Collman's

reagent.93 Unfortunately, these reduced species reacted with 1 via

redox reactions. The metal carbonyl starting material was regenerated

and could be identified by IR. Uncharacterized species from

decomposition of the Mo dimer were also produced. Apparently, these

anionic carbonyl complexes are too strongly reducing to form clusters.

This problem often occurs in the reaction of anionic complexes even with

other low oxidation state carbonyl compounds. For example,

(C5H5)Fe(CO)2- and V(CO)6- are not usable in these reactions because

they are such strong reducing agents.90 Furthermore, some clusters are,

like 1, easily reduced. For example, Fe3(C0)12, Fe2Ru(CO)12 and

FeRu2(CO)12 are easily reduced and fragmented.90 Thus, it appears that

the reaction of higher oxidation state metal-metal bonded dimers with

reduced organometallic species is not a facile means of synthesizing

metal custers.

In addition to 1, [Mo2(02CCH3)2(CH3CN)4](CF3SO3)2, the synthesis of

other complexes containing the Mo-Mo quadruple bond was iv-estigated.

One approach would be to use a solvent other than acetonitrile. As

discussed previously, the strong acid needed for carboxylate protonation

precludes the use of many solvents. The problems with THF

(polymerization) and nitriles (oligomerization) have already been

mentioned. Solvents that would solvate the cationic species generated

by carboxylate protonation but be only weakly coordinating are

nitromethane, propylene carbonate and sulfolane (tetrahydrothiophene-

1,1-dioxide). The first two decompose readily upon addition of CF3SO3H;

however sulfolane appears not to decompose. Addition of CF3SO3H (~4

equivalents) to suspensions of Mo2(02CCH3)4 in these three solvents

leads to a faint red color indicative of aquo-coordinated Mo dimer

species. However, the bulk of the molybdenum acetate does not dissolve

and addition of more acid does not lead to more dissolution, only to

solvent decomposition. Clearly, the only reaction occurred because of

the presence of water in these solvents. A reasonably good donor

solvent, such as acetonitrile, is needed to stabilize any cationic

molydenum complexes produced and so drive the carboxylate protonation

reaction to completion.

Another parameter that can be varied besides solvent is the acid

used. It was desired in this work to avoid the use of aqueous solvent

systems since those had been studied previously58 and cationic Mo dimer

complexes were not isolated except with strong ligands such as

ethylenediamine. This solvent choice limits the variety of acids

usable. Furthermore, acids containing halide ions are to be avoided

since the Mo dimer readily coordinates halides. For example,

Mo2(02CCH3)4 reacts with Ph4AsCl in dilute HC1 to give [Mo2(02CCH3)2C14]
(Ph4As)2.83 Other completing acids would produce similar species,

resulting in a Mo dimer coordinatively saturated by strong, anionic

ligands. A non-complexing, nonaqueous acid that is readily available,

besides CF3SO3H, is fluoroboric acid as the diethylether adduct,

C(CH3CH2)20]HBF4. This acid is very difficult to handle since it is

very viscous and hygroscopic. Furthermore, it is difficult to purify

and may be of varying composition, as will be shown below. Addition of

approximately four equivalents of [(Et20)]HBF4 to an acetonitrile

suspension of Mo2(02CCH3)4 leads to formation of an intensely colored

magenta solution. From this solution an air sensitive, hygroscopic

magenta compound can be isolated that is best formulated as

[Mo2(O2CCH3)2(CH3CN)5](BF3OH)2, (2). This complex was characterized in
the same manner as 1. The oxidation state of Mo was found to be 2+.

The UV-visible absorption spectrum of 2 in acetonitrile shows bands at

527 (E=890), 370 (=205), and 269 nm (e=7000). This indicates that the

Mo-Mo quadruple bond is present. Exposure to air leads to

decomposition, as with 1, only it occurs more rapidly with 1. This

process is shown in Figure 2-9.

The IR spectrum of 2 is of interest and supports the above

formulation based on elemental analysis. The spectrum (Nujol mull) is

shown in Figure 2-10. Three strong bands corresponding to v(CN) of

coordinated acetonitrile are observed at 2308, 2282, and 2258 cm1.

Elemental analysis indicated that there were five acetonitriles in 2 as

opposed to four in 1. In 1 there are two v(CN) bands whereas in 2 a

third band results from CH3CN in either a different coordination

environment or from different isomers. Two possible isomeric structures

for 2 are shown in Figure 2-11. As can be seen by comparison with

Figure 2-8, in I the acetonitriles are equivalent while in 2 they are

not. Comparison of the IR absorption bands corresponding to the acetate

demonstrates again the difference between I and 2. Complex 1 showed no

band corresponding to vasy(C02). By contrast, 2 shows bands at 1647,

350 400 450 500 550 600

X (nm)

Figure 2-9. UV-visible absorption spectrum of 2 in
acetonitrile (6.5 x 10-4 M) showing changes
on exposure to air.

33NVlHIASNlai %

1540, and 1500 cm-1. The first is most likely Vasy(CO2) for monodentate

acetate, the latter two for bridging acetate. Comparison with known

compounds with bridging acetate, such as Cr2(02CCH3)4(H20)2 which has

vasy(C02) at 1575 cm-1 94 and those with monodentate acetate, such as

Ru(02CCH3)2(CO)2(PPh3)2 which has vasy(CO2) at 1613 cm-1,95 shows that a

band at this high frequency is characteristic of monodentate acetate.

Very sharp, intense bands are observed at 680 and 685 cm-1 corresponding

to S(C02). If one of the acetates is monodentate this would allow
coordination of an additional acetonitrile as shown in Figure 2-11. It

is possible that the fifth acetonitrile is axially coordinated; however

this site in Mo carboxylate dimers is only weakly coordinating. Even a

strong Lewis base such as pyridine only weakly binds to this position in

Mo2(02CCF3)4,96 which is a stronger Lewis acid than Mo2(02CCH3)4. A
weak, but distinct, band at 405 cm-1 may correspond to the Mo-Mo

stretch. This would be IR allowed in 2 since no centrosymmetric isomers

are possible as can be seen in Figure 2-11. A band is observed at 720

cm-1 corresponding to an acetate or acetonitrile vibration as in 1.

The remaining bands correspond to the counterion and support its

formulation as BF3OH-. A strong band is seen at 1060 cm-1 with weak,

but distinct, bands at 950, 765, 520, 378, and 360 cm-1. Vibrational

absorptions for BF4- are at 1070 (v3, vasy(BF)), 777 (vl, vs(BF)), 533

(v4, 6asy(FBF)), and 360 cm-1 (v2, 5asy(FBF)).97 These same bands for
B(OH)4- are at 945, 754, 533, and 379 cm1.98 All of these modes are

Raman allowed, but only '3 and v4 are IR allowed in these tetrahedral

complexes. The bands observed in 2 at 1060 and 520 cm-1 correspond to

these two IR allowed vibrations. The band at 950 cm-1 may be v3 for 8-0.

