Synthesis and characterisation of ferrocene-containing β-diketonato complexes of rhodium(I) and rhodium(III)

Article abstract of DOI:10.1016/S0020-1693(01)00731-9

The synthesis of new β-diketonato rhodium(I) complexes of the type [Rh(FcCOCHCOR)(CO)₂] and [Rh(FcCOCHCOR)(CO)(PPh₃)] with Fc = ferrocenyl and R = Fc, C₆H₅, CH₃ and CF₃ are described. ¹H, ¹³C and ³¹P NMR data showed that for each of the non-symmetric β-diketonato mono-carbonyl rhodium(I) complexes, two isomers exist in solution. The equilibrium constant, K_c, which relates these two isomers in an equilibrium reaction, are concentration independent but temperature and solvent dependent. Δ_rG, Δ_rH and Δ_rS values for this equilibrium have been determined and a linear relationship between solvent polarity on the Dimroth scale and K_c exists. The relationship between Rh-P bond lengths, d(Rh-P), and ³¹P NMR peak positions as well as coupling constants ¹J(³¹P-¹⁰³Rh) has been quantified to allow calculation of approximate d(Rh-P) values. Variations in d(Rh-P) for [Rh(RCOCHCOR′)(CO)(PPh₃)] complexes have also been related to the group electronegativities (Gordy scale) of the terminal β-diketonato R groups trans to PPh₃. A measure of the electron density on the rhodium centre of [Rh(RCOCHCOR′)(CO)(PPh₃)] may be expressed in terms of the IR carbonyl stretching wave number, ν(CO), the sum of the group electronegativities of the R and R′ groups, (χ_R + χ_R′), or the observed pK′_a values of the free β-diketones RCOCH₂COR′. An empirical relationship between ν(CO) and either pK′_a or (χ_R + χ_R′) has also been quantified.

Full text of DOI:10.1016/S0020-1693(01)00731-9

www.elsevier.com/locate/ica
Inorganica Chimica Acta 328 (2002) 191–203
Synthesis and characterisation of ferrocene-containing b-diketonato
complexes of rhodium(I) and rhodium(III)
Jeanet Conradie, Gert J. Lamprecht, Stefanus Otto, Jannie C. Swarts *
Department of Chemistry, Uni6ersity of the Orange Free State, Bloemfontein 9300, South Africa
Received 5 April 2001; accepted 16 October 2001
Abstract
The synthesis of new b-diketonato rhodium(I) complexes of the type [Rh(FcCOCHCOR)(CO)2] and [Rh(FcCOCH-
1
COR)(CO)(PPh3)] with Fc=ferrocenyl and R=Fc, C6H5, CH3 and CF3 are described. H, 13C and 31P NMR data showed that
for each of the non-symmetric b-diketonato mono-carbonyl rhodium(I) complexes, two isomers exist in solution. The equilibrium
constant, Kc, which relates these two isomers in an equilibrium reaction, are concentration independent but temperature and
solvent dependent. DrG, DrH and DrS values for this equilibrium have been determined and a linear relationship between solvent
polarity on the Dimroth scale and Kc exists. The relationship between RhꢀP bond lengths, d(RhꢀP), and 31P NMR peak positions
1
as well as coupling constants J(31Pꢀ103Rh) has been quantified to allow calculation of approximate d(RhꢀP) values. Variations
in d(RhꢀP) for [Rh(RCOCHCOR%)(CO)(PPh3)] complexes have also been related to the group electronegativities (Gordy scale) of
the terminal b-diketonato R groups trans to PPh3. A measure of the electron density on the rhodium centre of [Rh(RCOCH-
COR%)(CO)(PPh3)] may be expressed in terms of the IR carbonyl stretching wave number, w(CO), the sum of the group
electronegativities of the R and R% groups, (R+R%), or the observed pK% values of the free b-diketones RCOCH2COR%. An
a
empirical relationship between w(CO) and either pK% or (R+R%) has also been quantified. © 2002 Elsevier Science B.V. All
a
rights reserved.
Keywords: Rhodium complexes; Ferrocene complexes; Diketonato complexes
platinum group metal catalysts [13]. The rhodium and
iodide-catalysed carbonylation of methanol to acetic
acid is to date probably the most successful example of
an industrial process homogeneously catalysed by a
metal complex in solution [12]. Leipoldt [14] and Roodt
[15] have previously shown that some b-diketonato
complexes of rhodium(I) can facilitate the formation of
acyl species during oxidative addition of methyl iodide
to complexes of the type [Rh(b-diketonato)(CO)(PPh3)].
We were interested to establish whether the ferrocene-
containing b-diketones, FcCOCH2COR [16], may en-
hance acyl formation in complexes of the type
[Rh(FcCOCHCOR)(CO)(PPh3)]. Towards this end, we
report with this communication the synthesis and char-
acterisation of a variety of new ferrocene-containing
complexes of the type [Rh(FcCOCHCOR)(CO)2] and
[Rh(FcCOCHCOR)(CO)(PPh3)]. The physical implica-
tions of some of their properties are also quantified.
1. Introduction
Platinum group transition metal complexes are well
known for their medical applications [1] and are often
employed as either homogeneous [2] or heterogeneous
[3] catalysts in
a variety of chemical reactions.
Rhodium-based homogeneous catalysts in particular
may be used inter alia in carbon–carbon bond forma-
tion [4], cyclopropanation [5], decarbonylation of alde-
hydes
[6],
hydroboration
of
olefins
[7],
hydroformylation of olefins [8], hydrogenation [9], hy-
drosilation [10] and carbon monoxide insertion reac-
tions
[11,12].
Hydroformylation
and
alcohol
carbonylation are major industrial processes involving
* Corresponding author. Tel.: +27-51-401 2781; fax: +27-51-430
7805.
E-mail address: swartsjc@cem.nw.uovs.ac.za (J.C. Swarts).
0020-1693/02/$ - see front matter © 2002 Elsevier Science B.V. All rights reserved.
PII: S0020-1693(01)00731-9

