β-Diketones containing a ferrocenyl group: Synthesis, structural aspects, pKa1 values, group electronegativities and complexation with rhodium(I)

Article abstract of DOI:10.1039/a802398k

1-Ferrocenyl-4,4,4-trifluorobutane-1,3-dione (ferrocenoyltrifluoroacetone, Hfctfa, pK_a¹ = 6.53 ± 0.03), 4,4,4-trichloro-1-ferrocenylbutane-1,3-dione (trichloroferrocenoylacetone, Hfctca, pK_a¹ = 7.15 ± 0.02), 1-ferrocenylbutane-1,3-dione (ferrocenoylacetone, Hfca, pK_a¹ = 10.01 ± 0.02), 1-ferrocenyl-3-phenylpropane-1,3-dione (benzoylferrocenoylmethane, Hbfcm, pK_a¹ = 10.41 ± 0.02) and 1,3-diferrocenylpropane-1,3-dione (diferrocenoylmethane, Hdfcm, pK_a¹ = 13.1 ± 0.1) were prepared by Claisen condensation of acetylferrocene with an appropriate ester under the influence of sodium amide, sodium ethoxide or lithium diisopropylamide. The group electronegativity of the ferrocenyl group is 1.87 (Gordy scale) as inferred from a linear β-diketone pK_a¹-group electronegativity relationship as well as from a linear methyl ester IR carbonyl stretching frequency-group electronegativity relationship. Complexes [Rh(β-diketone)(cod)] were obtained in yields approaching 80% by treating the β-diketones with [Rh₂Cl₂(cod)₂], while the copper(II) chelates form just as readily. Treatment of all [Rh(β-diketone)(cod)] complexes with 1,10-phenanthroline (phen) and some of its derivatives resulted in substitution of the β-diketone ligand to form [Rh(cod)(phen)]⁺. The uncomplexed β-diketones are increasingly stable towards the OH^- nucleophile in the order Hdfcm (apparent most unstable) < Hfctfa < Hbfcm < Hfctca < Hfca (most stable). Asymmetric enolisation in the direction furthest from the ferrocenyl group was observed for all β-diketones. This finding is considered to be the result of resonance driving forces rather than inductive electronic effects of substituents on the pseudo-aromatic β-diketone core.

Full text of DOI:10.1039/a802398k

View Article Online / Journal Homepage / Table of Contents for this issue
â-Diketones containing a ferrocenyl group: synthesis, structural
1
aspects, pKa values, group electronegativities and complexation with
rhodium(I)
W. C. (Ina) du Plessis, Theunis G. Vosloo and Jannie C. Swarts*
Department of Chemistry, University of the Orange Free State, Bloemfontein, 9300,
Republic of South Africa
1-Ferrocenyl-4,4,4-trifluorobutane-1,3-dione (ferrocenoyltrifluoroacetone, Hfctfa, pKa1 = 6.53 0.03), 4,4,4-
trichloro-1-ferrocenylbutane-1,3-dione (trichloroferrocenoylacetone, Hfctca, pKa1 = 7.15 0.02), 1-ferrocenyl-
butane-1,3-dione (ferrocenoylacetone, Hfca, pKa1 = 10.01 0.02), 1-ferrocenyl-3-phenylpropane-1,3-dione
(benzoylferrocenoylmethane, Hbfcm, pKa1 = 10.41 0.02) and 1,3-diferrocenylpropane-1,3-dione (diferro-
cenoylmethane, Hdfcm, pKa1 = 13.1 0.1) were prepared by Claisen condensation of acetylferrocene with an
appropriate ester under the influence of sodium amide, sodium ethoxide or lithium diisopropylamide. The
1
group electronegativity of the ferrocenyl group is 1.87 (Gordy scale) as inferred from a linear β-diketone pKa –
group electronegativity relationship as well as from a linear methyl ester IR carbonyl stretching frequency–group
electronegativity relationship. Complexes [Rh(β-diketone)(cod)] were obtained in yields approaching 80% by
treating the β-diketones with [Rh2Cl2(cod)2], while the copper() chelates form just as readily. Treatment of
all [Rh(β-diketone)(cod)] complexes with 1,10-phenanthroline (phen) and some of its derivatives resulted in
substitution of the β-diketone ligand to form [Rh(cod)(phen)]ϩ. The uncomplexed β-diketones are increasingly
stabletowardstheOHϪ nucleophileintheorderHdfcm(apparentmostunstable) < Hfctfa < Hbfcm < Hfctca < Hfca
(most stable). Asymmetric enolisation in the direction furthest from the ferrocenyl group was observed for all
β-diketones. This finding is considered to be the result of resonance driving forces rather than inductive
electronic effects of substituents on the pseudo-aromatic β-diketone core.
β-Diketone complexes of transition metals have been the sub-
ject of many different studies ranging from synthetic,1 kinetic2
and structural3 topics to catalysis4 and many others.5 Ferrocene
and its derivatives, on the other hand, have been investigated1,6
equally well, inter alia because of their use as colour pigments,7
high burning rate catalysts8 in solid fuel, liquid fuel combustion
catalysts9 and smoke suppressant additives.10 A new field of
potential application of both β-diketone complexes of rho-
dium() and derivatives of ferrocene has evolved in recent years
with reports that some of these compounds show appreciable
antineoplastic activity. Thus it has been shown that, compared
to cisplatin [Pt(NH3)2Cl2],11 certain ferrocenium salts12 have
more favourable 50% lethal dosage (LD50) values, and that
[Rh(acac)(cod)] (acac = acetylacetonate) is more effective
against Erlich Ascite tumours.13 Today cisplatin is still one of
the most widely used metal-containing chemotherapeutic drugs
in the USA, Europe and Japan,14 but suffers, as do all other
chemotherapeutic agents, from many side effects. These include
inter alia high toxicity to the kidneys and bone marrow,15 loss of
appetite (anorexia),16 development of drug resistance after con-
tinued drug dosage,17 a high rate of excretion from the body,18
low aqueous solubility and, perhaps most important of all,
inability to distinguish between healthy and cancerous cells.19
To combat the negative aspects surrounding cisplatin and other
chemotherapeutic drugs, new antineoplastic materials are
continuously being synthesized and evaluated, new methods
of delivering an active drug to a cancerous growth are
being developed20 and combination chemotherapy has been
investigated in the hope of finding synergistic effects.21 Partly
because of this, partly because of our kinetic2 and structural3
programmes regarding the chemistry of β-diketonato com-
plexes of rhodium() and also because of our synthetic program
on ferrocene derivatives,22 we decided to investigate the struc-
tural, thermodynamic and chelate-forming properties of various
ferrocene-containing β-diketones with RhI. Since the ferrocene-
containing rhodium() chelates obtained are constructed
from more than one antineoplastic moiety within the same
molecule, they hold the promise of displaying synergistic effects
in chemotherapy without the need of administering two or
more types of antineoplastic drugs simultaneously to a tumour-
bearing mammal.
β-Diketones are normally prepared by Claisen condensation
of appropriate carbonyl-containing compounds.23 The strong
electron-donating properties of the ferrocenyl group lower the
acidity of the methyl hydrogen atoms of acetylferrocene (1)
which in turn necessitates the use of strong bases (i.e. metal
amides or alkoxides) to ensure reasonable yields. Thus Hauser
and co-workers24 synthesized various ferrocene-containing
β-diketones, including 1-ferrocenylbutane-1,3-dione (ferro-
cenoylacetone, Hfca) and 1-ferrocenyl-3-phenylpropane-
1,3-dione (benzoylferrocenylmethane, Hbfcm), by using potas-
sium amide as the active base additive in liquid ammonia
as solvent. Weinmayr1b utilised sodium methoxide as basic
initiator to prepare both Hfca and 1-ferrocenyl-4,4,4-trifluoro-
butane-1,3-dione (ferrocenoyltrifluoroacetone, Hfctfa) in
diethyl ether, while Cullen et al.4a favoured the use of the
sterically hindered base lithium diisopropylamide in the prep-
aration of Hfca.