The bands at 378 and 360 cm-1 may correspond to v2 for 8-0 and B-F




\ 0

0- -0 0O "

\ o

O --- Z

lo o
ZI u

CM 1
0-0 -

z I-
0 0

I -

bonds, respectively. In BF30H-, a complex with C3v symmetry, all

vibrations are IR allowed so these would be observed. Finally, two

strong bands assigned to v(OH) are seen at 3600 and 3530 cm-1. Thus,

the IR spectrum of 2 supports the formulation of the counterion as

BF30H- presumably resulting from an impurity in the [(Et20)]HBF4 used.

Support for this counterion formulation is also obtained by anion

exchange. Complex 2 can be dissolved in an acetonitrile solution of

excess (n-Bu)4NBF4 or (n-Bu)4PF6 and addition of toluene leads to

precipitation of primarily the BF4- or PF6- salt. This process can be

repeated to effect complete exchange.

The 1H NMR spectrum of 2 resembles that of I with signals observed

at 3.0 and 2.1 ppm in CD3CN. Thus, NMR does not distinguish between

different types of acetonitrile coordination.

Due to the more difficult synthetic procedure for 2 compared to 1,

as well as more uncertainty as to the exact structure of 2, reactivity

studies were not performed.

Interestingly, a complex analogous to 2 can be obtained using

CF3SO3H. After recrystallization of 1, the filtrate is often magenta

rather than purple. Addition of a small amount of toluene and allowing

the solution to sit overnight leads to formation of a. crystalline

magenta precipitate, (3). The amount of 3 varies greatly from one

preparation of i to the next. It is not clear as to the procedure for

selectively preparing one or the other, although use of freshly

distilled CF3SO3H leads to better yields of I over 3. A formula that

can be proposed for 3 is [Mo2(02CCH4)2(CH3CN)4X](CF3S03)2 where X=CH3CN

or H20. The oxidation state of Mo in 3 is 2+. The elemental analysis

of 3 favors X=H20. This is also supported by the fact that 3 is more

33NllIMSNVai %

- t

likely to be obtained with less pure, presumably water contaminated

CF3SO3H. However, the IR spectrum of 3, shown in Figure 2-12 (Nujol

mull), has no band corresponding to v(OH). A shoulder on the Nujol band

at 3250 cm-1 might be from this vibration. The bands assignable to

v(CN) at 2310, 2285, and 2255 cm-1 are virtually identical in frequency

and intensity pattern to those observed for 2. Furthermore, the bands

assignable to the CO2 vibrations are similar for 2 and 3. A weak band

is seen at 1640 with stronger bands at 1530 and 1508 cm-1. The first

can be assigned to vasy(C02) for monodentate acetate, the latter two to
bridging acetate. Well resolved bands at 680 and 690 cm-1 correspond to

6(C02). Strong, well resolved bands corresponding to all the vibrations
of non-coordinated CF3S03" are observed at 1280, 1230, 1150, 1030, 755,

635, 575, and 515 cm-1. The assignment of these bands has been

discussed previously and are the same as those found in 1. Without a

structure determination by single crystal x-ray diffraction, the

differences between complexes 1, 2, and 3 cannot be definitively

determined. Assuming that 2 and 3 contain monodentate and I bidentate

acetate, it is remarkable that these two types of carboxylate

coordination lead to such different colors. The exact orientation of

the monodentate acetate might give some clue to this. It is clear that

different anions do not lead to different properties.

In addition to the Mo2(O2CR)4 system, the Mo2(S2CR)4 system was
investigated. The facile synthesis of Mo2(S2CCH3)4 has been reported.99

Unfortunately, it shows no reaction with two to four equivalents of

CF3SO3H in acetonitrile. Overnight stirring of Mo2(S2CCH3)4 in neat

CF3SO3H leads to recovery of the starting material along with a small

amount of decomposition products. The CH3CS2 species binds very

strongly to Mo and is not readily protonated. Mo2(S2CR)4 complexes

could be used in calorimetric studies of Lewis base binding for

comparison with RCO2- complexes. Mo2(02CCH3)4 is soluble in THF and a

complex such as Mo2(S2CCH2CH2CH3)4 might be soluble in more poorly

coordinating solvents suitable for use in calorimetric work.


The addition of stoichiometric amounts of strong, non-complexing

acids to metal-metal bonded carboxylate dimers leads to protonation of

the bridging carboxylate and generation in solution of M2(02CR)22+

species. Spectroscopic evidence confirms that the metal-metal bond

remains intact and that two carboxylates are retained. The choice of

solvent is crucial since it must stabilize the resulting coordinatively

unsaturated cationic complex, but withstand the strong acid.

Acetonitrile fits these requirements and several acetonitrile

coordinated complexes of the molybdenum dimer are reported here and

elsewhere.61 With rhodium it was not possible to isolate such a complex

as was previously found by workers using aqueous solvents.32,52 Using

strong donors such as pyridine as described here, and related ligands as

reported elsewhere,53,54 it is possible to isolate a cationic rhodium

carboxylate dimer. However, these ligands may not be sufficiently

labile for subsequent reactivity studies on the rhodium system. It may

be that even acetonitrile coordinates too strongly to the Mo dimer,

since the complex reported here does not show reactivity towards

oxidative addition in contrast to various organometallic metal-metal

bonded complexes. Another interesting possibility is that only

organometallic dimers, containing relatively electron donating ligands

and with metals in a low oxidation state, can undergo these reactions

which resemble those found in organic chemistry. The metal carboxylate

dimer with acetonitrile coordination differs from organometallic

complexes and undergoes reactions such as Lewis base coordination and

ligand substitution resembling those found in classical coordination

chemistry. Nevertheless, the photochemical and photophysical properties

of the complexes described here may be of interest. Analogous systems

studied by Gray and co-workers such as Mo24+(aq)59, the metal-metal

bonded diphosphite bridged Pt(II)/(III) dimers72 and the non-metal-metal

bonded isonitrile bridged Rh(I) dimers100 have shown interesting

photochemical behavior. Furthermore, the ligand substitution reactions

of the metal-metal bonded complexes described here which contain

accessible equatorial sites could be investigated in a quantitative

manner as was previously done for systems containing only axial

coordination sites.

Experimental Section

Operations were carried out under nitrogen using Schlenk techniques

or an inert atmosphere box except as otherwise noted. Solvents were

distilled before use. Trifluoromethanesulfonic acid was distilled under

reduced pressure. Tetrafluoroboric acid diethyletherate (Pfaltz and

Bauer) was used without further purification. Rhodium acetate was

synthesized from RhCI3(H20)3 by literature methods.101


{h2(02CCH3)4 (0.5 g, 1.1 mmol) was refluxed for 6 h in n-butyric

acid (14 mL) and n-butyric anhydride (1 mL). The solution was

concentrated to 3 mL and cooled at -20 C overnight. The resulting crude

Rh2(but)4 was recrystallized from hot toluene, washed with cold hexane

and dried over P205 overnight to yield 0.5 g (0.9 mmol, 80%). Anal.

Calcd. for Rh2C16H2808: C, 34.68; H, 5.09. Found: C, 34.74; H, 4.99.

Rh7(but)?2+ Solution

Rh2(but)4 (0.328 g, 0.59 mmol) was dissolved in CH3CN (5.00 mL) to

give a purple solution. To this was added CF3SO3H (0.21 mL, 2.37 mmol)

leading to an immediate slight color change towards dark red. Similar

solutions using CD3CN were used for the NMR work.

