Intermediates and anion effects in the activation of carbon-fluorine bonds by η5-pentamethylcyclopentadienylrhodium halide complexes; Crystal structure of [{η5-C5Me3[CH2C 6F4P(C6F5)CH2] 2-1,3}RhBr]+·Br-

Article abstract of DOI:10.1016/S0022-328X(98)01206-6

The reaction between [(η⁵-C₅Me₅)RhBr(μ-Br)]₂ and the diphosphine, (C₆F₅)₂PCH₂CH₂P(C ₆F₅)₂ (dfppe), in benzene proceeded via the intermediate cation [(η⁵-C₅Me₅)RhBr(dfppe)]⁺, which underwent C-F and C-H bond activation and C-C bond formation to give sequentially [{η⁵-C₅Me₄CH₂C₆F ₄P(C₆F₅)CH₂CH₂P(C ₆F₅)₂}RhBr]⁺ and then [{η⁵-C₅Me₃[CH₂C ₆F₄P(C₆F₅)CH₂] ₂-1,3}RhBr]⁺, as evidenced by mass spectrometry and NMR spectroscopy. The bromide salt of the final product (4c) has been structurally characterized by X-ray diffraction. Compound 4c crystallizes in the triclinic space group P1 with a=10.616(1), b=13.904(2), c=14.911(1) A, α=66.86(1), β=86.38(1), γ=84.72(1)° and Z=2. Refinement gave final R1 and wR2 [I=2σ(I)] values of 0.0581 and 0.1641, respectively, for 6837 unique reflections. In contrast to the BF₄^- salt, the Cl^- and BPh₄^- salts of cation [(η⁵-C₅Me₅)RhCl(dfppe)]⁺ undergo reaction upon thermolysis in benzene to give the cation [{η⁵-C₅Me₃[CH₂C ₆F₄P(C₆F₅)CH₂] ₂-1,3}RhCl]⁺.