Results and Discussion
Upon utilising sodium amide as basic initiator, Hauser’s
method24 resulted in Hfca (2, R = CH3, Scheme 1) in yields not
exceeding 22% based on the initial amount of acetylferrocene
used. In contrast, 1,3-diferrocenylpropane-1,3-dione (Hdfcm)
(2, R = ferrocenyl) could only be obtained in trace amounts
(ca. 5%) via this route. By adapting the method of Cullen et al.4a
to a one-pot procedure, we found the LiNPri2 route effective
for Hfca (38% yield), Hbfcm (2, R = phenyl, 28% yield), Hdfcm
(30% yield) and 4,4,4-trichloro-1-ferrocenylbutane-1,3-dione
(trichloroferrocenoyl acetone, Hfctca) (2, R = CCl3, 16% yield)
synthesis provided the added base was never the limiting
J. Chem. Soc., Dalton Trans., 1998, Pages 2507–2514
2507

View Article Online
route 1
NaNH2, NH3(l)
O
O
O
O
OH
H +
RCOOR1
route 2
NaOEt, ether
R
R
Fe
Fe
Fe
route 3
LiNPri2, thf
H +
– H2O
1
2
CH3CO2Na
Cu(CH3CO2)2
H +
O
R
R
O
O
O
O
Cu
Fe
Fe
Cl
Cl
Rh
Rh
Fe
Fe
3
phen, NaClO4
4
R
+
O
Rh
O
N
N
phen,
[ClO4] –
Rh
Fe
NaClO4
5
6
Scheme 1 Claisen condensation of acetylferrocene 1 with appropriate esters (R1 = ethyl or methyl) give the β-diketones Hfctfa (2, R = CF3), Hfctca
(2, R = CCl3), Hfca (2, R = CH3), Hbfcm (2, R = phenyl) and Hdfcm (2, R = ferrocenyl); NMR studies indicated that asymmetric enolisation as
shown dominates. Self-aldol condensation of 1 leads to the side product 3. Copper and rhodium complexation proceed with ease to give 4 and 5 while
1,10-phenanthroline substituted the β-diketone in 5 to exclusively give [Rh(cod)(phen)]ϩ
6
reagent (to minimise self aldol condensation of acetylferrocene)
and rigorous Schlenk conditions were adhered to. Weinmayr’s
relatively simple alkoxide method1b takes much longer than the
more involved amide routes but still results in reasonable yields
of Hfca (≈35%) as well as Hfctfa (2, R = CF3; ≈54%). However,
we were unsuccessful in obtaining Hfctca and Hbfcm via the
alkoxide route. 1,3-Diferrocenylbut-2-en-1-one (1-ferrocenoyl-
2-ferrocenylpropene) 3, the dehydrated aldol condensation
product25a of acetylferrocene, was also isolated from the
Hdfcm, Hbfcm and Hfctca reaction mixtures in yields р23%.
Exhaustive column chromatography of compounds 2 and/or
fractional precipitation of the copper complexes 4 followed by
free β-diketone generation with 6 mol dmϪ3 HCl (Scheme 1)
were needed to separate 2 from 3. The rhodium() complexes,
[Rh(β-diketone)(cod)] 5, were easily prepared in high yield
(>75% in dmf from the rhodium() dimer [Rh2Cl2(cod)2] but
3 does not react with either CuII or RhI. All five complexes
of 5 reacted with 1,10-phenanthroline (phen) to generate
[Rh(phen)(cod)]ϩ 6. The fca is also substituted in 5 (R = CH3) if
it is treated with the 5-nitro-, 4,7-dichloro-, 5,6-dimethyl-, 4,7-
dimethyl-, 2,9-dimethyl- and 3,4,7,8-tetramethyl-phenanthro-
line derivatives to generate the corresponding substituted
phenanthroline complex. Confirmation of these reactions,
which are the subject of a detailed kinetic study in a forth-
coming paper, was obtained by comparing the IR and 1H NMR
spectra of the derivatives of 6 so obtained with that of an
authentic sample prepared from [Rh2Cl2(cod)2] (Scheme 1).
The crystal structure of 5 (R = CH3) has recently been
reported.3b Two molecules within the same unit cell were
observed, one of which (molecule A) approached very much the
extreme case of asymmetric co-ordination with rhodium. As a
result, for molecule A, the geometry of the co-ordinated fca is
very similar to that found for the asymmetrically enolised free
(unco-ordinated) Hfca.26 The geometry of the fctfa ligand co-
ordinated to the RhI was also recently established.3c Structural
investigations of fctca-, bfcm- and dfcm-containing rhodium()
and iridium() complexes are currently in progress. Here we use
the published26 crystallographic data of Hfca to explain the
observed asymmetric enolisation for all β-diketones in a direc-
tion opposite to the aromatic ferrocenyl side group and the
following properties of this structure are pertinent. Bell et al.26
found that for Hfca, in the solid state, asymmetric enolisation
takes place in the direction away from the aromatic ferrocenyl
side group. The dihedral angle of 4.9(2)Њ between the pseudo-
aromatic β-diketone plane and the planar cyclopentadiene ring
of the ferrocenyl group attached to the β-diketone skeleton
indicates appreciable conjugation between the two groups. The
longer bond distance between the β-diketone skeleton and the
methyl group, 1.490(8) Å, as compared to β-diketone core/
ferrocenyl group distance of 1.468(7) Å, also indicates that the
ferrocenyl group conjugates well with the pseudo-aromatic
β-diketone core. Regarding the side product 3, a single crystal
structural determination27 has shown that the ferrocenyl and
ferrocenoyl groups are situated trans with respect to each other.
By comparing the average dihedral angles of the cyclopenta-
dienyl planes of the ferrocenyl and ferrocenoyl groups of 3
linked to plane 1 [consisting of the atoms COCHC(CH3),
Scheme 1], 12.5(3) and 19.3(3)Њ respectively, one can conclude
that, although both ferrocenyl moieties show appreciable con-
jugation into the planar (sp2 conjugated) carbon backbone of
plane 1, they do so much less effectively than the ferrocenyl
group of Hfca conjugating into the β-diketone core. This is
expected, for the β-diketone plane is pseudo-aromatic while
plane 1 is not.
2508
J. Chem. Soc., Dalton Trans., 1998, Pages 2507–2514

View Article Online
Table 1 pKa1 Values and % enol tautomer of various β-diketones. Htfaa = 1,1,1-trifluoroacetylacetone
1 a
1 a
Compound
pKa
% Enolb
Compound
pKa
% Enolb
Hhfaa
Htfaa
Hbtfa
Hfctfa
Hfctca
Hba
4.35c
100c
>99
>99
>99
≈95
92d
Haa
Hfca
Hdbm
Hbfcm
Hdfcm
8.95 0.08c–e
10.01 0.02d,e
9.35c
91d, f
6.3c
86d,g
6.3c
>99d
≈95d
>99d
6.53 0.03d,e
7.15 0.02d,e
8.7c
10.41 0.02d,e
13.1 0.1d,e
a At 21 ЊC. b In CDCl3 at 298 K. c Ref. 2(b). d This work. e In water containing 10% acetonitrile, I = 0.1 mol dmϪ3 (NaClO4). f However, see also ref. 28.
g However, see also ref. 26.