Sulfide Complex

Elemental sulfur (0.0236 g, 0.74 mmol) was suspended in THF (1

mL). To this Super-Hydride (LiBH(CH2CH3)3, 1.5 mL, 1 M in THF, Aldrich)

was added dropwise. Gas evolution was vigorous and a pale yellow

solution resulted. This solution was added to the above Rh2(but)22+

solution (2 mL, 0.092 M in rhodium dimer). A black precipitate formed

immediately. Filtration, washing with THF and drying under vacuum at

100 C afforded 0.8 g of a black, completely insoluble solid. Anal.

Calcd. for Rh2(02CCH2CH2CH3)2S: C, 23.32; H, 3.43; S, 7.78; C:H,

6.80. Found: C, 24.39; H, 3.70; S, 12.53; C:H, 6.81. The high sulfur

analysis results from SH units and bridging polysulfide. The selenium

compound was prepared in the same manner and gave an even less

satisfactory elemental analysis.

Pyridine Complex

To the above Rh2(but)22+ solution (3 mL, 0.03 M in rhodium dimer)

was added pyridine (0.16 mL, 2.0 mmol). An orange color immediately

resulted. Attempts to obtain a solid by cooling and evaporation yielded

only an oil. Addition of NH4PF6 (0.16 g, 1.0 mmol) dissolved in water

(1 mL) and subsequent evaporation and cooling led to formation of an

orange-red precipitate. This procedure was not always repeatable; oils
often resulted. Furthermore, IR indicated the presence of CF3S03- as

well as PF"6. Anal. Calcd. for [Rh2(02CCH2CH2CH3)2(C5H5N)4](PF6)2: C,
34.09; H, 3.47; N, 5.68. Found: C, 33.99; H, 3.87; N, 6.18. The above

procedure using Rh2(02CCH3)4 allowed isolation of an orange solid

without addition of PF6-. Anal. Calcd. for [Rh2(O2CCH3)2(C5H5N)4]

(CF3S03)2: C, 33.27; H, 2.79; N, 5.97. Found: C, 32.73; H, 2.98; N,


This complex was synthesized following the procedure of Martin and
co-workers77 which gives much higher yields than the original

method.102 Mo(CO)6 (1 g, 3.8 mmol) was added to o-dichlorobenzene (30

mL). Acetic acid (8 mL) and acetic anhydride (1 mL) were added and the

solution refluxed overnight during which time the solution turned almost

black. The heating was stopped and the solution allowed to cool without

removal of the heating mantle for 8 h. Filtration and washing with

ethanol followed by diethylether led to isolation of beautiful yellow

needle crystals of Mo2(02CCH3)4 (0.65 g, 1.5 mmol, 80%). Anal. Calcd.

for Mo2C8H1208: C, 22.45; H, 2.83. Found: C, 22.45; H. 2.90.

Molybdenum acetate should be used as soon as possible since it

decomposes even under inert atmosphere or vacuum over a period of days
to green and eventually black products.

Bis(trifluoromethylsulfonate), (U)

Mo2(02CCH3)4 (0.40 g, 0.93 mnol) was suspended in acetonitrile (4
mL). It is important that the acetonitrile be degassed using freeze-

pump-thaw cycles with the final vacuum broken by nitrogen, otherwise

decomposition of molybdenum acetate often occurs giving a brown

solution. To this was added CF3SO3H (0.17 mL, 1.92 mmol). An intensely

colored purple solution formed immediately and was stirred for 10 min.

Removal of solvent by pumping left a dark purple solid which was

dissolved in a minimum amount of acetonitrile (~2 mL) and filtered.

Addition of toluene (~3 mL) led to formation of a purple precipitate

after 1 h. The solid was recrystallized from 1:1 acetonitrile/toluene

and washed with toluene followed by hexane to yield 0.5 g. Anal. Calcd.

for [Mo2(02CCH3)2(CH3CN)4](CF3S03)2: C, 21.77; H, 2.35; N, 7.25; S,

8.30; F, 14.76; Mo, 24.84; 0, 20.72. Found: C, 21.83; H, 2.38; N,

7.56; S, 8.28; F, 15.10; Mo, 24.00; 0 (by diff.), 20.85. From the

filtrate obtained in the above recrystallization a magenta, rather than

a purple, solution is often obtained. Addition of toluene (~1 mL) to

this leads to formation of a magenta precipitate, 3. Anal. Calcd. for

[Mo2(02CCH3)2(CH3CN)4(H20)](CF3S03)2: C, 21.27; H, 2.55; N, 7.09; S,
8.11; F, 14.42; Mo, 24.48; 0, 22.27. Found: C, 22.00; H, 2.65; N,

7.06; S, 8.08; F, 12.7; Mo, 22.59; 0 (by diff.), 24.92.

Bis(trifluorohydroxyborate), (2)

Mo2(02CCH3)4 (0.71 g, 1.66 mmol) was suspended in acetonitrile as
above. To this was added (Et20).HBF4 (0.8 mL, approx. 6 mmol). An

intensely colored magenta solution immediately resulted. Removal of

solvent by pumping left a magenta solid which was dissolved in

acetonitrile (~4 mL) and filtered. A small amount of yellow needle

crystals of unreacted Mo2(02CCH3)4 remained. When less acid is used,

more unreacted molybdenum acetate is recovered. To the filtrate was

added diethylether (5 mL) which caused rapid formation of a magenta

precipitate. The compound was recrystallized from 1:1

acetonitrile/toluene and washed with toluene followed by hexane to yield

0.9 g. Anal. Calcd. for [Mo2(02CCH3)2(CH2CN)5](BF30H)2: C, 24.55; H,

3.38; N, 10.23; F, 16.65; Mo. 28.02. Found: C, 23.79; H, 3.32; N,

10.02; F, 17.03; Mo, 26.36.

Anion Exchange

Complex 2 (0.3 g, 0.44 mmol) and (n-Bu)4NBF4 (1 g, 3.0 mmol) were

dissolved in acetonitrile (5 mL). To this was added toluene (5 mL)

leading to formation of a magenta precipitate. After two cycles of this

procedure, IR of the magenta precipitate showed a greatly diminished

v(OH) band and the other bands unchanged. Anal. Calcd. for

[Mo2(O2CCH3)2(CH2CN)5](BF4)2: C, 24.41; H, 3.07; N, 10.17; F, 22.06;
Mo, 27.86. Found: C, 24.49; H, 3.29; N, 11.77; F, 19.16; Mo, 38.13.

Molybdenum Trifluoromethylsulfate Complex

To Mo2(02CCH3)4 (0.2 g, 0.47 mmol) was added to CF3SO3H (10 mL).

The suspension was heated at 100 C with stirring for 1 h, by which time

all the solid dissolved. The acid was removed by pumping leaving a red

solid which presumably corresponds to the Mo2(03SCF3)4(CF3SO3H) complex

described by Abbott and co-workers.60 Further pumping with heating at

160 C led to formation of a tan solid which is presumably the

Mo2(03SCF3)4 complex.60 These intermediates were not isolated or
characterized. Addition of acetonitrile (10 mL) to the tan solid led to

formation of a bright blue solution. Addition of toluene (-7 mL) caused

formation of a blue precipitate. Elemental analysis of this compound

was not satisfactory, although it appeared to be an acetonitrile

coordinated Mo(II) dimer. IR spectroscopy indicated strong v(CN) and

v(OH) bands as well as bands corresponding to non-coordinated CF3SOg".