192
J. Conradie et al. / Inorganica Chimica Acta 328 (2002) 191–203
2. Experimental
solution through a sinter glass tube. The CO pressure in
the reaction vessel was maintained at approximately 1
cm above atmospheric pressure for 20 min. [Rh(b-dike-
tonato)(CO)2] was precipitated with cold water (100
ml), stirred for 15 min and centrifuged. The precipitate
was washed with water, filtrated and dried in a vacuum
dessicator. Spectroscopically pure product [Rh(b-diket-
onato)(CO)2] (2–5) was obtained in high (\75%) yield.
Selected characterisation data of complexes 2–5 may be
found in Tables 1 and 2, NMR data are as follows.
2.1. Materials and apparatus
Solid reagents used in preparations (Merck, Aldrich
and Sigma) were used without further purification.
Liquid reactants and solvents were distilled prior to use,
water was double distilled. DMF was dried according
to published methods [17]. Flash chromatography was
performed on Kieselgel 60 (Merck, grain size 0.063–0.2
mm, eluent ether–hexane 2:3 by volume unless other-
wise stated) utilising an overpressure that never ex-
ceeded 100 Torr (1 Torr=1 mmHg=133.32 Pa).
Melting points (m.p.) were determined with a Reichert
Thermopan microscope with a Koffler hot-stage and
are uncorrected. NMR measurements, at 298 K unless
otherwise stated, were recorded on a Bruker Advance
DPX 300 NMR spectrometer [1H (300 MHz), 13C
(75.476 MHz) and 31P (121.497 MHz)]. Chemical shifts
are reported as l values relative to SiMe4 (0 ppm) for
1
[Rh( fctfa)(CO)2] (2). H NMR (CDCl3): l 4.21 (5H,
s, C5H5), 4.64 (2H, t, C5H4), 4.86 (2H, t, C5H4) and 6.20
(1H, s, CH). 13C NMR (CDCl3): l 192.67 (s, CꢀFc),
165.82 (q, CꢀCF3, 2JCꢀF=32.6 Hz), 183.27 (d,
RhꢀCO,1JCꢀRh=73.8 Hz), 183.05 (d, RhꢀCO,1JCꢀRh
=
74.6 Hz). The crystal structure of 2 is as per ref. [21].
1
[Rh( fca)(CO)2] (3). H NMR (CDCl3): l 2.14 (3H, s,
CH3), 4.17 (5H, s, C5H5), 4.46 (2H, t, C5H4), 4.77 (2H,
t, C5H4) and 5.87 (1H, s, CH).
[Rh(bfcm)(CO)2] (4). The synthesis was performed
under Ar and in dry, freshly distilled DMF over 2.5 h,
Rf=0.70 (sefadex, 4 hexane+1 ethanol). 1H NMR
(CDCl3): l 4.15 (5H, s, C5H5), 4.49 (2H, t, C5H4), 4.85
(2H, t, C5H4), 6.50 (1H, s, CH), 7.4–7.5 (3H, m, C6H5)
and 7.9 (2H, m, C6H5).
1
the H spectra, relative to 85% H3PO4 at 0 ppm for the
31P spectra and referenced to CDCl3 (77.04 ppm) as
internal marker for 13C spectra. Infrared stretching
frequencies were recorded on a Hitachi 270-50 infrared
spectrometer either in a KBr matrix or in chloroform or
acetone solutions. Line fittings were performed with the
aid of the fitting program MINSQ [18].
1
[Rh(dfcm)(CO)2] (5). H NMR (CDCl3): l 4.18 (10H,
s, C5H5), 4.47 (4H, t, C5H4), 4.83 (4H, t, C5H4) and 6.18
(1H, s, CH).
2.2. Synthesis
2.2.2. [Rh(i-diketonato)(CO)(PPh3)] complexes (8–11)
The general procedure was as follows: To a boiling
solution of [Rh(b-diketonato)(CO)2] (0.2 mmol) in n-
hexane (30 cm3) was added a solution of PPh3 (0.2
mmol) in warm n-hexane (15 cm3). The resulting reac-
tion mixture was stirred for 30 s in a boiling water bath
until no more CO gas was released and filtered while
still warm. Pure crystals of the desired complexes were
obtained by slowly cooling the filtered reaction mixture.
The characterisation data of complexes 8–11 are as per
Tables 1 and 2, the crystal structure of 8 is as per ref.
[22]. 13C NMR (CDCl3) for 8: isomer 1, l 191.72 (s,
CꢀFc), 167.43 (q, CꢀCF3, 2JCꢀF=31.8 Hz), 189.00
The ferrocene-containing b-diketones Hfctfa, Hfca,
Hbfcm and Hdfcm [16], the [Rh(b-diketonato)(cod)]
complexes [16], and di-m-chloro-bis(h4-cycloocta-1,5-di-
ene) dirhodium(I), [Rh2Cl2(cod)2] [19], were prepared
according to published procedures.
2.2.1. [Rh(i-diketonato)(CO)2] complexes (2–5)
2.2.1.1. Route 1. The general procedure was as follows:
[Rh2Cl2(CO)4], (1), was prepared in situ by refluxing
RhCl3·3H2O (0.2 g, 0.76 mmol) in DMF (3 ml) until
the colour changed from red to yellow (ca. 30 min [20]).
The dimer-containing solution was allowed to cool on
ice and an equivalent amount of solid b-diketone (0.76
mmol) was slowly added while stirring. After 30 min of
stirring at r.t., the crude [Rh(b-diketonato)(CO)2], com-
plexes 2–4, were precipitated with an excess of water
and filtered, dried in air and recrystallised from hexane.
Purification of the bfcm complex 4 required flash chro-
matography of the reaction residue (Rf=0.70, sefadex,
4 hexane+1 ethanol). Complex 5 can only be obtained
in trace amounts utilising this route.
1
2
(d.d., RhꢀCO, JCꢀRh=79.3 Hz, JCꢀP=24.1 Hz); iso-
2
mer 2, l 190.34 (s, CꢀFc), 167.95 (q, CꢀCF3, JCꢀF
=
31.8 Hz), 189.07 (d.d., RhꢀCO, 1JCꢀRh=79.3 Hz,
2JCꢀP=24.1 Hz).
2.2.3. [Rh(i-diketonato)(CO)(PPh3)(CH3)(I)] complexes
(12–15)
The syntheses of [Rh(FcCOCHCOR)(CO)(PPh3)-
(CH3)(I)], 12, may serve as an example. Methyliodide
(1.9 g, 13.4 mmol, a 130 fold excess) was added to
[Rh(fctfa)(CO)(PPh3)] (8) (75 mg, 0.1 mmol) dissolved
in hexane (70 ml) at 30 °C. N2 was purged through the
reaction vessel before it was sealed and left at r.t. in the
dark. After 5 days crystals suitable for X-ray structure
2.2.1.2. Route 2. The general procedure was as follows:
[Rh(b-diketonato)(cod)] (7) (0.30 mmol) was dissolved
in acetone (80 ml). CO gas was purged through the

Table 1
31P and 1H NMR spectral parameters, IR carbonyl stretching wave numbers w(CO), RhꢀP bond lengths d(RhꢀP), equilibrium constants Kc and the percentage of the isomers (as illustrated in
Scheme 2) in CDCl3 at solutions at 298 K for selected [Rh(L,L%-BID)(CO)(PPh3)] and related rhodium(III) complexes
a
d
31P (ppm) 1J (Hz) b l1H (ppm)
I (%) c Kc
w(CO) (cm−1
)

tot
Atom trans to
PPh3
d(RhꢀP) (A)
References f
,
Compound
numbers
b-Diketonato
Isomer
l
e
numbers a
8
fctfa
1
2
1
48.04
48.04
49.05
176.4
176.4
193.9
7.4–7.8, 6.08,
4.83, 4.47, 4.22
7.4–7.8, 6.09,
4.25, 4.06, 3.98
7.38–7.78, 5.73, 19
4.79, 4.37, 4.19,
1.71
40
60
0.68 1986
1986
4.88
4.21
O
O
O
2.232(1)
[22]
9
fca
0.23 1980
2
47.53
174.0
7.38–7.78, 5.78, 83
4.15, 4.08, 3.91,
2.19
1980
O
10
bfcm
dfcm
1
2
49.28
47.43
47.36
179.0
176.7
175.6
7–8, 6.46, 4.90, 36
4.43, 4.23
7–8, 6.44, 4.21, 64
4.18, 3.95
0.56 1977
1977
4.08
3.74
O
O
O
/
11
12
7.40–7.82, 6.11,
4.87, 4.41, 4.23,
4.18, 4.16, 3.94
7.25–7.50, 5.31, 82
4.77, 4.48, 4.40,
4.30, 4.30, 1.78
7.25–7.50, 5.45, 18
4.80, 4.60, 4.38,
4.24, 4.24, 2.15
7.30–8.20, 5.83, 66
4.53, 4.60, 4.27,
1.50
1977
fctfa
1
2
1
29.33
27.09
30.10
33.31
116.5
115.2.
124.5
120.1
2056
2056
2077
2.319(3)
[23]
Rh(III)ꢀalkyl g
328
13
14
fca
(
Rh(III)ꢀalkyl g
1
2
1
7.30–8.20, 5.83, 34
4.40, 4.33, 1.42
2077
2078
203
bfcm
7.0–8.3, 6.55,
4.4–5.2, 4.30,
1.59
7.0–8.3, 6.48,
4.4–5.2, 4.36,
1.50
65
35
Rh(III)ꢀalkyl g
2
2078
2078
15
dfcm
7.35–8.20, 6.15,
4.92, 4.79, 4.68,
4.37, 4.41-4.45,
4.29, 1.50
Rh(III)ꢀalkyl g
16
17
tftma
tfdma
1
2
1
47.85
47.85
47.75
172.9
172.9
178.0
7.3–8, 6.04, 0.70 42
7.3–8, 6.04, 1.26 58
0.72 1992
1992
1.00 1984
O
O
O
2.238(3)
2.239(2)
[35]
[36]
7.3–8, 5.88,
50
2.30, 0.60