From a 1H NMR study, by comparing the relative intensities
of the CH2 (keto) and CH (enol) signals, the percentages of
enolised tautomers in solution of the prepared and some other
β-diketones were established (Table 1). Regarding Hfca, the
apparent absence of more than one set of signals for the ferro-
cenyl substituent as well as the two observed signals for the
methyl side group indicate that, as in the solid state, enolisation
in solution is predominantly away from the aromatic ferrocenyl
group. This result complements that obtained by Bell et al.26
who were unable to predict under their conditions the preferred
Hfca enol isomer in solution. It should be noted that, in order
to observe the keto isomer in solution, the concentration of the
β-diketone solution should be fairly low. High concentrations
drive the keto–enol equilibrium towards the enol side3f which
explains why Bell could not observe the keto isomer in sig-
nificant quantities. This solution behaviour is also observed3f
for other β-diketones. Enolisation for Hbfcm, in solution, pre-
dominantly took place towards the phenyl group as demon-
ferrocenyl and phenyl groups, enolisation to take place in the
direction of the least electronegative ferrocenyl group. This is
also in contrast to what was found. Clearly there is a different
driving force than the suggested electronic driving force that
determines the observed preferred enol configuration in β-
diketones where aromatic side groups are present. To explain
this observation the existence of a resonance driving force is
proposed. The resonance driving force implies that the
formation of different canonical forms of a specific isomer will
lower the energy of this specific isomer enough to allow it to
dominate over the existence of other isomers. Crystallographic
studies may be used to support the existence of the resonance
driving force as explained for Hfca below. However, the same
argument may also, with success, be applied to all the other
β-diketones.
It should first be noted that crystallographic determinations
of the structure of several compounds have indicated that both
the ferrocenyl and the phenyl groups conjugate very well with
adjacent carbonyl,27,30a alkene,27 aromatic6,30b and pseudo-
aromatic26 substituents in free compounds27,30 and in com-
plexed form.3 With respect to β-diketones, this conjugation can
take place in at least two ways as demonstrated for the two
theoretically possible enol forms I and II of Hfca in Scheme 2.
To understand why under our experimental conditions only
isomer II was observed, implying that the equilibrium between
I and II is shifted very far towards II, one may consider the
canonical forms III and IV. The relatively short C11᎐C111
bond length [1.468(7) Å, see Scheme 2 for atom labelling]
compared to the C᎐C bond length of alkanes [for example3b
lengths of 1.500–1.552(8) Å observed in the cod fragment of
5, R = CH3; bond lengths between the β-diketone pseudo-
aromatic core and the carbon atoms of adjacent alkyl groups
are usually3a,b,26,27 around or larger than 1.5 Å] indicates
substantial conjugation between the ferrocenyl group and the
β-diketone core. Inspection of the bond distances of the
cyclopentadienyl ring attached to the β-diketone skeleton26
indicates bond lengths of 1.435(6) for C111᎐C112, 1.415(7) for
C112᎐C113, 1.417(7) for C113᎐C114, 1.426(7) for C114᎐C115
and 1.420(7) Å for C111᎐C115 respectively. These bond lengths
would fit both canonical forms III and IV. However, the C᎐O
bond lengths fit only structure IV. Bond C13᎐O2 is long
[1.307(7) Å] and typical of an enol structure. In contrast bond
1
strated by two distinct sets of H NMR signals for the phenyl
group versus the single set of signals for the ferrocenyl group. In
the case of 1-phenylbutane-1,3-dione (benzoylacetone, Hba),
1
the lack of more than one set of phenyl H NMR signals and
two methyl signals indicate enolisation took place in a direction
away from the aromatic phenyl group. Excluding Hdbm (diben-
zoylmethane) and Hdfcm, in the case of β-diketones which are
>99% enolic (Table 1, these β-diketones have only one 1H NMR
active side group, ferrocenyl or phenyl) the relative high field
position of the aromatic signals indicates that for Hfctfa,
Hfctca and 4,4,4-trifluoro-1-phenylbutane-1,3-dione (Hbtfa)
enolisation most likely takes place predominantly away from the
aromatic substituent. In order to explain the dominance of the
observed enol isomer in each case, two different driving forces
that will control the conversion from β-diketone into an enolic
isomer may be defined. The first may by labelled an electronic
driving force in which the formation of the preferred enol
isomer is controlled by the electronegativity of the R and RЈ
substituents in the β-diketone RCOCH2CORЈ. When the electro-
negativity of R is greater than that of RЈ the carbon atom of
the carbonyl group adjacent to RЈ will be less positive in char-
acter than the carbon atom of the other carbonyl group. Con-
sequently, from an electronic point of view, the dominant enol
isomer should be RCOCH᎐C(OH)RЈ. If the documented group
᎐
electronegativities29 are correct, from the viewpoint of an elec-
tronic driving force as just described, only Hba {group electro-
negativity of the phenyl group, χphenyl = 2.43 (apparent) or 3.0
[corrected, see ref. 29(a)], while χmethyl = 2.3029a or 2.3429b} of all
the β-diketones just discussed exhibits the expected dominant
enol isomer. For all the other mentioned β-diketones having
only one aromatic side group one would expect enolisation to
take place predominantly in the direction of this aromatic sub-
stituent because in each case the electronegativity of the tri-
fluoromethyl29b (3.20), trichloromethyl29b (2.76) or methyl group
is larger than that of the other substituents: either the phenyl or
the ferrocenyl group (electronegativity ≈1.87 as per this work).
In the case of Hbfcm, which has two aromatic side groups, one
would expect, by considering the electronegativities of the
C11᎐O1 [1.287(6) Å] is not short enough for a typical C᎐O
bond [ca. 1.206 Å, see for example ref. 3(d)]. Indeed, this bond
is so long it more closely approaches typical single C᎐O than
᎐
double C᎐O bond lengths, a situation that can only arise if IV
᎐
makes an appreciable contribution to the overall structure of
Hfca. Thus, the crystal structure determination of Hfca indicates
that zwitterion IV arising from isomer II dominates over III
arising from isomer I. It should be noted that although the
canonical forms indicated in Scheme 2 explain the dominance
1
of isomer II over I by virtue of the observed H NMR and
crystallographic data, they by no means imply that other
canonical forms and relationships do not also exist. Based on
the above arguments, the indications are that with respect to
aromatic side groups the driving force which determines which
J. Chem. Soc., Dalton Trans., 1998, Pages 2507–2514
2509

View Article Online
O
O
electronic
resonance
Fe
driving force
driving force
HO
O
O
OH
Fe
Fe
Fig. 1 Absorbance dependence on pH of Hfctfa (0.1266 mmol dmϪ3
,
320 nm), Hfctca (0.1044 mmol dmϪ3, 330 nm), Hfca (0.1059 mmol
dmϪ3, 326 nm), Hbfcm (0.0639 mmol dmϪ3, 360 nm) and Hdfcm (9.230
mmol dmϪ3, 420 nm) at 21 ЊC. The absorbance indicated for Hfctfa is
exact, but those for all the other β-diketones were adjusted to a relative
value in order to show all curves on the same axis system. Only ᭿ data
were used in the fitting program. For Hfctfa, the points shown as ᭹ (at
I
II
HO
O –
– O1 O2H
115
+
1
pH close to 8; not included in the data set used for the pKa fitting)
+
11
111
13
114
indicate an OHϪ consumption to form a new compound speculated to
be a hydroxylated species similar to that formed by hexafluoroacetyl-
acetone (see Scheme 3). Solvent: water containing 10% acetonitrile,
I = 0.100 mol dmϪ3 (NaClO4)
12
14
112
113
Fe
Fe
III
IV
in Table 1 by having two aromatic substituents per molecule.