Bands corresponding to residual acetate were observed at 1615 cm-1

(vasy(C02)) and 675 cm-1 (6(C02)). A band at 415 cm-1 may be v(Mo2).

Slow evaporation of the filtrate obtained above resulted in formation of

purple cyrstals of what is most likely Complex 1. The IR

spectrumcorresponded to I as did the elemental analysis although the

precipitate may have been contaminated with species such as

[Mo2(02CCH3)(CH3CN)x](CF3S03)3 giving higher %S, %F, and %0. Anal.
Calcd. for [Mo2(02CCH3)2(CH3CN)4](CF3SO3)2: See above. Found: C,

21.33; H, 2.10; N, 6.23; S, 9.48; F, 15.48; Mo, 21.47; 0 (by diff.),



This complex was synthesized following the procedure of Cotton and

co-workers.99 CS2 (0.77 mL, 0.013 mol) was added to CH3MgBr (5.63 mmol,

as THF solution, Aldrich) in THF (10 mL). A pale yellow solution

resulted which was stirred for 45 min. To this was added Mo2(02CCH3)4

(0.60 g, 1.4 mmol). A dark red-brown solution formed immediately.

After stirring 15 min, methanol (20 mL, N2 purged) was added leading to

formation of an orange-red precipitate. Filtration and washing with

methanol afforded 0.44 g (56%). This complex, in contrast to

Mo2(02CCH3)4, is indefinitely stable and can be recrystallized in air
from THF. Anal. Calcd. for Mo2C8H12S8: C, 17.26, H, 2.17; S, 46.09;

Mo, 34.48. Found: C, 17.46; H, 2.43; S, 45.93; Mo (by diff.), 34.18.

Experimental Methods

Elemental analyses were performed by the Microanalytical Laboratory

of the University of Illinois, Urbana, IL, or by Galbraith Laboratories,

Knoxville, TN. Ultraviolet-visible spectra were recorded on a Cary 14

spectrometer using matched quartz 1.0 cm cells. Infrared spectra were

recorded on a Perkin-Elmer 5998 instrument using KBr cells. Fourier

transform 13C{1H} NMR spectra were recorded on a Varian Associates XL-

100 FT spectrometer operating at 25.2 MHz. The 13C chemical shifts were

measured with respect to the nitrile carbon of CD3CN (118.2 ppm relative

to TMS). Proton NMR spectra were recorded using a Varian HR-220 NMR

spectrometer equipped with a Nicolet Technology Corp. TT-220 Fourier

transform accessory. Precision-grade tubes were used for the 220 MHz

spectra so as to reduce spinning sidebands. The 1H chemical shifts were

measured with respect to internal TMS.



As discussed in the previous chapter, it is possible to effect

removal of the removal of the bridging carboxylate ligands in metal

carboxylate dimers by reaction with stoichiometric amounts of strong

non-complexing acids. This reaction allows Lewis bases and potentially,

substrates for catalytic processes, to coordinate to equatorial as well

as axial sites on the metal carboxylate dimer. An alternative approach

to achieving this type of coordination is to use a carboxylate ligand

with an electron-withdrawing group. This type of carboxylate would

donate less electron density to the metal dimer subunit rendering the

carboxylates more prone to displacement and the metals more susceptible

to attack by Lewis bases. The effect of an electron-withdrawing

carboxylate ligand, CF3CF2CF2CnO- (hfb), has been quantitatively shown

in earlier studies by Drago and co-workers.25 The erthalpies of axial

Lewis base adduct formation by Rh2(hfb)4 versus Rh2(but)4 were

examined. The hfb ligand greatly enhanced the Lewis acidity of the

rhodium system towards electrostatic interactions and increased the

acidity towards covalent interactions by almost as much. In addition to

this greater reactivity towards Lewis bases, the metal fluorocarboxylate

dimers have much greater solubility in non-coordinating organic solvents

than do the corresponding alkylcarboxylate systems. This facilitates

study of their solution chemistry. For example, since io"2(02CCH3)a is

completely insoluble in organic solvents and Rh2(02CCH3)4 only sparingly

so, a direct comparison of their solution properties is impossible. Use

of the trifluoroacetate ligand makes such a study possible. The

interest in a comparison of the solution chemistry of the rhodium and

molybdenum carboxylate dimers stems from the large difference in their

metal-metal interactions. As discussed earlier, the d8 Mo system has a

strong, short, relatively unpolarizable quadruple bond. The d14 Rh

system has a longer, weaker, more polarizable single bond. Furthermore,

since rhodium is more electronegative than molybdenum, the Rh-Rh

molecular orbitals are overall lower in energy than the Mo-Mo bond

orbitals. The result of all this is that in the rhodium dimer, the

frontier MO's are the 7* HOMO and the o* LUMO while for molybdenum those

metal-metal orbitals are vacant and high in energy while the 6 HOMO and

the 6* LUMO are of main importance. This implies that the covalent

interaction with axial bases should be strong for the rhodium

carboxylate dimer and much less for molybdenum. This has been

quantitatively confirmed by Drago and co-workers25 in a comparison of

the enthalpies of axial Lewis base adduct formation by Rh2(hfb)4 versus

Mo2(hfb)4. In addition, a r-backbonding interaction was observed

between the filled 7* orbitals on the rhodium dimer and vacant r*

orbitals on bases such as pyridine and acetonitrile. This interaction

was not seen for the molybdenum dimer as expected from the MO scheme

described above.

Another implication of the MO scheme is that given the opportunity,

Lewis bases should coordinate more readily to equatorial than to axial

sites on molybdenum, while this would be less likely for rhodium. This

type of reactivity has indeed been found with Mo when the

trifluoroacetate dimer is used, since this ligand allows access to the

equatorial sites. Girolami and co-workers33,103 synthesized and

characterized a number of adducts of molybdenum trifluoroacetate with

phosphines and other Lewis bases. They found that Mo2(02CCF3)4 not only

formed adducts in which there was coordination along the Mo-Mo axis, but

also some in which there was coordination in sites perpendicular to the

Mo-Mo axis. All of these complexes were of general formula

Mo2(02CCF3)4L2. Those with axial coordination were called Class I
adducts, those with equatorial coordination, Class II. Only Lewis bases

that are sterically small and good o-donors were reported to give

isolable equatorial adducts. Examples are trimethylphosphine (PMe3),

triethylphosphine (PEt3), and dimethylphenylphosphine (PMe2Ph).