Table 1 (Continued)
31P (ppm) 1J (Hz) b l1H (ppm)
I (%) c Kc
w(CO) (cm−1
)

tot
Atom trans to
PPh3
d(RhꢀP) (A)
References f
a
d
,
Compound
numbers
b-Diketonato
Isomer
l
e
numbers a
2
48.72
175.7
7.3–8, 5.89,
50
1984
O
2.71, 1.20
18
tta
1
2
1
2
1
47.78
47.84
49.28
49.37
49.14
176.7
177.7
175.8
174.8
177.2
47
53
32
68
40
O
O
O
O
O
[30]
[37]
[38]
2.245(3)
2.248(3)
2.249(3)
2.252(3)
19a
19b
20
ba
7–8, 6.08, 1.53
7–8, 6.13, 2.23
7.3–8, 5.85,
2.05, 0.56
0.47 1980
1980
0.67 1986
4.55
tfhd
2
1
2
47.86
49.57
47.72
174.8
177.7
176.9
7.3–8, 5.89,
2.50, 1.20
7.3–7.8, 5.84,
1.75
7.3–7.8, 5.88,
2.25
7–8, 6.57
60
35
65
1986
0.54 1983
1983
O
O
O
21
22
tfaa
tfba
5.35
5.22
2.231(3) h
[38]
1
2
49.94
48.06
49.56
48.84
175.8
177.0
177.9
175.7
50
50
1.00 1983
1983
O
O
O
O
/
7–8, 6.57
23
24
dbm
acac
7.1–8.1, 6.85
7.8–7.3, 5.45,
2.11, 1.62
1979
1983
4.42
4.68
2.237(7)
2.244(2)
[39]
[40]
25
26
27
28
bzaa
sacac
salnr
49.44
35.36
175.0
144.5
O
S
N
N
O
2.243(1)
2.300(2)
2.281(2)
2.275(1)
[41]
[42]
[43]
[44]
dmavk
1
2
41.45
54.91
41.25
149.7
172.0
152.2
29
30
dmavk
(Rh(III)ꢀacyl)
dmavk
(Rh(III)ꢀalkyl)
cacsm
macsm
2.260(4)
2.356(3)
[45]
[32]
328
21.35
45.2
107.1
144.6
148.9
(
31
32
33
1944
1956
1965
N
N
S
2.268(1)
2.269(1)
2.283(1)
[32]
[46]
[32]
1
hacsm
1
42.70
203
L,L%-BID, ligand containing co-ordinating donor atoms L and L% that forms a six-membered pseudoaromatic ring when co-ordinated to the Rh nucleus; tftma (1,1,1-trifluoro-5,5-dimethyl-2,4-hex-
anedionato); tfdma (1,1,1-trifluoro-5-methyl-2,4-hexanedionato); tfhd (1,1,1-trifluoro-2,4-hexanedionato); tfaa (1,1,1-trifluoro-2,4-pentanedionato); tfba (1,1,1-trifluoro-4-phenyl-2,4-butanedion-
ato); dbm (1,3-diphenyl-1,3-propanedionato); acac (2,4-pentanedionato); tta (thenoyltrifluoroacetonato); and bzaa (3-benzyl-2,4-pentanedionato). Other abbreviations are defined in the abstract
or Fig. 5.
a Only in compounds 8–10 do isomers 1 and 2 refer to Scheme 2. Compound 19 fits Scheme 2 when R=CH3 and the ferrocenyl group is replaced with a phenyl group, 16–18 and 20–22 fits
Scheme 2 if R=CF3 and if the ferrocenyl group is replaced with the appropriate group for the specified b-diketonato ligand. In all cases, isomer 1 represents the expected isomer utilising electronic
arguments as described in refs [22,30,31].
b For 1J(31Pꢀ103Rh) couple. 31P NMR parameters for 18, 26, 28–31, and 33 are from Ref. [32].
c Percent I=% isomer at 298 K.
d

tot=(R+R%) on the Gordy scale,  and  values are as per Refs [16,34].
R R%
e Donor atom of the co-ordinating ligand bound to the rhodium nucleus trans to PPh3.
f Refs for d(RhꢀP).
g During oxidative addition of CH3I, at least two alkyl and two acyl species were detected. From the kinetics study that we performed (results still to be published) the isolated products as
well as the crystal structure of 12 and the cited NMR data corresponds to the dominant alkyl species that formed.
h d(RhꢀP) for the P(p-C6H5C1)3 complex. A second molecule with d(RhꢀP)=2.231(2) A was also observed.
,

Table 2
Characterisation data of ferrocene-containing mono- and dicarbonyl rhodium complexes
a
/
Complex
Yield (%)
M.p. (°C)
Percent found (requires)
w(CO) (cm−1
)
Complex b
Yield (%)
M.p. (°C)
Percent found (requires)
C
H
C
H
(2) [Rh(fctfa)(CO)2] c
(3) [Rh(fca)(CO)2]
(4) [Rh(bfcm)(CO)2]
(5) [Rh(dfcm)(CO)2]
66; 83
45; 79
38; 75
B5; 77
228
164
205
\255
39.8(39.9)
44.8(44.9)
51.6(51.5)
50.1(50.2)
2.1(2.1)
3.0(3.1)
3.1(3.1)
3.3(3.2)
2008; 2074
2014; 2068
1998; 2074
2008; 2074
(8) [Rh(fctfa)(CO)(PPh3)] d
(9) [Rh(fca)(CO)(PPh3)]
(10) [Rh(bfcm)(CO)(PPh3)]
(11) [Rh(dfcm)(CO)(PPh3)]
90
72
93
43
256
184
165
188
55.2(55.3)
60.0(59.9)
62.8(63.0)
60.5(60.6)
3.6(3.5)
4.4(4.3)
4.1(4.2)
4.2(4.1)
a The first percent is for route 1, the second for route 2.
328
−1
b w(CO) values for 8–11 is in Table 1. For 12, [Rh(FcCOCHCOCF3)(CO)(PPh3)(CH3)(I)], d(RhꢀCcarbonyl)=1.834(9) A, w(CO)=2056 cm
.
,
)
c
,
RhꢀC bond lengths are 1.83(2) and 1.84(2) A (ref. [21]).
d
,
RhꢀC bond length is 1.801(5) A (ref. [22]).
9
203