Hence resonance stabilisation of Hdbm, Hbfcm and Hdfcm, as
just explained for Hfca, will increase greatly and account for the
observed seemingly unusual high percentage of enolisation in
Scheme 2 Electronic considerations in terms of electronegativity, χ,
favour I as the enol form of Hfca. However, structure II was shown by
crystallography and NMR spectroscopy to be dominant. A dihedral
angle of 4.9(2)Њ between the aromatic ferrocenyl group and the pseudo-
aromatic β-diketone core implies that energy-lowering canonical forms
such as IV make a noticeable contribution to the overall existence of
Hfca. For clarity the ferrocenyl group in II and IV is shown in just one
canonical form but in both cases the iron atom can be bound to any of
the five cyclopentadienyl carbon atoms as indicated in I. Likewise, the
positive charge of IV is also not confined to the single position shown
but rather oscillates between C112 and C115 (it cannot be on C111;
atom numbers are indicated next to individual atoms) to give rise to
four different canonical forms as indicated in III. χmethyl = 2.34,29b
χferrocenyl = 1.87
1
these three compounds. The newly determined pKa values of
the β-diketones cited in Table 1 were obtained from a least-
squares fit of UV absorbance/pH data using equation (1)31a
1
AHA10ϪpH ϩ AA10ϪpK
a
(1)
AT =
1
10ϪpH ϩ 10ϪpK
a
1
1
(Fig. 1, single pKa ) or (2)31b (Fig. 1, two pKa values) with
AT = total absorbance, AHA the absorbance of the β-diketone in
the protonated form, and AA the absorbance of the deproton-
1
ated (basic) form. In equation (2), pKa2 represents the appar-
CF3
L
CF3
L
III
Co
1
1
1
ϪpKa2
AHA(10ϪpH)2 ϩ AA(10ϪpK )(10ϪpH) ϩ AF(10ϪpK )(10
)
a
a
L
L
O
O
L
L
O
O
OR
III
+
– OR
AT =
Co
1
1
1
(10ϪpH)2 ϩ (10ϪpH)(10ϪpK ) ϩ (10
)(10ϪpKa2
)
ϪpKa
a
L
L
CF3
CF3
(2)
Scheme
3
Reversible hydroxylation (R = H) and methoxylation
1
ent pKa , or more likely the pH dependent formation constant
of an in situ formed, possibly hydroxylated species (compare
Scheme 3) with final absorption AF, specifically in the case of
Hfctfa and Hbfcm. Whether pH/absorbance data for Hfctfa
and Hbfcm were fitted by equation (1) (separate, single pKa1 fits
(R = methyl) for hexafluoroacetylacetone complexes of cobalt().
L = Ammonia or an amine, see ref. 32(a)
enolate is favoured is resonance stabilisation and not electronic
factors. Resonance stabilisation favours the formation of
intermediates such as II and IV while electronic factors favour
for each pKa ) or (2) (a single fit to determine both pKa1 values
simultaneously) made essentially no difference to the obtained
1
1
the formation of intermediates resembling I and III. Since H
1
1
NMR spectroscopy does not distinguish between a ferrocenyl
group on the keto side and one on the enol side of the enolic
form of Hdfcm [the same applies to the two phenyl groups of
1,3-diphenylpropane-1,3-dione (dibenzoylmethane, Hdbm)]
it is assumed that the keto–enol conversion in these two
compounds takes place on a timescale much faster than that
of NMR.
pKa values. The reason for pKa2 for Hfctfa and Hbfcm is
still unknown, but it is possible that reversible hydroxylation,
similar to that observed for 1,1,1,5,5,5,-hexafluoroacetylacetone
(Hhfaa)32a (Scheme 3) may take place. This subject is now under
further investigation. The electronic spectra of compounds 2, 5
and 6 (R = CH3) are shown in Fig. 2, while peak absorption
coefficients are in Table 2. The basic forms of 2 become pro-
gressively more stable towards alkaline solutions in the order
Hdfcm (most unstable) < Hfctfa < Hbfcm < Hfctca < Hfca
(most stable). The general instability of all β-diketones towards
aqueous alkali media (cleavage at the methine position takes
place33a) is probably the major reason for Hdfcm’s alkaline
instability. Since the β-diketones were not all well soluble in
pure or basic water, water–acetonitrile mixtures were used as
solvent. We have found that such mixtures have much less
The increasing tendency towards enolisation of all the β-
diketones having only one aromatic side group with stronger
1
acidity in the apparent pKa range of 4.35–10.01 (Table 1)
is expected when the electronegativity of the side groups is
considered. The term ‘apparent’ is used since in this study no
attempt was made to partition the experimentally obtained
1
pKa values between separate pKa values for the enol and keto
tautomers. All the β-diketones with trifluoroacetyl side groups
are acidic enough (pKa1 р 6.53) to ensure almost complete eno-
lisation (>99% at 298 K). The compounds Hdfcm, Hbfcm and
Hdbm differ in structure from all the other β-diketones listed
1
influence on β-diketone pKa determinations than32b do 1,4-
dioxane, methanol, ethanol or propan-2-ol. The effect of the
amount of acetonitrile or 1,4-dioxane in the solvent medium
2510
J. Chem. Soc., Dalton Trans., 1998, Pages 2507–2514

View Article Online
Table 2 Peak molar absorption coefficients, ε, at the corresponding wavelength, λmax, of free (in water containing 10% acetonitrile) and rhodium()
complexed β-diketones (in methanol)
λmax/nm (ε/dm3 molϪ1 cmϪ1
)
Compound
protonated
unprotonated
Compound
λmax/nm (ε/dm3 molϪ1 cmϪ1
)
Hfctfa
Hfctca
Hfca
Hbfcm
Hdfcm
469 (1130)
345 (6750)
473 (1010)
375 (10 320)
344 (11 360)
324 (6170)
[Rh(fctfa)(cod)]
[Rh(fctca)(cod)]
[Rh(fca)(cod)]
[Rh(bfcm)(cod)]
[Rh(dfcm)(cod)]
472 (1900)
495 (2360)
335 (8530)
475 (2380)
470 (1570)
335 (12 170)
330 (24 420)
345 (18 270)
369 (3130)
Fig. 3 Linear relationships observed between group electronegativ-
Fig. 2 The UV spectra of Hfca, both in acidic (——, pH 2.0) and
basic (––––, pH 13.0) form, recorded in water, superimposed on the
spectra of [Rh(fca)(cod)] (———) and [Rh(cod)(phen)]ϩ (ؒؒؒؒؒ) in
methanol
1
ities, χR, and pKa values of β-diketones of the type FcCOCH2COR,
Fc = ferrocenyl (upper) as well as carbonyl stretching frequencies,
ν(C᎐O) (lower), of methyl esters of the type RCOOMe; R is indicated
᎐
on each plot. For R = CF3 the ethyl ester was used. Points marked as ᭡
(this study) and ؉ (from ref. 29) were not used in the fitting
1
was tested using Hfca and Hfctfa as references. The pKa of
Hfca in water and water containing 10% acetonitrile were
9.96 and 10.01 respectively. For Hfctfa the pKa was 6.53 in
1
(Merck) were used without further purification. Liquid react-
ants and solvents were distilled prior to use; water was double
distilled. Diethyl ether and thf were dried by refluxing under
nitrogen over sodium wire and distilled directly before use.