Andersen estimated steric bulk by cone angle and o-donor strength by

v(CO) values as described by Tolman.104 The assignment of complexes

into the two classes was made on the basis of 19F and 31P NMR

spectroscopy which showed different signals resulting from phosphines in

different coordination sites. Infrared spectroscopy also showed

different CO2 stretches for the two types of CF3CO2" ligands. However

some controversy exists over the assignment of IR and NMR peaks for

these complexes. Cotton and Lay105 also prepared phosphine complexes of

Mo2(O2CCF3)4 and obtained spectra at variance with those of Girolami and
Andersen and co-workers.33,103 In addition, these two groups reported

different structures for the complex Mo2(02CCF3)4(PMePh2)2. Cotton and
Lay105 obtained a Class II (equatorial) adduct and Girolami and

Andersen103 a Class I (axial) adduct. Slight variations in synthetic

procedure led to this difference since PMePh2 is a phosphine inter-

mediate on the size and donor strength scales.

Other solution studies106,107 have been performed on Mo2(02CCF3)4
as well as a number of crystallographic studies.96,103,104,108 In

contrast, the analogous rhodium system has not been as extensively
investigated, particularly in solution.5,6 Crystal structures have been

determined for Rh2(02CCF3)4L2 where L=(CH3)2S02,109 (CD3)2SO,110

PPh3,111 P(OPh)3,111 CH3CH2OH,112 H20113, 2,2,6,6-tetramethylpiper-
idinolyl-1-oxyll3 and (CH3)2S0.114 In all these cases, as in those with

alkylcarboxylates, only Class I (axial) adducts were formed. However, a

systematic study of the Lewis base reactivitiy of Rh2(02CCF3)4 had not

been performed. For the reasons discussed above, that is ligand effects
and metal-metal bond effects, such a study was performed and is
described below. Furthermore, it was hoped that this study would shed

some light on the discrepancies in the interpretation of the
spectroscopy properties of the molybdenum systems described above.

Results and Discussion

The 19F NMR spectrum of Rh2(02CCF3)4 was obtained in both

nitromethane-d3 and toluene-d All of the 19F NMR data are summarized
in Table 3-1. A sharp singlet was found in both room and low

temperatures in both solvents which corresponds to the CF3 groups on the
four equivalent bridging trifluoroacetates. Nitromethane and toluene

are very weak bases and thus should coordinate weakly, if at all, to the
rhodium dimer. What is significant is that these signals occurred in

the -73 to -75 ppm range (relative to internal CFC13). The signals were
somewhat solvent and temperature dependent. Earlier workers33,106,107
have assigned peaks in the -72 to -74 ppm range to monodentate CF3CO2
and peaks at -70 ppm to bidentate CF3C02" in Mo2(02CCF3)4 complexes.

The IR spectrum of Rh2(02CCF3)4 shows a single vasy(CO2) band in
solution and in the solid state. All the major IR data are summarized

in Table 3-11 and the IR spectrum of Rh2(O2CCF3)4 is shown in Figure 3-

1. However, this stretch occurs at a higher frequency (1650 to 1670

cm-1) than vasy(C02) for bidentate CF3CO2" in Mo2(02CCF3)4 complexes

(-1600 cm-1). Thus, there is no direct correspondence between the

location of the 19F NMR and IR signals for the rhodium and molybdenum

systems. Nevertheless, mono- and bidentate CF3CO2 give significantly

different spectra in the rhodium complexes as will be shown below.

Oxygen Donors

Emerald green Rh2(O2CCF3)4 forms blue 2:1 complexes with oxygen

donor bases such as tetrahydrofuran (THF), dimethylsulfoxide (DMSO),

N,N-dimethylformamide (DMF), and trimethylphosphine oxide (OPMe3). The

THF adduct is quite stable but heating at 100 C under vacuum effects

quantitative removal of THF to give base-free starting material.

The 19F NMR and IR spectra are characteristic of a Class I adduct. A

singlet is observed in the 19F NMR spectrum at -75 ppm and vasy(C02)

occurs at about 1660 cm-1 in both the solid adduct and in solution. The

IR spectrum of Rh2(02CCF3)4(THF)2 is shown in Figure 3-2. These results

are similar to those for the free Lewis acid, Rh2(02CCF3)4.

The crystal structure of Rh2(02CCF3)4(DMSO)2 showed a Class I
adduct, with 0-bonded DMSO.109 There was nothing to indicate otherwise

in solution since a single peak was observed in the 19F NMR spectrum.

Both axial S-coordination and equatorial 0- or S-bonding would most

likely lead to additional signals. The equivalence of the solution and

solid state structures was confirmed by IR, which showed a single

Vasy(C02) band at 1662 cm-1 (Nujol mull) and at 1655 cm-1 (CHC13

Table 3-1. 19F NMR Data for Rh2(02CCF2)4 Complexesa










19F Chemical Shiftb









-74.69, -75.44

-74.08, -74.87

-74.73, -75.34

-73.0, -73.2

-73.4, -73.8

-74.2, -74.6

-73.98, -74.90

-74.41, -74.75








Solvent Temperature(oC)


tol uene-d

tol uene-d.









tol uene-d8




Table 3-I. continued

1F Chemical Shiftb


Solvent Temperature(oC)




-72,73, -74.51

-75.42, -75.92


-74.68, -74.83

(-75.10, -75.45


-73.40, -73.88

-74.65, -75.25

1:1:2:1 CDC13


2.5:1 CDC13

3:2 CDC13

3:2 CDC13

a. Data are reported for those complexes that were obtainable as
and the P(OMe)3 complex. Solution studies were undertaken on
and are described in the text.

genuine adducts
other systems

b. All chemical shifts are with respect to internal CFC13.

c. These signals are of very low intensity and may indicate the beginning of
complex decomposition.


Table 3-11. Infrared Data for Rh2(02CCF3)4 Complexes




asy(C02) (cm-1)a
CHC13 Solution

1670 s, 1660 s
1705 s, 1642 m
1693 s br, 1658 s
1715 w, 1654 m, 1652 m
1715 w, 1660 m
1658 m, 1648 w

Nujol mull



s, 1650 s
s, 1655 s
m br, 1660 m br

1663 m, 1653 w

a. s = strong, m = medium, w = weak, br = broad
b. includes amide v(CO), resolved in solution
c. included to contrast with Rh2(O2CCF3)4(PPh3)2













14 U



33NV111HSW"1 h











solution) in agreement with the assignment as an axial adduct. The IR

spectrum of Rh2(02CCF3)4(DMSO)2 is shown in Figure 3-3. The DMSO bands

are also of interest. In the solid state, two doublets were observed at

1005 (s) and 995 cm-1 (s) and at 945 (s) and 937 cm-1 (s). In CHC13

solution these occurred at 1020 (m) and 1000 cm-1 (m) and at 948 (m) and

931 cm-1 (w). Free DMSO in CHC13 solution has v(SO) at 1055 cm-1 and a

6(CH3) band at 946 cm-1. Oxygen coordination lowers v(SO) and sulfur

coordination raises the frequency of this band. No bands were observed

for Rh2(02CCF3)4 in the 1050 to 1150 cm-1 region where S-coordinated

DMSO would obsorb. It should be noted that Rh2(02CCH3)4 binds DMSO via

the sulfur atom (v(SO) at ~1090 cm-1),62'115 further evidence of ligand

effects on Lewis acidity of the rhodium carboxylate dimer.115 Cotton

and Felthouse114 have reported bands for this complex at 943 and 939

cm-1 (Nujol mull) which they assigned to v(SO) of 0-coordinated DMSO,

while the higher frequency doublet was not mentioned. Their assignment

was based on earlier work by Cotton and co-workers,116 who proposed that

the band at ~950 cm-1 in DMSO complexes corresponds to v(SO) while that
at ~1000 cm-1 to 6(CH3). However, Drago and Meek117 reversed this

assignment since the band at ~1000 cm-1 is more sensitive to the type of

metal coordinated. The IR spectrum of Rh2(02CCF3)4(DMSO-6 )2 was
obtained here in the hope of clarifying the assignment of these two

bands. The IR spectrum of this deuterated complex was identical to the

original complex with respect to the bands related to CF3C02-.