196
J. Conradie et al. / Inorganica Chimica Acta 328 (2002) 191–203
determination [23] were obtained from the reaction
mixture. Characterisation data of these complexes are
as per Table 1.
the product [Rh(b-diketonato)(CO)(PPh3)]. The fer-
rocene-containing complexes 2–5 and 8–11 were all
acid sensitive. They decomposed on silicagel and NMR
studies had to be performed in CDCl3 which were
passed through basic alumina before use. The rhodi-
um(I) complexes 8–11 all undergo oxidative addition
with CH3I to give [Rh(b-diketonato)(CO)(CH3)(I)-
(PPh3)] (complexes 12–15) with b-diketonato previ-
ously defined. Complex 12 was isolated and stable in
the solid state for months on end, but in solution IR
spectroscopy indicated that CO insertion to an acyl
species occurs at a slow rate.
3. Results and discussion
Two routes were utilised to obtain new dicarbonyl
rhodium(I) complexes of general formula [Rh(Fc-
COCHCOR)(CO)2]. The first consisted of the direct
interaction between [Rh2(Cl)2(CO)4] (1) and the b-dike-
tones ferrocenoyltrifluoroacetone (Hfctfa), ferrocenoy-
lacetone (Hfca), benzoylferrocenoylmethane (Hbfcm)
or diferrocenoylmethane (Hdfcm) (route 1, Scheme 1)
to give the red complexes [Rh(FcCOCHCOR)(CO)2] (2)
(R=CF3, 66% yield), (3) (R=CH3, 45%), (4) (R=
C6H5, 38%) and (5) (R=Fc, impure trace amounts).
To improve yields of 5, an alternative synthetic method
involving the cyclooctadiene (cod) dimer of rhodium(I)
was employed. In this alternative route, [Rh2(Cl)2(cod)2]
(6) was converted to [Rh(dfcm)(cod)] (7) which in an
equilibrium process can undergo carbon monoxide ex-
change to liberate, in the presence of an excess CO,
pure 5 in high (77%) yields (route 2, Scheme 1). To
prevent 5 converting back to 7, it proved useful to
precipitate it from the CO-saturated solvent, acetone,
with water. This procedure results in liberated cod
drifting as an oil on top of the aqueous phase while the
desired Rh complex settles on the bottom and allows
separation of 5 from the reaction mixture without
regeneration of starting material, 7, in an equilibrium
reaction during work up. Complexes 2–4 were also
obtained efficiently (\75% yield) in this manner.
The new deep-red [Rh(b-diketonato)(CO)(PPh3)]
complexes 8–11 with b-diketonato=fctfa (8), fca (9),
bfcm (10) and dfcm (11), respectively, were obtained by
reacting an equivalent amount of PPh3 with [Rh(b-dike-
tonato)(CO)2] in a n-hexane medium. This high-yield
reaction (up to 92%) is immediate with precipitation of
1H NMR spectra showed that for each of the [Rh(b-
diketonato)(CO)(PPh3)] complexes 8–10 two isomers
exist in solution (Fig. 1 and Scheme 2). Only 11, which
has a symmetrical b-diketonato ligand co-ordinated to
the rhodium(I) nucleus, exists as a single species in
solution. Regarding the two possible isomers, we chose
to label the isomer with CO trans to the oxygen atom
nearest to the more electron-donating ferrocenyl group
[16] of the b-diketonato chelate ring as isomer 1, while
the other isomer is labelled isomer 2 (Scheme 2). Fo-
1
cussing on the positions of the signals of the H NMR
spectrum of [Rh(fca)(CO)(PPh3)] between 3.9 and 4.9
ppm (Fig. 1), the set of signals belonging to an isomer
in the more downfield position is assigned to isomer 1.
The other set of signals belongs to isomer 2. This
assignment was done utilising the following arguments:
It is known that electron-withdrawing substituents on
ferrocene derivatives move the position of the proton
signals of the substituted cyclopentadienyl ring
downfield [24]. Acetylferrocene, by way of example,
exhibits signals at 4.78 ppm (2H, C5H4), 4.52 ppm (2H,
C5H4) and 4.21 ppm (5H, C5H5), compared with the
position of the singlet signal of ferrocene itself at 4.18
ppm. Conversely, any electron-donating substituents
move the signal of the protons of both cyclopentadienyl
rings to a more upfield position. For example, ethylfer-
rocene exhibits signals at 4.01–4.05 ppm (4H, C5H4)
Scheme 1. Synthetic routes for the synthesis of the [Rh(b-diketonato)(CO)2] complexes 2–5 and [Rh(b-diketonato)(CO)(PPh3)] complexes 8–11
from RhCl3·3H2O. Route 2 gave higher yields of 2–5. Oxidative addition of CH3I to 8–11 to yield 12–15 was shown by 1H NMR to be a slow
equilibrium reaction. If 12–15 are not isolated in the solid state, carbonyl insertion to form an acyl species (not shown) takes place over several
days. Fc, ferrocenyl, C10H9Fe.

J. Conradie et al. / Inorganica Chimica Acta 328 (2002) 191–203
197
Fig. 1. 1H NMR (left, referenced to SiMe4 at 0 ppm) and 31P NMR spectra (right, referenced to 85% H3PO4 at 0 ppm) of [Rh(fca)(CO)(PPh3)]
(9) in CDCl3 illustrate the existence of two isomers for this complex. Isomers 1 and 2 as well as Kc are defined in Scheme 2. 1H NMR assignments:
isomer 1: l 1.71 (3H, s, CH3), 4.19 (5H, s, C5H5), 4.37 (2H, t, C5H4), 4.79 (2H, t, C5H4), 5.73 (1H, s, CH) and 7.38–7.78 (15H, m, 3×C6H5);
isomer 2: l 2.19 (3H, s, CH3), 3.91 (5H, s, C5H5), 4.08 (2H, t, C5H4), 4.15 (2H, t, C5H4), 5.78 (1H, s, CH) and 7.38–7.78 (15H, m, 3×C6H5).
31P NMR assignments: isomer 1: l 49.05 (d, JPꢀRh=193.9 Hz); isomer 2: l 47.53 (d, JPꢀRh=174.0 Hz).
nato)(CO)(PPh3)] complexes 16–25, obtained Kc values
are summarised in Table 1. It is notable that complexes
11 and 23–25, which have symmetrical b-diketonato
ligands, have only one isomer in solution. 31P NMR
also differentiate between isomers 1 and 2 (Fig. 1 and
Table 1). The time needed for each complex to reach
the solution equilibrium position according to Scheme 2
varied. For both the new complexes [Rh(fctfa)(CO)-
(PPh3)] (8) and [Rh(bfcm)(CO)(PPh3)] (10) the equi-
librium set in quickly (less than 115 s, the time required
to dissolve the complex and to record data for a H
NMR spectrum), but [Rh(fca)(CO)(PPh3)] (9) needed
10 min for the equilibrium to set in.
The ratio of isomers, i.e. Kc, is independent of con-
centration in the concentration range 0.0028–0.0200
mol dm−3, but temperature and solvent dependent.
The variation of the Kc with temperature for the equi-
librium shown in Scheme 2 may be mathematically
quantified by
and 4.08 ppm (5H, C5H5), respectively [24]. Utilising
these observations, the influence of the electron-with-
drawing CO and the electron-donating PPh3 groups [25]
(relative to each other) on the b-diketonato carbonyl
group trans to the O-atom nearest to the ferrocenyl
moiety in isomers 1 and 2 (Scheme 2) can be evaluated.
In the case of isomer 2, the PPh3 group donates elec-
tron density via a d-orbital s bond to the rhodium
nucleus and accordingly also to the oxygen trans to
PPh3 and nearest to the ferrocenyl group, while for
isomer 1, the CO group withdraws electron density
from the carbonyl group neighbouring the ferrocenyl
moiety in the b-diketonato ligand. The net result of this
electron withdrawing and electron donating action is a
relatively higher electron density on the ferrocenyl
group of isomer 2 than on the ferrocenyl group of
isomer 1. Therefore, the higher electron density in the
ferrocenyl group of isomer 2 manifest in 1H NMR peak
positions of its protons in a more upfield position
compared with those of isomer 1. By applying the
above concepts to each of the complexes 8–10, an
equilibrium constant defined as Kc=[isomer 1]/[isomer
2] and applicable to the equilibrium shown in Scheme 2,
may be determined by calculating the ratio of peak
integral values of non-overlapping corresponding sig-
nals of each isomer and averaging all answers. For
[Rh(fca)(CO)(PPh3)], Kc, may be calculated from the
peak integrals of the methine proton of the b-diketo-
nato ligand fca, giving Kc=0.22/1.00=0.22. Other
combinations of peak integration result in Kc=0.46/
2.15=0.21 or 0.49/2.12=0.23 or 1.11/5.03=0.22 (inte-
gral values were taken from the NMR shown in Fig. 1).
The average of these values is Kc=0.22. The described
1H NMR technique to determine Kc values was also
1
D H
R
1
1
ꢀ ꢁ
r
ln Kc2=ln Kc1−
−
T2 T1
with Kc2 and Kc1 the equilibrium constants at tempera-
tures T2 and T1, R=8.314 J K−1 mol−1 and DrH the
Scheme 2. Equilibrium between the two isomers of unsymmetrical
complexes of the type [Rh(FcCOCHCOR)(CO)(PPh3)] with R=CF3,
CH3 and C6H5. Kc=[isomer 1]/[isomer 2].
applied to
a
series of other [Rh(b-diketo-