Full text of DOI:10.1016/S0022-328X(98)01206-6

www.elsevier.nl/locate/jorganchem
Journal of Organometallic Chemistry 582 (1999) 163–172
Intermediates and anion effects in the activation of
carbon–fluorine bonds by h5-pentamethylcyclopentadienylrhodium
halide complexes; crystal structure of
[{h5-C5Me3[CH2C6F4P(C6F5)CH2]2-1,3}RhBr]+·Br−
Malcolm J. Atherton a, John Fawcett b, John H. Holloway b, Eric G. Hope b,
David R. Russell b, Graham C. Saunders b,c,
*
a F2 Chemicals, Springfields, Salwick, Preston PR4 0XJ, UK
b Department of Chemistry, Uni6ersity of Leicester, Leicester LE1 7RH, UK
c School of Chemistry, The Queen’s Uni6ersity of Belfast, Da6id Keir Building, Belfast BT9 5AG, UK
Received 9 October 1998
Abstract
The reaction between [(h5-C5Me5)RhBr(m-Br)]2 and the diphosphine, (C6F5)2PCH2CH2P(C6F5)2 (dfppe), in benzene proceeded
via the intermediate cation [(h5-C5Me5)RhBr(dfppe)]+, which underwent CꢀF and CꢀH bond activation and CꢀC bond formation
to give sequentially [{h5-C5Me4CH2C6F4P(C6F5)CH2CH2P(C6F5)2}RhBr]+ and then [{h5-C5Me3[CH2C6F4P(C6F5)CH2]2-
1,3}RhBr]+, as evidenced by mass spectrometry and NMR spectroscopy. The bromide salt of the final product (4c) has been
structurally characterized by X-ray diffraction. Compound 4c crystallizes in the triclinic space group P1 with a=10.616(1),
,
b=13.904(2), c=14.911(1) A, h=66.86(1), i=86.38(1), k=84.72(1)° and Z=2. Refinement gave final R1 and wR2 [I=2|(I)]
values of 0.0581 and 0.1641, respectively, for 6837 unique reflections. In contrast to the BF4− salt, the Cl− and BPh4− salts of
cation [(h5-C5Me5)RhCl(dfppe)]+ undergo reaction upon thermolysis in benzene to give the cation [{h5-
C5Me3[CH2C6F4P(C6F5)CH2]2-1,3}RhCl]+. © 1999 Elsevier Science S.A. All rights reserved.
Keywords: CꢀF bond activation; CꢀH bond activation; CꢀC bond formation; Rhodium; 1,2-Bis{bis(pentafluorophenyl)phosphino}ethane
F5)CH2]2-1,3; Fig. 1) in which two aryl CꢀF bonds and
two pentamethylcyclopentadienyl CꢀH bonds have
been cleaved and two CꢀC bonds have been formed
(Eq. (1)) [1,2].
1. Introduction
As part of a programme to develop the synthetic
utility of CꢀF bond activation in the synthesis of
organometallic complexes we have been studying the
intriguing reaction between the metal complexes [(h5-
C5Me4R)MCl(m-Cl)]2 (M=Rh, Ir; R=H, Me, Et) and
the diphosphine (C6F5)2PCH2CH2P(C6F5)2 (dfppe). We
have reported that dfppe reacts with [(h5-
C5Me5)RhCl(m-Cl)]2 under mild conditions in benzene
or ethanol to afford, in quantitative yield, the cationic
complex [LRhCl]+ (11, L=h5-C5Me3[CH2C6F4P(C6-
12[(h5-C5Me5)RhCl(m-Cl)]2+dfppe
“[LRhCl]+·X− +2HF
(1)
(X− =Cl− 1a, X− =BF4− 1b)
The reaction exhibits complete regiospecificity: only
one ortho CꢀF bond of each P(C6F5)2 moiety and CꢀH
bonds of methyl groups exclusively in a 1,3 disposition
are activated. Further, remarkable regioselectivity was
observed in the reaction between [(h5-C5Me4H)RhCl(m-
Cl)]2 and dfppe, in which a racemic mixture of the
chiral-at-metal cation [{h5-C5HMe2-2,4-[CH2C6F4P(C6-
* Corresponding author.
E-mail address: g.saunders@qub.ac.uk (G.C. Saunders)
1 For salts, the number refers to the cation and the letter indicates
the anion: a=Cl-, b=BF4−, c=Br−, d=BPh4−, e=PF6−
.
0022-328X/99/$ - see front matter © 1999 Elsevier Science S.A. All rights reserved.
PII: S0022-328X(98)01206-6

164
M.J. Atherton et al. / Journal of Organometallic Chemistry 582 (1999) 163–172
F5)CH2]2-1,3}RhCl]+ was formed selectively in high
yield [3]. In order to extend the scope of such reactivity
it is important to understand the mechanism of reaction
(1) and related reactions.
[(h5-C5Me5)RhCl(dfppe)]+ · BF4−
(2b)
“[LRhCl]+ · BF4− +2HF
(1b)
(2)
Similar observations were noted for reactions involv-
ing (h5-C5Me5)IrCl [2] and (h5-C5Me4Et)RhCl species
[3] in ethanol. However, there is no evidence to support
this sequence for the reaction in benzene. Subsequent in
situ NMR spectroscopic studies, performed on the reac-
tion between [(h5-C5Me5)RhCl(m-Cl)]2 and dfppe in
C6D6, have failed to detect the presence of either 2 or 3.
(Spectroscopic data indicate that the cations [(h5-
C5Me4H)RhCl(dfppe)]+ and [{h5-C5Me3HCH2C6F4-
P(C6F5)CH2CH2P(C6F5)2}RhCl]+ are present in the
final product mixture of the reaction between [(h5-
C5Me4H)RhCl(m-Cl)]2 and dfppe in benzene [3], but
their presence may indicate their existence as by-prod-
ucts and not as intermediates.) Furthermore, 2b does
Reaction (1) is complete in refluxing benzene or
ethanol after a few hours, but also occurs at room
temperature and in dichloromethane, albeit more
slowly. In ethanol two intermediate cations have been
identified, and it has been established that the reaction
proceeds via cation 2, which undergoes CꢀF and CꢀH
bond activation and CꢀC bond formation with loss of
HF to give the singly activated cation [L%RhCl]+ (3,
L%=h5-C5Me4CH2C6F4P(C6F5)CH2CH2P(C6F5)2, Fig.
1) and then 1 (Scheme 1) [2]. Furthermore, the te-
trafluoroborate salt of 2 (2b), prepared independently,
undergoes reaction on thermolysis in ethanol to yield
1b via 3b (Eq. (2)) [2].
Fig. 1. Ligands L and L%.
Scheme 1.