Unfortunately, in the area of interest, there was also little change.

Fairly strong, broad bands were seen at 1020 and 950 cm-1, with the

latter more intense. A new band occurred at 825 cm-1 which may


correspond to a 6(CD3) band at 811 cm-1 in free DMSO-d. Thus, a

definitive assignment of these two bands cannot be made.

When DMF was added to Rh2(02CCF3)4, a purple color initially
appeared, but the solution then rapidly turned blue. The 19F NMR

spectrum of this complex showed a single peak at -74 ppm at room and low

temperature. The IR spectrum was also characteristic of a Class I

adduct (vasy(C02) at 1662 cm'1 in CHC13 solution). In solution the

amide v(CO) could be resolved from the vasy(C02) band (this was not

possible in the solid) and was shifted from 1685 cm-1 in free DMF (CC14
solution) to 1643 cm-1, indicative of 0-coordination.118 The IR

spectrum of Rh2(02CCF3)2(DMF)2 is shown in Figure 3-4. Addition of
excess DM1F did not change the 19F NMR spectrum. Rh2(02CCF3)4 was

indefinitely stable in excess DMF. This is in contrast to DMSO.

Kitchens and Bear119 reported that addition of excess DMSO to

Rh2(02CCF3)4 led to formation of a yellow decomposition product, which

was also observed here. This is probably due to eventual sulfur


The reaction of Rh2(02CCF3)4 with 10 equivalents of OPMe3 in 1:1

toluene/dichloromethane afforded a blue solid contaminated with

crystalline, white, unreacted OPMe3. The IR spectrum of this blue solid

(Nujol mull) showed a strong, broad vasy(CO2) band at 1650 cm-1 and a
band assignable to 6(C02) at 725 cm-1. A very strong, broad band at
-1200 cm-1 included vasy(CF3) and v(PO). Thus, a Class I adduct is most

likely formed as with the above oxygen donors. This is to be expected
since using the E and C parameters,27-29 OPMe3 is a Lewis base roughly

comparable to DMSO. Due to limited availability of OPMe3, further

studies were not performed.







0.. 0


0- 0
Cr r

o u

r 0

1- 3


0 -

33NVlllH NSUy *I.

Nitrogen Donors

A variety of nitrogen donor bases were reacted with Rh2(02CCF3)4

with varying results. Isolable analytically pure complexes could not be

obtained with piperidine and N-methylimidazole (N-MeIm). When these

bases were added to toluene solutions of Rh2(02CCF3)4, a red color

immediately resulted indicative of nitrogen base coordination. However,

after several hours the solutions turned yellow and evaporation gave an

intractable yellow tar in both cases. This indicates dimer

decomposition forming Rh(I) and/or Rh(III) as was described with DMSO.

Some 19F NMR studies were performed on solutions of Rh2(02CCF3)4 with

these bases. A freshly prepared solution containing 10:1 N-

MeIm/Rh2(02CCF3)4 showed single peaks at -74.6 ppm at 27 C and at -74.1

ppm at -61 C in CDC13. Thus, a 2:1 Class I adduct was initially present

and remained for a few hours. The spectrum became complex as the yellow

color appeared. A freshly prepared solution of 10:1

piperldine/Rh2(02CCF3)4 showed a major signal at -74.2 ppm and smaller

peaks at -68.4 and -80.3 ppm indicating that rapid decomposition

occurred. The major peak is presumably from axially coordinated dimer,

the other two from decomposition products. It might be possible to

prepare adducts with these two bases if only stoichiometric amounts were

used. By contrast, use of excess triethylamine led to facile isolation

of Rh2(O2CCF3)4(Et3N)2. This complex had a singlet in the 19F NMR

spectrum, even with excess base, at -75 ppm at both room and low

temperatures. This complex had an IR spectrum characteristic of

bidentate CF3CO2- in solution and in the solid state. The latter is

shown in Figure 3-5.


2 a8



4 -



^^"^ '



- ^



With pyridine an interesting product was formed that is

intermediate between the Class I adduct formed with Et3N and the

decomposition products formed with piperidine and N-MeIm. This complex

is a stable red 4:1 adduct containing both axially (Class I) and

equatorially (Class II) coordinated Lewis base and thus may be
considered a new type of complex which will be called Class III. These

three structures are shown in Figure 3-6. Only one of six isomers of

Class II and III adducts are depicted. This behavior can be contrasted

with the reaction of pyridine with Mo2(02CCF3)496 and Rh2(O2CCH3)4120 in

which Class I adducts form. A precedent for this Class III compound

exists. Webb and Dong106 have performed solution studies on

Mo2(O2CCF3)4 with varying amounts of pyridine and found 19F NMR signals
in the places predicted33 for mono- and bidentate CF3CO2- (-70.5, -75.3

ppm, respectively) and IR adsorption bands at 1713, 1617, and 1611 cm-1

corresponding to vasy(CD2) of mono- and bidentate CF3CO2". Only one Mo-

Mo stretch was observed in the Raman spectrum (343 cm-1) indicating the

presence of only one centrosymmetric isomer. The 19F NMR spectrum of

Rh2(O2CCF3)4(pyr)4 prepared here showed signals at -74.1 and -74.9 ppm
in toluene-dg and at -74.7 and -75.4 ppm in CDC13, both at -60 C. In

both solvents the peak ratio was 1:1. Here mono- and bidentate CF3CO2

are separated by less than 1 ppm, whereas in the molybdenum work they

were separated by about 3 ppm. However, there are examples of mono- and
bidentate CF3CO2- with all resonances in the -74 to -76 ppm range. King
and Kapoor121 have synthesized a large number of complexes such as

(CsH5)Fe(CO)2(CF3CO2) which has monodentate CF3C02- and gives a 19F NMR
signal at -74.2 ppm in CDC13. Creswell and co-workers122 have prepared

compounds such as Os(CO)(PPh3)2(CF3C02)2 which has two 19F NMR signals

Figure 3-6. Possible structures for Lewis base adducts of
Rh2(02CCF3)4. Only one of the six isomers of
both Class II and Class III adducts are shown.