198
J. Conradie et al. / Inorganica Chimica Acta 328 (2002) 191–203
Fig. 3. Linear dependence of the equilibrium constant Kc at 298 K for
the equilibrium between the isomers of the [Rh(fctfa)(CO)(PPh3)]
complex in different solvents as a function of solvent polarity, ET(30).
Polarity data (on the Dimroth scale) are from ref. [29].
Fig. 2. Temperature dependence of Kc for the equilibrium (Scheme 2)
between the two isomers of [Rh(fctfa)(CO)(PPh3)] (8) ꢀ,
[Rh(fca)(CO)(PPh3)] (9) “, and [Rh(bfcm)(CO)(PPh3)] (10) ".
1H NMR spectra also distinguish between at least
two main isomers, see Table 1, for the [Rh(b-diketo-
nato)(CO)(CH3)(I)(PPh3)] complexes 12–15. However,
a thermodynamic study to determine thermodynamic
data for the rhodium(III) isomers was not undertaken
as these solutions were not stable. After 24 h, the
existence of a slow equilibrium between the alkyl rhodi-
um(III) complexes 12–15, acyl rhodium(III) complexes
of the type the [Rh(b-diketonato)(COCH3)(I)(PPh3)]
(not shown in Scheme 1) and the initial reactant rhodi-
um(I) complexes 8–11 became apparent. Effects of
temperature variations on the value of Kc between
different isomers of complexes 12–15 will therefore not
be uniquely attributable to only one equilibrium step,
but will be a composite of effects due to the multiple
equilibria observed for these solutions.
It is noteworthy that the equilibrium constants, Kc in
CDCl3, summarised in Table 1, favour isomer 2 in
Scheme 2, yet the crystal structure of [Rh(fctfa)(CO)-
(PPh3)] that was previously reported [22] was that of
isomer 1. In our previous communication [22] we ar-
gued that isomer 2 should be the dominant isomer
based on electronic considerations [30] (trans influence)
according to the polarisation theory put forward by
Grinberg [31]. When we found in the structure determi-
nation that isomer 1 was isolated, we attributed the
unexpected result to steric implications of the transition
state during substitution of the carbonyl group in 2
with PPh3. This study, however, reveals that there is a
reaction enthalpy as defined elsewhere [26,27]. The
above equation also implies that a graph of ln Kc versus
1/T should be linear [27], with slope −DrH/R. Fig. 2
illustrates this linearity in CDCl3 for [Rh(fctfa)-
(CO)(PPh3)], [Rh(fca)(CO)(PPh3)] and [Rh(bfcm)(CO)-
(PPh3)]. DrH values at 298 K in CDCl3 for the [Rh(b-
diketonato)(CO)(PPh3)] complexes 8–10 are sum-
marised in Table 3. The thermodynamic quantities
Gibbs free energy, DrG, and reaction entropy, DrS,
calculated [27] from the equations DrG= −RT ln Kc
and DrG=DrH−TDrS at 298 K are also presented. It
is instructive to note that the large temperature depen-
dence of Kc found in this study appears to be a direct
consequence of the presence of the ferrocenyl group in
our ligands. In contrast to what we found for our
ferrocene-containing complexes, temperature studies by
Trzeciak [28] and others on non-ferrocene-containing
b-diketonato complexes noted that temperature does
not influence the position of the equilibrium between
1
different isomers of this type of complexes much. A H
NMR study of the equilibrium position between the
two isomers of [Rh(fctfa)(CO)(PPh3)] in different sol-
vents revealed that Kc is dependent on the polarity of
the solvent, ET(30), as measured on the Dimroth [29]
scale. The observed relationship between ET(30) and Kc
(Fig. 3) is empirical and must at least in part reflect
improved solvation of [Rh(fctfa)(CO)(PPh3)] by solvent
molecules of higher solvent polarity.
Table 3
Equilibrium constants Kc at 298 K and in CDCl3 for the equilibrium shown in Scheme 2 as well as the thermodynamic data, also at 298 K,
relevant to this equilibrium for [Rh(b-diketonato)(CO)(PPh3)] complexes
Complex
Kc
DrH (kJ mol−1
)
DrG (kJ mol−1
)
DrS (J mol−1 K−1
)
[Rh(fctfa)(CO)(PPh3)]
[Rh(fca)(CO)(PPh3)]
[Rh(bfcm)(CO)(PPh3)]
0.68(2) a
0.23(3)
0.56(3)
6.2(3)
7.6(3)
8.6(3)
0.97(4)
3.0(1)
1.4(3)
17.6(6)
15(1)
24(2)
a Kc for the fctfa complex varied as follows in the deuterated solvents (solvent polarity [29] are given in brackets): cyclohexane, 1.03 (30.9);
benzene, 0.86 (34.3); acetone, 0.33 (42.2); DMSO, 0.20 (45.1) and acetonitrile, 0.16 (54.6).

J. Conradie et al. / Inorganica Chimica Acta 328 (2002) 191–203
199
solvents such as acetone or acetonitrile (Fig. 3). Unfor-
tunately, to date, we were not successful in isolating
crystals suitable for crystallography from more polar
media.
fast equilibrium between isomers 1 and 2, which largely
invalidates our previous explanation. Previously, we
have not given consideration to the possibility that the
observed isomer simply crystallised because it had a
more favourable crystallisation energy. The isomer that
dominates in solution will definitely be the one that is
thermodynamically more stable in solution. It is, how-
ever, also true that the isomer that dominates in solu-
tion will not necessarily be the one that possesses the
most favourable crystallisation energy. The solvent
dominant isomer will therefore not necessarily be the
isomer that crystallises first from solution. Factors that
may influence crystallisation energy include solvent
temperature and solvent polarity. It will also be af-
fected by crystal packing in the solid state that will be
dependent on the relative orientations of the PPh3 and
ferrocenyl groups. Our previous explanation for the
observed structure would only be correct if the ob-
tained isomer also possesses the lowest energy of crys-
tallisation in the particular solvent (hexane) employed
and at the experimental temperature at which the crys-
tal was grown. It is, however, interesting to note that
the polarities [29] of hexane, 31.0, and cyclohexane,
30.9, are virtually identical, and that isomer 1 actually
is present in slight excess in cyclohexane solutions
(Kc,cyclohexane=1.03). This implies that, for the fctfa
complex, the isomer which is thermodynamically fa-
voured in solution may also be favoured in the solid
state, and it would explain why we isolated isomer 1 of
8 from hexane for our structure determination and not
isomer 2. In principle, this begs the question if isomer 2
may be obtained in crystalline form from more polar
In previous communications [32], Roodt et al. re-
ported that there exists an approximately linear rela-
tionship between RhꢀP bond distances, d(RhꢀP), and
(a) chemical shift positions, l31P (ppm), as well as (b)
1
coupling constants, J(31Pꢀ103Rh), in 31P NMR spectra
of rhodium complexes of the type [Rh(L,L%-
BID)(CO)(PPh3)] with L,L%-BID=a symmetric or
asymmetric bidentate ligand, and L or L% the coordinat-
ing donor atoms O, S or N that co-ordinates to the
rhodium nucleus. The complex [Rh(fctfa)(CO)(PPh3)]
(8) prepared in this study belongs structurally to the
series reviewed by Roodt [32] but represents the first
example in which the co-ordinating ligand, fctfa, con-
tains the organometallic ferrocenyl group. We have
found that the 31P NMR parameters of 8 also fall on
the linear trend observed as shown in Fig. 4. Fig. 4
clearly illustrates that larger d(RhꢀP) values are associ-
1
ated with smaller J(31Pꢀ103Rh) coupling constants as
well as smaller l31P peak positions, in ppm. A fit on the
data of Fig. 4 implies that approximate d(RhꢀP) values
for the class of compounds under discussion may be
calculated utilising the equations
d(RhꢀP)= −0.0014(1)×1J(31Pꢀ103Rh)+2.49(2)
= −0.0041(3)×l31P+2.44(1)
These two equations are applicable for d(RhꢀP) values
,
between 2.23 and 2.36 A. Utilising these equations, it
follows from 31P NMR data (Table 1) that the average
RhꢀP bond length for isomers 1 and 2 of the fca and
bfcm complexes 9 and 10 should be in the order of 2.24
,
and 2.25 A, respectively. The RhꢀP distance of the
dfcm complex, 11, is calculated to be approximately
,
2.25 A. The errors on the derived equations prohibit
meaningful more accurate bond lengths. The data in
1
Table 1 indicate that coupling constants, J(31Pꢀ103Rh),
in complexes with PPh3 trans to a N or a S atom are
approximately equal in size, but appreciably smaller
1
than J(31Pꢀ103Rh) values for complexes in which PPh3
is trans to an O atom. RhꢀC bonds distance in these
square-planar complexes follow a similar trend as the
1
RhꢀP bonds when compared with J(13Cꢀ103Rh) values,
but is much less accurate for reasons adequately ex-
plained elsewhere [32].
The RhꢀP bond distance in square-planar complexes
of the type [Rh(L,L%-BID)(CO)(PPh3)] is also a good
indicator of the relative trans influence (a ground state
thermodynamic phenomenon that is defined as the abil-
ity of a ligand to weaken the metal–ligand bond trans
to it) of the donor atoms L and L% in the bidentate
ligand (L,L%-BID) [32]. Atoms co-ordinated to the
rhodium nucleus with a larger trans influence will
Fig. 4. The linear relationship between the coupling constants,
1J(31Pꢀ103Rh), or the position of the phosphorous signal, l31P, and
the RhꢀP bond distance, d(RhꢀP) of the indicated rhodium phos-
phine complexes. Data and complex numbers are as per Table 1.