Flash chromatography was performed on Kieselgel 60 (Merck,
grain size 0.063–0.2 mm, eluent ether–hexane 2:3 by volume)
or, if stated eluent 2, Sephadex LH-20 (Pharmacia, hexane–
ethanol 6:1 by volume) utilising an overpressure that never
exceeded 100 Torr (1 Torr = 1 mmHg = 133.32 Pa).
water containing 10% acetonitrile, 6.9 in water containing 5%
aqueous 1,4-dioxane and 7.0 in 50% aqueous acetonitrile
[at 21 ЊC and I = 0.1 mol dmϪ3 (NaClO4)].
In an attempt to use the obtained pKa1 values of compounds
2 to determine the group electronegativity of the ferrocenyl
group, χFc, from known χR values (Gordy scale29b) for the
other R substituents on the β-diketones an inconsistency was
observed for the published χPh value. Using literature values29
for group electronegativities it would not fit on the straight line
1
generated by the fit of pKa and χR (R = Ph, CH3 or CCl3).
Acid dissociation constant determinations and spectroscopy
Utilising the stretching frequency of the carbonyl group
Proton NMR spectra at 298 K were recorded either on a Bruker
AM-300 or WP-80 instrument, with chemical shifts presented
as δ values referenced to SiMe4 at 0.00 ppm. Electronic spectra
were recorded on a Hitachi Model 150-20 and IR spectra (KBr
pellets unless otherwise stated) on a Hitachi Model 270-50
instrument. The pKa1 values were determined by measuring the
absorbance at different pH during an acid–base titration in
water or acetonitrile–water mixtures, 1:9 by volume, I = 0.100
mol dmϪ3 (NaClO4) at 21.0 ЊC. β-Diketone concentrations were
where possible less than 0.2 mmol dmϪ3 and are indicated in
Fig. 1. A linear response by the pH meter (Orion model SA
720), fitted with a glass electrode, was ensured by calibration
with commercial buffers (Sigma) at pH = Ϫlog αH = 4.01, 7.00
[ν(C᎐O)] of the methyl esters of the indicated compounds
(Fig. 3) the effective or apparent χPh was redetermined as 2.21
by extrapolation of the linear relationship between χR and
᎐
ν(C᎐O) while χFc and χCF were found to be 1.87 and 3.01
᎐
R
3
respectively. The obtained effective CF3 group electronegativity
stands in contrast to literature values29 of 3.2 and 3.35. The
scattering of CF3 group electronegativities is the direct result of
the uncertainty in the structure observed for trifluoroacetyl-
containing molecules and is inter alia also echoed in reported
pKa values of 4.35–5.3 for hexafluoroacetylacetone.32a The
1
newly obtained χPh fitted the linear plot of pKa versus χR,
R = CCl3, CH3 or phenyl, very well (Fig. 3). Extrapolation
of this line also resulted in an apparent χFc value of 1.87 which
1
and 12.00 respectively, αH = activity of Hϩ. A test pKa deter-
is mutually consistent with the value derived from the ν(C᎐O)
᎐
mination was then performed by titrating the well characterised
compound acetylacetone with sodium hydroxide. A least
squares fit of the obtained UV absorbance/pH data for this
titration using equation (1), utilising the fitting program
vs. group electronegativity relationship. A polarographically
determined χFc value of 2.08 reported elsewhere33b is in close
agreement with the above mentioned results. Once again the
value obtained for the CF3 group did not fit the trend set by the
other R groups for reasons adequately described above.
1
MINSQ,33c resulted in a pKa of 8.95 0.08 in water. This
was within experimental error the same as the best available
1
published pKa for acetylacetone in water (8.878 0.005 when
I = 1 mol dmϪ3 and 8.98 when I = 0.0172 mol dmϪ3).32b It
was therefore concluded that the electrode was calibrated to
measure hydrogen ion concentration under the conditions used.
It is not expected that the electrode would behave differently for
Experimental
Materials
Ferrocene (Strem), n-butyllithium and other solid reagents
J. Chem. Soc., Dalton Trans., 1998, Pages 2507–2514
2511

View Article Online
1
1
any of the other pKa determinations because only pKa values
of a series of β-diketones were determined. For Hfca, titration
was performed with HClO4 from high pH, adjusted with
NaOH, to low pH since compound 2 (R = CH3) would not
dissolve in non-alkaline water. Owing to the slow rate of dis-
solving OH-unstable Hdfcm and Hfctca, solutions of these
β-diketones in acetonitrile were treated with aqueous NaOH to
a final mixed solvent constitution of 90% water and a pH high
enough to ensure that the β-diketones were fully in the depro-
tonated form, before titrations commenced. The influence of
acetonitrile on the pKa1 determination was studied by repeating
the determination for Hfca in water containing 10% aceto-
suspension for 1 h, dry ether (3 cm3) was carefully added
followed by slowly heating the mixture to 40 ЊC to remove
all NH3(l). The resulting suspension was refluxed for 15 min,
the residue filtered off, washed with ether (4 × 5 cm3) and dis-
solved without delay in warm (60 ЊC) water. By adjusting the
pH to 2 (HCl) a red-brown precipitate was obtained which was
extracted with ether (3 × 5 cm3), washed with water (10 cm3)
and dried (MgSO4). Removal of the solvent followed by flash
chromatography (Rf = 0.58) afforded Hdfcm (0.061 g, 7%) as
deep red needles after solvent removal, m.p. 157 ЊC (Found: Fe,
25.2. C23H20Fe2O2 requires 25.38%); ν max(KBr)/cmϪ1 1640 and
1710 (C᎐O); δ (300 MHz, CDCl ) 4.19 (10 H, s, 2 × C H ), 4.49
᎐
H
3
5
5
1
nitrile. The obtained pKa was within 99.5% of the value
obtained in pure water. The influence of 1,4-dioxane was much
(4 H, t, 2 × C5H4), 4.82 (4 H, t, 2 × C5H4) and 5.96 (1 H, s, enol
CH).
more pronounced. Water containing only 5% 1,4-dioxane
Hfctfa (2, R = CF3), sodium ethoxide method, an adaptation of
a published procedure.1 A suspension of solid sodium ethoxide38
(3.4 g, 50 mmol) in an ether solution (40 cm3) of acetylferrocene
(5.47 g, 24 mmol) was stirred for 15 min before ethyl trifluoro-
acetate (6.82 g, 48 mmol), prepared, purified and dried in the
same way as described for ethyl acetate,39 was slowly added. The
resulting orange-red precipitate was filtered off after 12 h of
stirring, washed with dry ether and dissolved in lukewarm (50–
60 ЊC) water. The aqueous solution was filtered without delay,
the pH immediately lowered to 2 with HCl and extracted with
ether (3 × 100 cm3). Washing of the ether extract with water
(3 × 100 cm3), followed by drying (NaSO4) and solvent removal
under pressure, afforded solid Hfctfa (4.22 g, 54%); m.p. 102 ЊC
(lit.,1 102 ЊC). Flash column chromatography (Rf = 0.48)
afforded spectroscopically pure Hfctfa (Found: Fe, 17.3.
1
caused a pKa drift for Hfctfa of more than 7%. Acetonitrile
1
induced a drift this large in the pKa of Hfctfa only when the
mixed solvent contained 50% acetonitrile (see text). It was best
1
to determine the pKa of Hfctfa and Hbfcm by titration from
low to high pH due to an unknown, but apparently semi-
reversible, side reaction taking place on the basic side of the pKa
of the ligand. Therefore Hfctfa and Hbfcm were first dissolved
in acetonitrile followed by addition of water and HClO4 to a
final mixed solvent constitution of 10% acetonitrile, pH 3.5
before finally titrating with NaOH.