0 0 CF3 0 -
/0 \3
-C, c \ 0C-OCF
L- -h 0 L 0
0RRh Rh C Rh C h
0 \ CF, L

CF3 0 0 o 0


\o %
0 C-CF

CF3 o L/
0 0


at -75.44 and -75.22 ppm in CDC13 assigned to mono- and bidentate

CF3CO2-. Thus, not only do the locations of resonances in the rhodium

dimer differ from molybdenum, but the chemical shift differences between

CF3CO2-'s in different environments do not correspond. The IR data for

the pyridine complex are also in agreement with the Class III

formulation. Absorption bands for vasy(C02) were observed at 1705 and
1642 cm-1 (CHCl3 solution) and at 1705 and 1655 cm-1 (Nujol mull)

corresponding to mono- and bidentate CF3C02-. Another characteristic IR
band is 6(C02) which occurs at 740 cm-1 in Rh2(02CCF3)4. Free pyridine

has bands at 740 and 693 cm-1. In the pyridine complex bands were

observed at 760, 750, 738, 725, and 690 cm-1. It is likely that at

least two of the first three bands correspond to 6(C02) for mono- and

bidentate CF3C02-. The other absorption bands could be assigned to

either pyridine or Rh2(02CCF3)4 and the latter showed little change from

the base-free rhodium dimer. The IR spectrum of Rh2(O2CCF3)4(pyr)4 is

shown in Figure 3-7. Addition of excess pyridine, up to 20 equivalents,

caused no change in the 19F NMR spectrum. Two sharp signals of equal

intensity were still observed at -74.7 and -75.3 ppm in the toluene-d_

at 27 C. By contrast, in the molybdenum case106 the two peaks coalecse

at 30 C, indicating fast exchange. However, the slower exchange

observed here is not unusual since in the M(CO)(PPh3)2(CF3C02)2
complexes studied by Creswell and co-workers,122 separate resonances

corresponding to mono- and bidentate CF3CO2- were observed at room

temperature. As a final note, it should be mentioned that the synthesis
of "Rh2(02CCF3)4(pyr)2" was reported123 a number of years ago, but the

complex characterized only by C and H analysis. This procedure was

repeated here and a compound was isolated that was most likely a mixture

3Ol I INSNfl I%

of pyridine adducts. This was the result of using ethanol as solvent

rather than toluene, used here to obtain the pure 4:1 adduct. (See

Experimental Section.)

It is difficult to draw general conclusions from the above results

based on criteria such as steric size, a-donor, and n-acceptor abilities

of the bases. Triethylamine was the strongest a-donor used; it is bulky

and has no i-acceptor ability. It formed axial adducts. Similarly,

quinuclidine (a Lewis base very comparable to Et3N) formed a Class I

(axial) adduct with Mo2(02CCF3)4 although its size and a-basicity would

favor Class II. Pyridine, a base with less a-donor ability than Et3N,

has i-acceptor ability and formed a stable Class III (axial and

equatorial) adduct. N-methylimidazole, a stronger a-donor but a poorer

n-acceptor than pyridine caused dimer cleavage although via a Class I

adduct. Piperidine is a strong a-donor, but reactivity is most likely

due to the protonic nature of the base.

A final nitrogen donor base, acetonitrile, was used. It is a weak

a-donor, but a i-acceptor. Bear and co-workers40 were unable to isolate

a stable acetonitrile adduct of Rh2(02CCH3)4. These workers claimed

that evaporation of a CH3CN solution of rhodium acetate gave only

starting material.40 It was found here, by contrast, that stable purple

Rh2(02CCH3)4(CH3CN)2 was formed upon evaporation of an acetonitrile

solution of the rhodium dimer. (See Experimental Section.) However,

although Rh2(02CCF3)4(CH3CN)2 can be similarly prepared, it readily

loses acetonitrile and is hydrated to a blue-green material upon

standing in air. The freshly prepared complex 19F NMR showed a singlet

at -74.1 ppm and a doublet at -74.5 ppm in CDC13 at -60 C. The area

ratios were 2:1:1. Addition of excess CH3CN (-10 equivalents) led to

signals at -74.7 and -75.4 ppm in equal area ratios. IR data showed

Class I bridging CF3CO2- bands. The solution structure of Rh2(02CCF3)4

in the presence of acetonitrile is thus uncertain. However, the weaker

coordination of CH3CN to the fluorocarboxylate as opposed to the

alkylcarboxylate rhodium dimer is to be expected. This is due to r-

backbonding interactions as discussed earlier and shown by Drago and co-


Carbon Donors

Reactions with the isoelectronic carbon donors t-butylisonitrile

and carbon monoxide were investigated. The reaction of t-BuNC with a

variety of metal carboxylate dimers was studied by Girolami and

Andersen.76 They found that only monomeric complexes were obtained with

Mo2(02CCH3)4, Mo2(02CCF4)4, Re2(02CCH3)4C12, and Ru2(02CCH3)4C1.
However, with Rh2(02CCH3)4 only the Class I adduct Rh2(O2CCH3)4(t-BuNC)2

was produced. It was of interest to determine what effect replacement

of CH3CO2- by CF3CO2- would have in the dirhodium system. It was found

here that reaction of Rh2(03CCF3)4 with t-BuNC (~10 equivalents) led to

isolation of an air stable orange-brown complex best formulated as

Rh2(02CCF3)4(t-BuNC)4. Unfortunately, in contrast to the pyridine

complex which had the same stoichiometry and easily interpretable NMR

and IR spectra, t-BuNC gave complicated results, as will be discussed

below. This is most likely due to the presence in solution of a variety

of species including more than one isomer of a Class III adduct and

possibly monomeric species. Although t-BuNC and pyridine have similar

a-donor properties, the isonitrile is a better T-acceptor and somehow

this may lead to a variety of isomers of comparable stability. The 19F

NMR spectrum of this compound showed six peaks occurring between -73.0

and -74.6 ppm in CDC13 at -60 C. The1H NMR spectrum showed signals at

1.61 and 1.43 ppm in CDC13 at -50 C. All NMR data for nuclei other

than 19F are summarized in Table 3-III. At room temperature the peaks

occurred at 1.60 and 1.40 ppm, but instead of a 3:2 area ratio the ratio

was 2:1. Thus, at different temperatures different isomers

predominated, but specific assignment of the signals was not possible.

The IR spectrum of this complex showed bands assignable to vasy(C02) at

1693 and 1658 cm-1 in CHC13 solution and at 1720 and 1660 cm-1 in the

solid state. A single strong 6(C02) band was observed at 725-730 cm-1

(Nujol mull). There may have been more than one 6(C02) band, but
resolution was not possible. Very strong absorption bands corresponding
to v(NC) occurred at 2234 and 2167 cm-1 (CHC13 solution) and at 2212 and

2132 cm-1 (Nujol mull) as opposed to 2127 cm-1 for free t-BuNC. This

shift to higher frequency is expected for end-on isonitrile

coordination. The other absorption bands were assignable to either t-

BuNC or Rh2(02CCF3)4. The IR spectrum of Rh2(02CCF3)4(t-BuNC)4 is shown

in Figure 3-8. Although the solution and solid state IR spectra were

qualitatively the same, the fairly large difference for a given band

such as vasy(CO2) or v(NC) may indicate a different structure in

solution. Further studies with this complex would be needed to

unequivocally determine its structure. However, it seems clear that a

Class I adduct is not formed in contrast to Rh2(02CCH3)4.76 It is not
surprising that a 4:1 complex is formed since t-BuNC is a good o-donor

and an excellent 7-acceptor. As found with pyridine, the CF3CO2- ligand
was needed to allow coordination to the equatorial sites.