200
J. Conradie et al. / Inorganica Chimica Acta 328 (2002) 191–203
Fig. 5. Structures of two isomers of [Rh(ba)(CO)(PPh3)] (19a) and (19b); [Rh(sacac)(CO)(PPh3)] (26); [Rh(salnr)(CO)(PPh3)] (27); [Rh(d-
mavk)(CO)(PPh3)] (28); [Rh(cacsm)(CO)(PPh3)] (31); [Rh(macsm)(CO)(PPh3)] (32) and [Rh(hacsm)(CO(PPh3)] (33). Ba, benzoylacetone; sacac,
thioacetylacetonato; salnr, N-o-tolylsalicylaldiminato; dmavk, dimethylaminovinylketonato; cacsm, methyl(2-cyclohexylamino-1-cyclopentene-1-
dithiocarboxylato); macsm, methyl(2-methyl-amino-1-cyclopentene-1-dithiocarboxylato); hacsm, methyl(2-amino-1-cyclopentene-1-dithiocarboxy-
lato).
lengthen the bond, here the RhꢀP bond, trans to it. For
the cited complexes in Table 1 the RhꢀP bond dis-
tances, d(RhꢀP), vary in the order
with the Pauling atomic electronegativity of O (3.5), N
(3.0) and S (2.5). It is evident that there is a direct
relationship between d(RhꢀP) and the co-ordinating
atoms of a ligand trans to the PPh3 group provided that
the complexes are structurally related. This explains
why d(RhꢀP) for the hacsm complex 33, an N,S co-or-
d(RhꢀP) trans to O atoms
Bd(RhꢀP) trans to N atoms
Bd(RhꢀP) trans to S atoms
which implies S has a larger trans influence than N,
while O has a smaller trans influence than N. One way
of quantifying this observed tendency is to establish a
relationship between atomic electronegativities of O, N
and S, utilising the Pauling scale, and the observed
RhꢀP bond lengths trans to these co-ordinating atoms.
To establish this relationship, the compounds consid-
ered need to be structurally related to each other and
d(RhꢀP) should be as accurate as possible. To satisfy
these criteria, we considered three classes of O,O lig-
ands, two classes of N,O ligands and the only available
example of an S,O class of ligands. For the O,O
ligands, i.e. b-diketonates, the first subclass is repre-
sented by the complexes 8, 16 and 17 and contains the
CF3 group adjacent to the carbonyl group trans to
PPh3. The second subclass has a CH3 group in this
position, compounds 19a, 21, 24 and 25 are representa-
tive of this type. The third subclass in the O,O class has
a phenyl group in this position and compound 19b
represents the only suitable example of this subclass.
The two available examples of N,O family of ligands
are represented by the salnr and dmavk complexes 27
and 28, while the only suitable available example of the
family of complexes having a S,O ligand is the sacac
complex 26 (Fig. 5 and Table 1). The average d(RhꢀP)
of the specified CF3 containing b-diketonates was 2.235
Fig. 6. Upper graph: Relationship between d(RhꢀP) and the Pauling
atomic electronegativity of co-ordinating atoms O
the PPh3 group in complexes of the type [Rh(L,L%BID)(CO)(PPh3)].
L,L%-BID, bidentate ligands possessing O,O or N,O or S,O co-ordi-
nating donor atoms. The indicated RN groups in the N
,O and N,S
series represents the third substituent on N. The individual com-
pounds used are 27 and 28 (N,O series) or 31 and 32 (N,S series),
while the S,O entry is for the sacac complex 26. Ph, phenyl; Me,
6 , N6 and S6 trans to
6
6
6
6
6
6
6
6
methyl; Ch, cyclohexyl. Bottom graph: Relationship between
d(RhꢀP) and group electronegativities of the R groups CF3, Me or Ph
adjacent to the carbonyl group of the b-diketonato ligand trans to the
PPh3 group in complexes of the type [Rh(RCOCHCOR%)(CO)(PPh3)].
The CF3 point is from the average d(RhꢀP) of complexes 8, 16 and
17, the Me point is from the average d(RhꢀP) for complexes 19a, 12,
24 and 25, while the Ph point is from 19b. The same compound sets
,
,
A, for the CH3 b-diketonates it is 2.245 A and for the
,
indicated phenyl-containing b-diketonate it is 2.248 A.
were also used in the upper graph of the O
influence the position of the points of either series.
6 ,O series. R% does not
Fig. 6, upper graph, demonstrates how d(RhꢀP) varies