Atomic absorption iron analysis
Samples were prepared by heating ferrocenyl-containing
compound (10–30 mg) with 27.5% nitric acid (0.8 cm3) and
water (1 cm3) to gentle simmering. The colour changed from
blue to orange-yellow. After 10 min, water (5 cm3) was added
and the new solution again heated to gentle simmering
for 30 min. The cooled solution was diluted to 10 cm3 in a
volumetric flask and the iron content (between 5 and 15 ppm)
of the solution determined against standard iron() solutions
of concentration 5, 10 and 15 ppm on a Varian SpectrAA-300
atomic absorption spectrophotometer fitted with coprocessor.
A blank (10 cm3 55% HNO3 diluted to 250 cm3) correction
was made.
C14H14F3FeO2 requires 17.23%); ν max(KBr)/cmϪ1 1620 (C᎐O);
᎐
δH(80 MHz, CDCl3) 4.20 (5 H, s, C5H5), 4.65 (2 H, t, C5H4),
4.85 (2 H, t, C5H4) and 6.07 (1 H, s, CH).
Hbfcm (2, R = phenyl), lithium diisopropylamide method.
Rigorous Schlenk conditions were adhered to. A light yellow
solution of LiNPri2 was prepared by adding n-butyllithium
(4.21 cm3 of a 1.6 mol dmϪ3 solution in hexane) to an ice-cooled
solution of freshly distilled diisopropylamine (0.73 g, 7.2 mmol)
in thf (15 cm3). This was added to a solution of acetylferrocene
(1.46 g, 6.4 mmol) in thf (10 cm3) and stirred at room temper-
ature for 20 min before methyl benzoate (0.81 g, 6 mmol)
dissolved in thf (10 cm3) was added. Stirring of the resulting
reaction mixture continued for 4 h before it was shaken with
HCl (50 cm3, 1 mol dmϪ3) and immediately extracted with ether
(5 × 80 cm3). The combined ether extracts were thoroughly
washed with water, dried (MgSO4) and evaporated to dryness
under reduced pressure. Flash chromatography of the residue
(Rf = 0.58) afforded Hbfcm (0.56 g, 28%), m.p. 107 ЊC (lit.,24
106 ЊC) (Found: Fe, 16.6. C19H16FeO2 requires 16.81%);
Syntheses
Acetylferrocene 1. This was prepared (in 81% yield) by treat-
ing ferrocene with acetic anhydride according to a published
procedure6b with care being taken to maintain the internal
temperature34 of the reaction mixture between 100 and 105 ЊC.
Recrystallisation from hexane gave a sufficiently pure product
for β-diketone syntheses.
ν max(KBr)/cmϪ1 1640 and 1710 (C᎐O); δ (300 MHz, CDCl )
᎐
H
3
Methyl ferrocenoate. Ferrocenoic acid was prepared via the
lithium intermediate as described elsewhere35 and esterified by
refluxing it (1.6 g, 7 mmol) in methanol (100 cm3) in the pres-
ence of concentrated H2SO4 (0.04 cm3) under nitrogen for 48 h.
The resulting liquid was poured onto ice (150 g) and extracted
with ether (3 × 100 cm3). The combined ether extracts were
washed with water, 5% aqueous NaHCO3 and again water to
afford, after removal of the dried (NaSO4) solvent, the ester
(1.19 g, 70%), m.p. 70 ЊC (lit.,36 70 ЊC); δH(80 MHz, CDCl3) 3.78
(3 H, s, CH3), 4.17 (5 H, s, C5H5), 4.35 (2 H, t, C5H4) and 4.76
(2 H, t, C5H4).
3.91 (0.1 H, s, keto CH2), 4.20 (5 H, s, C5H5), 4.54 (2 H, t,
C5H4), 4.88 (2 H, t, C5H4), 6.48 (0.95 H, s, enol CH), 7.41–7.51
and 7.88–8.11 (5 H, m, C6H5). Compound 3 was isolated
from the reaction mixture in yields up to 20%, Rf = 0.54,
m.p. 116 ЊC (Found: Fe, 26.1. C24H22Fe2O requires 26.21%);
ν max(KBr)/cmϪ1 1630 (C᎐O); δ [300 MHz, (CD ) CO] 2.55 (3
᎐
H
3 2
H, d, CH3), 4.19 (5 H, s, C5H5 on the alkenyl side), 4.22 (5 H, s,
C5H5 of ferrocenoyl), 4.46 (2 H, t, C5H4 on the alkenyl side),
4.57 (2 H, t, C5H4 of ferrocenoyl), 4.79 (2 H, t, C5H4 on the
alkenyl side), 4.88 (2 H, t, C5H4 of ferrocenoyl) and 6.89 (1 H, q,
CH).
Hfctca. The LiNPri2 method resulted in Hfctca (m.p. 52 ЊC,
16% yield). It took several days to solidify after solvent
removal. Eluent 2 was used to prevent acid induced degradation
of Hfctca during chromatography (Found: Fe, 14.8.
â-Diketones 2. Only one representative example is provided
for each of the three methods.
Hdfcm (2, R = ferrocenyl), sodamide method. Solid acetyl-
ferrocene 1 (0.456 g, 2 mmol) was slowly added during 30 min
with vigorous stirring to a suspension of NaNH2 (0.215 g, 5.5
mmol) in distilled37 ammonia (15 cm3). The resulting yellow
suspension was stirred for 15 min before freshly prepared dry
methyl ferrocenoate (0.976 g, 4 mmol) dissolved in dry ether
(10 cm3) was added dropwise. After stirring the cooled red
C14H11Cl3FeO2 requires 14.95%); ν max(KBr)/cmϪ1 1620 (C᎐O);
᎐
δH(300 MHz, CDCl3) 4.14 (0.1 H, s, keto CH2), 4.22 (5 H, s,
C5H5), 4.62 (2 H, t, C5H4), 4.86 (2 H, t, C5H4) and 6.35 (0.95 H,
s, enol CH).
Characterisation data for Hfca (Found: Fe, 20.6. C14H14FeO2
requires 20.68%); ν max(KBr)/cmϪ1 1620 (C᎐O); δ (80 MHz,
᎐
H
2512
J. Chem. Soc., Dalton Trans., 1998, Pages 2507–2514

View Article Online
CDCl3) 2.02 (2.64 H, s, enol CH3), 2.27 (0.36 H, s, keto CH3),
3.80 (0.24 H, s, keto CH2), 4.12 (5 H, s, C5H5), 4.46 (2 H, t,
C5H4), 4.74 (2 H, t, C5H4) and 5.69 (0.86 H, s, enol CH).
Method 2, by treating compound 5 with phen. Equimolar
amounts of the β-diketonate 5 and phen, each dissolved in the
minimum of acetone, were mixed. Addition of an excess of a
saturated solution of NaClO4 in acetone, precipitated 6 in ca.
64% yield.
Copper(II) complexes 4. Crude compounds 2 could be puri-
fied with good effect via copper() complexation. The general
procedure was as follows: to a solution of crude solid β-
diketone (up to 30 mmol) in acetone (20 cm3) was added cop-
per() acetate (10.7 g, 54 mmol, super saturated) and sodium
acetate (2.2 g, 2.7 mmol) dissolved in water (120 cm3). The
precipitate 4 that formed was filtered off after 45 min of stirring
at room temperature, thoroughly washed with water and dis-
solved in chloroform (80 cm3). Liberation of the free β-diketone
was accomplished by shaking the chloroform solution of 4 with
an equal volume of HCl (6 mol dmϪ3), washing with water and
evaporating to dryness under reduced pressure.