The complex Rh2(02CCH3)4(CO)2 has been isolated and structurally
characterized by x-ray crystallography.124 The v(CO) band occurs at


- o ___ j_ ,


i- 4
0 0


- / I

s <. -s i

i ^ c,

^ ~ ~" ----

3N3111-SNVYi %

Table 3-111. 1H and 31p 1H} NMR Data for Rh2(O2CR)4 Complexes




100 MHz

100 MHz, CCl4 solution

300 MHz


300 MHz

Nucleus Chemical Shift (ppm)b

1H 1.61 s, 1.43 s (3:2)
1.60 s, 1.40 s (2:1)


+32.81 d

-14.25 (very weak)

-24.42 d

(major signals 1:1)

+34.78 d

-14.8 t (weak)

-23.66 d

(major signals 1:1)

+34.25 d of d

-15.1 d of d

-23.18 d of quart



- 7.80


d of d

d of d

d of quart





J=166.0 27

J= 92.7

J=153.0 27

J= 37


2j= 11.65 -50
1J= 47.4
2J= 33.8
J= 91.9
2= 11.7 (outer peaks)

15.1 (inner peaks)
J=165.2 27
J= 12.0

2J= 49.3
J= 34.8

J2= 88.9
2J= 12.0 (all peaks)


Table 3-III. continued

Nucleus Chemical Shift (ppm)b


100 MHz

300 MHz

300 MHz

300 MHz



(empircal formula)

eight peaks in +161 to

+75 ppm range, -18.2

+70.00 d of d

+69.72 d of d

+69.53 t

-18.91 br n

+58 m

+20 m

-72 t

J= 50 27

one observed

J 20
Javg 20
J= 50

a. All complexes are in CDC13 solution except as otherwise noted.

b. 3P chemical shifts relative to external 85% H3PO4. 1H chemical shifts
relative to internal TMS.

c. Decomposition occurring during data collection. Chemical shift of
OP(OPh)3 is ca. -18 ppm.




2105 cm-1, below that of free CO (2143 cm-1), indicative of ir-

backbonding. The mechanism. of this was discussed earlier. Rh2(O2CCF3)4
also forms a 2:1 adduct with CO although the CO is much more weakly

bound. Indeed, it was not possible to isolate a CO adduct of

Rh2(O2CCF3)4, CO was too readily lost. However, other workers125
reported isolation of this adduct as a light brown solid. The IR

spectrum of this complex prepared as a KBr pellet under 1 atm of CO
showed v(CO) at 2150 cm-1 and vasy(C02) at 1644 cm-1'125 It was found

here that bubbling CO through a solution of Rh2(O2CCF3)4 in CH2C12 led

to appearance of a purplish blue color, resembling that formed with

similar weak donors such as acetonitrile. The brown solid is surprising

since this resembles complexes formed with strong donors such as
phosphines and phosphites. The IR spectrum of this CH2C12 solution
showed bands assignable to v(CO) at 2160 cm-1 (m) and to vasy(CO2) at

1660 (s) and 1760 cm-1 (m). The former vasy(CO2) band may correspond to

CO free Rh2(O2CCF3)4. This positive shift in v(CO) from free CO was
taken as evidence of no r-backbonding in Rh2(O2CCF3)4.125 However, this

is not a definitive argument. If there were no r-backbonding it is
unlikely that CO would coordinate at all. As shown by Drago,26 BF3,
which using the E and C analysis,27-29 is a stronger Lewis acid than

Rh2(02CCF3)4, but does not bind CO since BF3 cannot provide any T-
backdonation. The perturbation from o effects could cause an increase
in v(CO) in Rh2(02CCF3)4(CO)2 comparable to the decrease cause by n

effects since both effects are small.

Phosphorus Donors

As mentioned previously, a large number of phosphine derivatives of

Mo2(02CCF3)4 have been reported.33,105 However, phosphites do not form
adducts with Mo2(02CCF3)4 presumably since they are not strong enough o-
donors. They do form axial complexes with Rh2(02CCH3)4 since in
contrast to the molybdenum system there is a significant r-backbonding

stabilization. It was of interest to extend this work to Rh2(02CCF3)4
since only triphenylphosphine and triphenyl phosphite adducts of

Rh2(02CCF3)4 have been reported.111 These complexes were studied by x-
ray crystallography and found to be Class I adducts. However, their

solution properties have not been investigated. The phosphorus donors
used here were dimethylphenylphosphine (PMe2Ph), triphenylphosphine

(PPh3), tricyclohexylphosphine (P(c-Hx)3), triphenyl phosphite (P(OPh)3)
and trimethyl phosphite (P(OMe)3).

PMe2Ph forms a Class II adduct with Mo2(02CCF3)4 due to its small
size and strong basicity.33 Thus, it would be a good candidate to form

a Class III adduct with Rh2(02CCF3)4, Unfortunately, the reaction of

Rh2(02CCF3)4 with four equivalents of PMe2Ph yielded only an intractable
orange oil indicating dimer decomposition.

PPh3 lies far outside the size and basicity range described by
Andersen33 for Class II adduct formation. Furthermore, in the solid

state Rh2(02CCF3)4(PPh3)2 is a typical Class I adduct.111 Thus, this
complex would be unlikely to show unusual solution behavior and one
would expect a simple 19F NMR spectrum such as that found for the THF

adduct. This was not the case. A freshly prepared solution of

Rh2(02CCF3)4(PPh3)2 showed sharp 19F NMR resonances at -74.4, -74.9, and
-75.9 ppm in CDC13 at 27 C in area ratios of 1:1:2. There was also a

small peak at -75.3 ppm. At -50 C there were still three sharp, major

peaks only in an area ratio of 1:1:1.3. That there was little change

over this temperature range indicates that the same species were

present, although perhaps in differing amounts. Assignment of these

peaks is difficult, presumably they corresond to mono- and bidentate

CF3CO2-. However, the situation differs from that observed with the

pyridine adduct and from the solution studies on Mo2(02CCF3)4 with

pyridine.106 In those cases there were two peaks representing one Class

III isomer with 1:1 mono- and bidentate CF3CO2. The more complex

spectrum observed here could be the result of a mixture of iscmers

containing axially and equatorially coordinated PPh3. That there would

be anything other than axial coordination in solution is surprising.

However, it is possible that in solution the dimer may dissociate to

some extent. The molecular weight of Rh2(02CCF3)4(PPh3)2 in CH2C12 was

found to be 590, half the expected value of 1183. This value could

result from the existence of Rh2(02CCF3)4(PPh3) and free PPh3 in

solution. However, if these were the major solution species, then only

one 19F NMR resonance would be observed, although perhaps weak signals

corresponding to 2:1 and base free species would be seen with similar

chemical shifts. Furthermore, a singlet corresponding to free PPh3

would be observed in the 31p{1H} NMR spectrum or a single broad peak

corresponding to fast exchange between free and coordinated PPh3. Such

behavior was found by Boyar and Robinsonl26 who very recently reported

the 31p{lH} NMR spectrum of Rn2(u2CCH3)4(P(OMe)3)2 in dichloromethane-d~

solution. These workers126 observed a single broad peak at room

temperature. The 31P{1H} NMR spectrum of Rh2(but)4(PPh3)2 in CDC13 at

room temperature was obtained here and it also exhibited a single broad