J. Conradie et al. / Inorganica Chimica Acta 328 (2002) 191–203
201
dinated ligand with PPh3 trans to S, deviates so sub-
stantially from that of the S,O complex 26. A further
point of interest is observed in the N,S co-ordinated
complexes 31 and 32 with PPh3 trans to N. Although
they do not belong to the N,O class of compounds with
PPh3 trans to N, they would fit into this latter series of
compounds in terms of d(RhꢀP). These observations
seem to imply that, provided the atomic or group
electronegativity of the atoms or groups cis to the RhꢀP
bond is not substantially larger than those of the atoms
or groups trans to the RhꢀP bond, the influence of the
cis group or atom can be ignored. This is to be ex-
pected, because the shape of d orbitals allows for more
effective trans influences than cis influences. It also
correlates well with the fact that the trans effect (a
kinetic phenomenon that is defined as the ability of a
ligand to influence the rate of substitution of a ligand
bound to a metal trans to it) dominates the cis effect
substantially [33]. The observed good relationship be-
tween d(RhꢀP) and the Pauling atomic electronegativi-
ties of O, N and S in the indicated main classes of
complexes begs the question how variations within a
particular main group, such as the CF3, CH3 and Ph
sub-members of the O,O family of b-diketonato com-
plexes, may manifest itself. The bottom graph in Fig. 6
demonstrates the relationship between group elec-
tronegativities of the indicated groups adjacent to the
carbonyl group of the b-diketonato ligand trans to the
PPh3 group and d(RhꢀP). The smaller spread of
d(RhꢀP) values within this family of b-diketonato of
nucleus and this manifests in a single, larger w(CO)
stretching wave number than that observed for the
monocarbonyl(phosphine)rhodium(I) complexes. By
noting that shorter wavelengths (i.e. longer wave num-
bers) represent stronger CꢀO bond energies in carbonyl
groups, and therefore weaker RhꢀC bonds in RhꢀCO
complexes, it follows that the Rh(III)ꢀC bond in [Rh(b-
diketonato)(CO)(PPh3)(CH3)(I)] complexes should be
weaker (i.e. longer) than Rh(I)ꢀC bonds in [Rh(b-diket-
onato)(CO)(PPh3)] complexes. On the same basis, both
of the Rh(I)ꢀC bonds in [Rh(b-diketonato)(CO)2)] com-
plexes should be weaker than the Rh(I)ꢀC bond in the
corresponding [Rh(b-diketonato)(CO)(PPh3)] com-
plexes. The crystallographic determined bond lengths
(Table 2) for [Rh(fctfa)(CO)2], [Rh(fctfa)(CO)(PPh3)]
and [Rh(fctfa)(CO)(PPh3)(CH3)(I)] are in agreement
with all the described predictions.
From the foregoing it follows that for all the com-
plexes of the type [Rh(RCOCHCOR%)(CO)(PPh3)], the
group electronegativity of the terminal R and R% groups
on the b-diketonato ligand directly influences the elec-
tron density on the rhodium metal centre. Since the
change in electron density on the metal is also reflected
in changes in carbonyl stretching frequencies (increased
electron density on a metal centre results in lower CO
infrared frequencies),
a relationship between this
parameter and group electronegativities of R and R%
should exist. We have previously determined apparent
group electronegativities, R, on the Gordy scale from
the formal reduction potential E°% of five ferrocene-con-
taining b-diketones as well as the IR stretching frequen-
,
complexes, 2.235–2.248 A, as compared with the much
cies of eight different methyl esters [16,34]. These 
R
larger spread over the three main groups of ligand,
values were used to determine the sum of the group
electronegativities of R and R%, that is R+R%, of the
b-diketonato ligand (RCOCHCOR%)− in Table 1. Fig.
7 illustrates, as was expected, that in complexes of the
type [Rh(RCOCHCOR%)(CO)(PPh3)] the infrared CO
stretching frequency, w(CO), increases linearly as R and
R% are replaced by more electron withdrawing groups,
i.e. groups with a higher group electronegativity. It
follows that a measure of the electron density on the
rhodium nucleus of this class of compounds may be
expressed in terms of the sum of the group electronega-
tivities of R and R% of the b-diketonato ligand
(RCOCHCOR%)− co-ordinated to the rhodium nucleus.
The larger this quantity is, the less will the electron
density on the rhodium core be. Finally, we have
previously shown that there is a linear relationship
,
2.235–2.300 A, are readily understood upon recognis-
ing that the latter is the result of an electronic effect by
an atom directly bonded to the rhodium nucleus (a first
order effect), while the first is an electronic effect as a
result of a group of atoms separated from the rhodium
nucleus by two atoms, i.e. a third order effect. Were it
not for the fact that the b-diketonato skeleton is
pseudo-aromatic which allows electronic communica-
tion via conjugation, the latter effect would probably
not have been noticeable at all.
The IR spectrum of the complexes [Rh(FcCOCH-
COR)(CO)2] (2–5) has two distinctive separate peaks
(Table 2) at higher wave numbers than the observed
single IR carbonyl stretching frequency for the mono-
carbonyl-phosphine rhodium(I) complexes 8–11 (Table
1). The lower CO wave number, w(CO), for monocar-
bonyl-phosphine rhodium(I) compounds is a result of
the higher electron density on the rhodium centre due
to electron donation of PPh3 through the d orbital s
bond that binds it to the rhodium(I) nucleus. Oxidative
addition of methyl iodide to the monocarbonyl(phos-
phine)rhodium(I) complexes results inter alia in the
rhodium(III) complexes 12–15. The rhodium(III) nu-
cleus is more electron deficient than the rhodium(I)
between pK% of ferrocene-containing b-diketones and
a
the group electronegativities of substituents on them
[16,34]. One would therefore also expect a linear rela-
tionship between w(CO) of [Rh(b-diketonato)(CO)-
(PPh3)] complexes and pK% of the free b-diketones. This
a
was found to be the case as illustrated in the insert of
Fig. 7. From Fig. 7, it follows that w(CO) of complexes
of the type [Rh(b-diketonato)(CO)(PPh3)] may be cal-
culated by the equations