Acknowledgements
The authors gratefully acknowledge financial support from the
Foundation for Research and Development and the Central
Research Fund of the University of the Orange Free State.
Generous financial support over a 4 year period by Henkel
South Africa is also greatly acknowledged.
References
1 (a) C. J. Pedersen, N. J. Salem and V. Weinmayr, US Pat., 2 875 223,
1959; (b) V. Weinmayr, Naturwissenschaften, 1958, 45, 311.
2 (a) J. G. Leipoldt, S. S. Basson, G. J. van Zyl and G. J. J. Steyn,
J. Organomet. Chem., 1991, 418, 241; (b) J. G. Leipoldt and E. C.
Grobler, Transition Met. Chem. (Weinheim, Ger.), 1986, 11, 110; (c)
J. G. Leipoldt, G. J. Lamprecht and E. C. Steynberg, J. Organomet.
Chem., 1991, 402, 259.
3 (a) A. Roodt, J. G. Leipoldt, J. C. Swarts and G. J. J. Steyn, Acta
Crystallogr., Sect. C, 1992, 48, 547; (b) J. C. Swarts, T. G. Vosloo,
J. G. Leipoldt and G. J. Lamprecht, Acta Crystallogr., Sect. C, 1993,
49, 760; (c) G. J. Lamprecht, J. C. Swarts, J. Conradie and J. G.
Leipoldt, Acta Crystallogr., Sect. C, 1993, 49, 82; (d) C. Glidewell
and C. M. Zakaria, Acta Crystallogr., Sect. C, 1994, 50, 1673; (e)
A. Haaland and J. Nilsson, Chem. Commun., 1968, 88; ( f ) A. Yogev
and Y. Mazur, J. Org. Chem., 1967, 32, 2162.
Di-ì-chloro-bis(ç-cycloocta-1,5-diene)dirhodium(I). The com-
plex was prepared according to a published procedure.40 δH(300
MHz, CDCl3) 1.30 (4 H, m, half of 4CH2), 2.08 (4 H, m, half of
4CH2) and 4.42 (4 H, m, 4CH).
[Rh(â-diketone)(cod)] complexes 5. The general procedure
was as follows. To a stirred yellow, near saturated solution
of [Rh2Cl2(cod)2] (0.5 g, 1 mmol) in dmf (6 cm3) was added
solid β-diketone 2 (2 mmol). After 5 min of stirring the crude
product 5 was precipitated with an excess of water, filtered
off and dissolved in ether. The ether solution was washed
with water, dried (MgSO4) and evaporated to dryness. Flash
column chromatography gave 5 spectroscopically pure in high
yield.
[Rh(fctfa)(cod)]. Rf = 0.73; 87% yield. δH(300 MHz, CDCl3)
1.84 (4 H, m, half of aliphatic C8H12 protons), 2.49 (m, 4 H,
other half of aliphatic C8H12 protons), 4.18 (m, 9 H, four
olefinic protons of C8H12 and 5 H of C5H5), 4.45 (2 H, t, C5H4),
4.70 (2 H, t, C5H4) and 5.93 (1 H, s, CH).
[Rh(fctca)(cod)]. Rf = 0.73; 57% yield. δH(300 MHz, CDCl3)
1.84 (4 H, m, half of aliphatic C8H12 protons), 2.49 (4 H, m,
other half of aliphatic C8H12 protons), 4.18 (9 H, m, 4 olefinic
protons of C8H12 and 5 H of C5H5), 4.43 (2 H, t, C5H4), 4.68
(2 H, t, C5H4) and 6.39 (1 H, s, CH).
[Rh(fca)(cod)]. Rf = 0.79; 73% yield. δH(300 MHz, CDCl3)
1.75 (4 H, m, half of aliphatic C8H12 protons), 2.09 (3 H, s,
CH3), 2.38 (4 H, m, half of aliphatic C8H12 protons), 4.08 (5 H,
s, C5H5), 4.18 (2 H, t, C5H4), 4.45 (4 H, m, olefinic protons of
C8H12), 4.73 (2 H, t, C5H4) and 5.80 (1 H, s, CH).
4 (a) W. R. Cullen, S. J. Rettig and E. B. Wickenheiser, J. Mol. Catal.,
1991, 66, 251; (b) W. R. Cullen and E. B. Wickenheiser, J. Organomet.
Chem., 1989, 370, 141.
5 R. C. Mehrotra, R. Bohra and D. P. Gaur, Metal β-Diketonates and
Allied Derivatives, Academic Press, London, 1978, pp. 268–277.
6 For reviews see (a) M. Rosenblum, Chemistry of the Iron Group
Metallocenes: Ferrocene, Ruthenocene, Osmocene, Wiley, New York,
1965, part 1; (b) D. E. Bublitz and K. L. Rinehart, jun., in Organic
Reactions, ed. W. G. Dauben, Wiley, New York, 1969, vol. 17, p. 1;
(c) W. E. Watts, in Comprehensive Organometallic Chemistry, ed.
G. Wilkinson, Pergamon, Oxford, 1982, vol. 8, ch. 59, pp. 1013–
1071; (d) A. J. Deeming, in Comprehensive Organometallic
Chemistry, ed. G. Wilkinson, Pergamon, Oxford, 1982, vol. 4, ch. 31,
pp. 475–491.
7 A. N. Nesmeyanov and N. S. Kotchetkova, Russ. Chem. Rev., 1974,
43, 710.
8 A. S. Tompa, Thermochim. Acta, 1984, 77, 133.
9 B. S. Chittawadgi and A. N. Voinof, Indian J. Technol., 1968, 6, 83.
10 E. W. Neuse, J. R. Woodhouse, G. Montaudo and C. Puglisi, Appl.
Organomet. Chem., 1988, 2, 53.
11 J. M. Ward, M. E. Grabin, E. Berlin and D. M. Young, Cancer Res.,
1977, 37, 1238.
12 P. Köpf-Maier, H. Köpf and E. W. Neuse, J. Cancer Res. Clin.
Oncol., 1984, 108, 336; E. W. Neuse and F. Kanzawa, Appl.
Organomet. Chem., 1990, 4, 19.
13 T. Giraldi, G. Sava, G. Betoli, G. Mestroni and G. Zassinovich,
Cancer Res., 1977, 37, 2662; G. Sava, S. Zorzet, L. Perissin,
G. Mestroni, G. Zassinovich and A. Bontempi, Inorg. Chim. Acta,
1987, 137, 69.
14 S. E. Sherman and S. J. Lippard, Chem. Rev., 1987, 87, 1153.
15 M. Rozenswerg, D. D. van Hoff, M. Slavik and J. Chrisholm, Ann.
Int. Med., 1977, 86, 803.
[Rh(bfcm)(cod)]. Rf = 0.78; 77% yield. δH(300 MHz, CDCl3)
1.84 (4 H, m, half of aliphatic C8H12 protons), 2.49 (4 H, m,
other half of aliphatic C8H12 protons), 4.17 (9 H, m, 4 olefinic
protons of C8H12 and 5 H of C5H5), 4.35 (2 H, t, C5H4), 4.70
(2 H, t, C5H4), 6.25 (1 H, s, CH), 7.37 (3 H, m, C6H5) and 7.78
(2 H, m, C6H5).