202
J. Conradie et al. / Inorganica Chimica Acta 328 (2002) 191–203
[2] W.A. Herrmann, Kontakte 1 (1991) 22.
[3] W.A. Herrmann, Kontakte 3 (1991) 29.
[4] C. Mercier, P. Chabardes, Pure Appl. Chem. 66 (1994) 1509.
[5] (a) J. Adams, D.M. Spero, Tetrahedron 47 (1991) 1765;
(b) A.J. Anciaux, A. Demonceau, A.F. Noels, R. Warin, A.J.
Hubert, P. Teyssie´, Tetrahedron 24 (1983) 2169.
[6] J.M. O’Connor, J. Ma, Tetrahedron 57 (1992) 5075.
[7] (a) K. Burgess, M.J. Ohlmeyer, Chem. Rev. 91 (1991) 1179;
(b) D.A. Evans, G.C. Fu, A.H. Hoveyda, J. Am. Chem. Soc. 114
(1992) 6671.
[8] (a) N. Sakai, S. Mano, K. Nozaki, H. Takaya, J. Am. Chem.
Soc. 115 (1993) 7033;
(b) G.D. Cuny, S.L. Buchwald, J. Am. Chem. Soc. 115 (1993)
2066.
[9] (a) M.J. Burk, Y.M. Wang, J.R. Lee, J. Am. Chem. Soc. 118
(1996) 5142;
(b) D.A. Evans, G.C. Fu, Chem. Rev. 93 (1993) 1307;
(c) W.R. Cullen, S.J. Rettig, F.B. Wickenheiser, J. Mol. Catal.
66 (1991) 251.
Fig. 7. Relationship between the carbonyl stretching wave number,
w(CO), and the sum of the group electronegativities of R and R%
(R+R%) of the b-diketonato ligand (RCOCHCOR%)− co-ordinated
to [Rh(RCOCHCOR%)(CO)(PPh3)]. The b-diketonato ligands are as
[10] (a) W.R. Cullen, F.B. Wickenheiser, J. Organomet. Chem. 370
(1989) 141;
(b) in: L.A. Paquette (Ed.), Encyclopedia of Reagents for Or-
ganic Synthesis, vol. 6, Wiley, West Sussex, 1995, p. 4409.
[11] J.H. Jones, Platinum Met. Rev. 44 (2000) 94.
[12] P.M. Maitlis, A. Haynes, G.J. Sunley, M.J. Howard, J. Chem.
Soc., Dalton Trans. (1996) 2187.
indicated. The insert shows how w(CO) varies with the apparent pK%
a
values of the free b-diketone ligands RCOCH2COR% of the same
complexes. Values of individual group electronegativities,  and 
,
R
R%
are as per refs [16,34].
[13] J.D. Atwood, Coord. Chem. Rev. 83 (1988) 93.
[14] (a) J.A. Venter, J.G. Leipoldt, R. van Eldik, Inorg. Chem. 30
(1991) 2207;
w(CO)=5(1)(R+R%)+1959(6)
(b) J.G. Leipoldt, S.S. Basson, L.J. Botha, Inorg. Chim. Acta
168 (1990) 215;
= −1.3(3)pK% +1993(2)
a
(c) S.S. Basson, J.G. Leipoldt, J.T. Nel, Inorg. Chim. Acta 84
(1984) 167.
[15] (a) G.J.J. Steyn, A. Roodt, J.G. Leipoldt, Rhodium Ex. 1 (1993)
25;
4. Conclusions
New rhodium complexes of the type [Rh(FcCOCH-
COR)(CO)2], [Rh(FcCOCHCOR)(CO)(PPh3)] and
[Rh(FcCOCHCOR)(CO)(PPh3)(CH3)(I)] were synthe-
sised and characterised. Equilibrium constants between
two possible isomers of a large number of unsymmetri-
cal rhodium(I) phosphine complexes have been deter-
mined, and the influence of temperature and solvent on
(b) G.J.J. Steyn, A. Roodt, J.G. Leipoldt, Inorg. Chem. 31
(1992) 3477;
(c) S.S. Basson, J.G. Leipoldt, A. Roodt, J.A. Venter, Inorg.
Chim. Acta 128 (1987) 31;
(d) J.G. Leipoldt, S.S. Basson, L.J. Botha, Inorg. Chim. Acta
168 (1990) 215.
[16] W.C. du Plessis, T.G. Vosloo, J.C. Swarts, J. Chem. Soc.,
Dalton Trans. (1998) 2507.
[17] B.S. Furniss, A.J. Hannaford, P.W.G. Smith, A.R. Tatchell,
Vogel’s Textbook of Practical Organic Chemistry, 5th ed., Wiley,
New York, 1994, p. 409.
[18] MINSQ, Non-linear parameter estimation V3.12, MicroMath Sci-
entific Software, Salt Lake City, UT, 1990.
[19] J. Chatt, L.M. Venanzi, J. Chem. Soc. A (1957) 4735.
[20] (a) Y.S. Varshavskii, T.G. Cherkasova, Russ. J. Inorg. Chem. 12
(1967) 899;
(b) A. Rusina, A.A. Vlcek, Nat. London 206 (1965) 295.
[21] J. Conradie, G.J. Lamprecht, T.S. Cameron, M.A.S. Aquino,
J.C. Swarts, unpublished work.
this equilibrium was quantified. The influence that pK%
a
values of b-diketone ligands and group electronegativi-
ties of substituents R on b-diketonato ligands has on
d(RhꢀP) or w(CO) stretching frequencies was demon-
strated and a relationship between d(RhꢀP), and 31P
NMR peak positions as well as coupling constants
1J(31Pꢀ103Rh) have been quantified.
Acknowledgements
[22] G.J. Lamprecht, J.C. Swarts, J. Conradie, J.G. Leipoldt, Acta
Crystallogr., Sect. C 49 (1993) 82.
Financial assistance by the South African NRF and
the Central Research Fund of the University of the
Free State is gratefully acknowledged, as is Professor
Andreas Roodt for helpful discussions.
[23] J. Conradie, G.J. Lamprecht, A. Roodt, J.C. Swarts, unpub-
lished work.
[24] (a) P.J. Swarts, M. Immelman, G.J. Lamprecht, S.E. Greyling,
J.C. Swarts, S. Afr. J. Chem. 50 (1997) 208;
(b) N.F. Blom, E.W. Neuse, H.G. Thomas, Transit. Met. Chem.
12 (1987) 301;
(c) J.C. Swarts, E.W. Neuse, G.J. Lamprecht, J. Inorg.
Organomet. Pol. 4 (1994) 143;
References
(d) J.C. Swarts, P.N. Nonjola, unpublished work.
[25] T. Allmann, R.G. Goel, Can. J. Chem. 60 (1982) 716.
[1] S.E. Sherman, S.J. Lippard, Chem. Rev. 87 (1987) 1153.

J. Conradie et al. / Inorganica Chimica Acta 328 (2002) 191–203
203
[26] S.H. Maron, J.B. Lando, Fundamentals of Physical Chemistry,
Macmillan, New York, 1974, pp. 376–383.
T.S. Cameron, M.A.S. Aquino, J.C. Swarts, Can. J. Chem. 77
(1999) 378.
[27] P.W. Atkins, Physical Chemistry, 5th ed., Oxford University
Press, Oxford, 1994, pp. 141–154, 276 and 288–289.
[28] (a) Y.S. Varshavsky, I.A. Poletaeva, T.G. Cherkasova, I.S. Pod-
korytov, Rhodium Ex. 14 (1995) 10;
[35] J.G. Leipoldt, S.S. Basson, J.H. Potgieter, Inorg. Chim. Acta 117
(1986) L3.
[36] J.G. Leipoldt, S.S. Basson, J.T. Nel, Inorg. Chim. Acta 74 (1983)
85.
(b) I.A. Poletaeva, T.G. Cherkasova, L.V. Osetrova, Y.S. Var-
shavsky, A. Roodt, J.G. Leipoldt, Rhodium Ex. 3 (1994) 21;
(c) A.M. Trzeciak, J.J. Zio´lkowski, Inorg. Chim. Acta 96 (1984)
15.
[37] W. Purcell, S.S. Basson, J.G. Leipoldt, A. Roodt, H. Preston,
Inorg. Chim. Acta 234 (1995) 153.
[38] E.C. Steynberg, G.J. Lamprecht, J.G. Leipoldt, Inorg. Chim.
Acta 133 (1987) 33.
[29] (a) K. Dimroth, C. Reichardt, T. Siepmann, F. Bohlman,
Liebigs Ann. Chem. 661 (1963) 1;
[39] D. Lamprecht, G.J. Lamprecht, J.M. Botha, K. Umakoshi, Y.
Sasaki, Acta Crystallogr., Sect. C 53 (1997) 1403.
[40] J.G. Leipoldt, S.S. Basson, L.D.C. Bok, T.I.A. Gerber, Inorg.
Chim. Acta 26 (1978) L35.
[41] A. Roodt, J.G. Leipoldt, J.C. Swarts, G.J.J. Steyn, Acta Crystal-
logr., Sect. C 48 (1992) 547.
[42] L.J. Botha, S.S. Basson, J.G. Leipoldt, Inorg. Chim. Acta 126
(1987) 25.
[43] J.G. Leipoldt, S.S. Basson, E.C. Grobler, A. Roodt, Inorg.
Chim. Acta 99 (1985) 13.
(b) Y. Marcus, Chem. Soc. Rev. 22 (1993) 409;
(c) J. March, Advanced Organic Chemistry: Reactions, Mecha-
nism and Structure, 4th ed., Wiley, New York, 1992, p. 360.
[30] J.G. Leipoldt, L.D.C. Bok, J.S. van Vollenhoven, A.I. Pieterse,
J. Inorg. Nucl. Chem. 40 (1978) 61.
[31] A.A. Grinberg, Acta Physiochim. USSR 3 (1935) 573.
[32] (a) A. Roodt, G.J.J. Steyn, Recent Res. Dev. Inorg. Chem. 2
(2000) 1;
(b) G.J.J. Steyn, A. Roodt, I. Poletaeva, Y.S. Varshavsky, J.
Organomet. Chem. 536 (1997) 197.
[44] L.J. Damoense, W. Purcell, A. Roodt, J.G. Leipoldt, Rhodium
Ex. 5 (1994) 10.
[33] K.F. Purcell, J.C. Kotz, Inorganic Chemistry, Saunders,
Philadelphia, 1977, pp. 700–708.
[34] W.C. du Plessis, J.C. Erasmus, G.J. Lamprecht, J. Conradie,
[45] L.J. Damoense, W. Purcell, A. Roodt, Rhodium Ex. 14 (1995) 4.
[46] G.G.J. Steyn, A. Roodt, J.G. Leipoldt, Inorg. Chem. 31 (1992)
3477.