[Rh(dfcm)(cod)]. Rf = 0.84; 79% yield. δH(300 MHz,
CDCl3) 1.86 (4 H, m, half of aliphatic C8H12 protons), 2.51
(4 H, m, other half of aliphatic C8H12 protons), 4.10 (4 H, m,
olefinic protons of C8H12), 4.15 (10 H, s, 2C5H5), 4.33 (4 H, t,
half of 2C5H4), 4.67 (4 H, m, half of 2C5H4) and 5.92 (1 H,
s, CH).
16 J. H. Burchenal, K. Kalaher, T. O’Toole and J. Chrisholm, Cancer
Res., 1977, 37, 3455.
17 W. Wolf and R. C. Manaka, J. Clin. Hemotol. Oncol., 1977, 7, 79.
18 J. Drobnik and P. Horacek, Chem.-Biol. Interact. I, 1973, 7, 223.
19 R. J. Woodsman, J. M. Venditti and S. A. Schepartz, Proc. Am.
Assoc. Cancer Res., 1971, 12, 24; G. R. Gale and C. M. Atkinds,
Bioinorg. Chem., 1978, 8, 445; H. W. Bruckner, C. C. Cohen and
G. Deppe, J. Clin. Hemotol. Oncol., 1977, 1, 619.
20 J. C. Swarts, E. W. Neuse, A. G. Perlwitz, A. S. Stephanou and
G. J. Lamprecht, Angew. Makromol. Chem., 1993, 207, 123.
21 G. R. Gale, L. M. Atkins and S. J. Meischen, Cancer Treat Rep.,
1977, 61, 445.
22 J. C. Swarts, E. W. Neuse and G. J. Lamprecht, J. Inorg. Organomet.
Pol., 1994, 4, 143.
23 A. I. Vogel, Practical Organic Chemistry including Qualitative
Organic Analysis, Longman, London, 3rd edn., 1977, pp. 861–865;
[Rh(cod)(phen)][ClO4] 6. Method 1, by treating [Rh2Cl2(cod)2]
with phen. The product obtained in each reaction according
to method 2 was the same (as judged by IR and H NMR
1
spectroscopy) as that obtained by allowing compounds
[Rh2Cl2(cod)2] and 5 to react with phen according to a liter-
ature procedure2b,40 to obtain 6. δH[300 MHz, (CD3)2SO]
2.15 (4 H, m, half of aliphatic C8H12 protons), 2.56 (4 H, m,
other half of aliphatic C8H12 protons), 4.78 (4 H, m, olefinic
protons of C8H12), 8.03 (2 H, dd, C12H8N2), 8.22 (2 H,
s, C12H8N2), 8.43 (2 H, dd, C12H8N2) and 8.89 (2 H, dd,
C12H8N2).
J. Chem. Soc., Dalton Trans., 1998, Pages 2507–2514
2513

View Article Online
1978, 40, 1387; S. Aygen, E. F. Paulus, Y. Kitamura and R. van
C. R. Hauser and B. E. Hudson, jun., in Organic Reactions, ed.
R. Adams, Wiley, New York, 1942, vol. 1, ch. 9, pp. 266–302; C. R.
Hauser, F. W. Swamer and J. T. Adams, in Organic Reactions, ed.
R. Adams, Wiley, New York, 1954, vol. 8, ch. 3, pp. 59–196.
24 C. R. Hauser and C. E. Cain, J. Org. Chem., 1958, 23, 1142; C. E.
Cain, T. A. Mashburn and C. R. Hauser, J. Org. Chem., 1961, 26,
1030.
25 R. T. Morrison and R. N. Boyd, Organic Chemistry, Allyn and
Bacon, Boston, 5th edn., 1987, (a) pp. 913–916; (b) p. 929.
26 W. Bell, J. A. Crayston, C. Glidewell, M. A. Mazid and M. B.
Hursthouse, J. Organomet. Chem., 1992, 434, 115.
27 J. J. C. Erasmus, G. J. Lamprecht, J. C. Swarts, A. Roodt and
A. Oskarsson, Acta Crystallogr., Sect. C, 1996, 52, 3000.
28 J. L. Burdett and M. T. Rogers, J. Am. Chem. Soc., 1964, 86, 2105.
29 (a) P. R. Wells, Prog. Phys. Org. Chem., 1968, 6, 111; (b) R. E.
Kagarise, J. Am. Chem. Soc., 1955, 77, 1377; (c) N. Inamoto and
S. Masuda, Chem. Lett., 1982, 1003; (d) W. Gordy, Phys. Rev., 1946,
69, 604.
30 (a) G. J. Palenic, Inorg. Chem., 1970, 9, 2424; Y. Luo, R. J. Barton
and B. E. Robertson, Acta Crystallogr., Sect. C, 1990, 46, 1388; (b)
A. C. Macdonald and J. Trotter, Acta Crystallogr., 1964, 17, 872;
F. H. Allen, J. Trotter and C. S. Williston, J. Chem. Soc. A, 1970,
907.
Eldik, Inorg. Chem., 1987, 26, 769; W. Purcell, J. G. Leipoldt and
Y. Kitamura, Inorg. Chim. Acta, 1990, 177, 151; (b) A. E. Martell,
Stability Constants of Metal-Ion Complexes, The Chemical Society,
London, Special Publication No. 25, 3rd edn., part II, 1971, p. 365.
33 (a) B. S. Furniss, A. J. Hannaford, P. W. G. Smith and A. R. Tatchell,
Vogel’s textbook of practical organic chemistry, Longman, Harlow,
5th edn., 1989, p. 622; (b) M. A. Bannett, M. I. Bruce and
T. W. Matheson, in Comprehensive Organometallic Chemistry,
ed. G. Wilkinson, Pergamon, Oxford, 1982, vol. 4, ch. 32, p. 765;
(c) L. Helm, MINSQ, Non-linear parameter estimation and
model development, least squares parameter optimisation V3.12,
MicroMath Scientific Software, Salt Lake City, UT, 1990.
34 H. Van Ryswyk and G. R. Van Hecke, J. Chem. Educ., 1991, 68, 878.
35 S. I. Goldberg, L. H. Keith and T. S. Prokopv, J. Org. Chem., 1963,
28, 850.
36 K. Schlögl, Monatsh. Chem., 1957, 88, 601.
37 B. S. Furniss, A. J. Hannaford, P. W. G. Smith and A. R. Tatchell,
Vogel’s textbook of practical organic chemistry, Longman, Harlow,
5th edn., 1989, pp. 116, 462.
38 B. S. Furniss, A. J. Hannaford, P. W. G. Smith and A. R. Tatchell,
Vogel’s textbook of practical organic chemistry, Longman, Harlow,
5th edn., 1989, 634.
31 (a) R. G. Wilkins, The study of kinetics and mechanism of reactions
of transition metal complexes, Allyn and Bacon, Boston, 1974, pp.
44–45; (b) A. Roodt, J. G. Leipoldt, L. Helm and A. E. Merbach,
Inorg. Chem., 1992, 31, 2864.
39 F. G. Mann and B. C. Saunders, Practical Organic Chemistry,
Longman, New York, 1981, p. 95.
40 J. Chatt and L. M. Venzi, J. Chem. Soc., 1957, 4735.
32 (a) M. Ellinger, H. Duschner and K. Starke, J. Inorg. Nucl. Chem.,
1978, 40, 1063; H. Duschner and K. Starke, J. Inorg. Nucl. Chem.,
Received 30th March 1998; Paper 8/02398K
2514
J. Chem. Soc., Dalton Trans., 1998, Pages 2507–2514