M.J. Atherton et al. / Journal of Organometallic Chemistry 582 (1999) 163–172
165
Scheme 2. (i) dfppe, C6H6, reflux 5 days; (ii) NH4BF4, MeOH; (iii) dfppe, CH2Cl2; (iv) EtOH, reflux.
not undergo reaction on thermolysis in benzene and
other aprotic solvents such as dichloromethane or dry
acetone [2]. Clearly the solvent exerts a strong influ-
ence in these reactions. We were interested in deter-
mining whether reaction (1) proceeds by a different
mechanism in benzene to that in ethanol and whether
the anion also exerts a large effect in this reaction.
Our previous studies into the mechanism and scope
of these reactions have concentrated on variation of
the metal [2] and the cyclopentadienyl ligand [3,4].
Here, we report the effect of changing the halide lig-
and, which has allowed the determination of the in-
termediates in the reactions between [(h5-C5Me4R)-
RhX(m-X)]2 and dfppe in benzene, and the effect of
the anion on the thermolysis of [(h5-C5Me5)-
RhCl(dfppe)]+·X−.
of the product and isolated in moderate yield and
also by an X-ray diffraction study of a crystal of
bromide salt, 4c, which was crystallized from the
crude product mixture. The spectroscopic data are en-
tirely consistent with the formulation of 4b and simi-
lar to those of 1a [1]. Unfortunately it was not
possible to obtain elemental analysis of 4b due to
traces of other products with similar solubilities from
which it could not be separated.
The tetrafluoroborate salt of the postulated inter-
mediate bromo-diphosphine cation, [(h5-C5Me5)-
RhBr(dfppe)]+·BF4− (5b), was prepared in good yield
by addition of dfppe to
a
solution of [(h5-
C5Me5)RhBr(m-Br)]2 and NH4BF4 in methanol. Salt
5b was characterized by elemental analysis, mass spec-
trometry and multinuclear NMR spectroscopies. The
1H-, 19F- and 31P{1H}-NMR spectra show only minor
differences to those of the chloride analogue 2b [2].
For example, the phosphorus resonance of 5b is
shifted to lower frequency by only 1.1 ppm. As with
2b, thermolysis of 5b in ethanol induces CꢀF and
CꢀH bond activation/CꢀC bond formation and the
diactivated product 4b was obtained almost quantita-
tively after 4.5 h (Scheme 2).
2. Results and discussion
2.1. Synthesis and characterization of [LRhBr]+·BF−4
(4b) and [(p5-C5Me5)RhBr(dfppe)]+·BF−4 (5b)
The reaction between [(h5-C5Me5)RhBr(m-Br)]2 and
dfppe proceeded in refluxing benzene with activation
of two CꢀF and CꢀH bonds and the formation of
two CꢀC bonds to afford a mixture of the salts
[LRhBr]+·X−, X− =Br− or BF4− (Scheme 2). The
complex was characterized by mass spectrometry and
multinuclear NMR spectroscopies of the tetrafluorob-
orate salt, 4b, which was formed by anion metathesis
Attempts to prepare the iodo analogues of 4b and
5b from [(h5-C5Me5)RhI(m-I)]2 were unsuccessful. Pre-
sumably, the larger size of iodide (covalent radius:
,
,
1.33 A [5]) in comparison with chloride (0.99 A) and
,
bromide (1.14 A) prevents coordination of the bulky
dfppe ligand. Simple molecular modelling supports
this hypothesis.

166
M.J. Atherton et al. / Journal of Organometallic Chemistry 582 (1999) 163–172
2.2. Crystal structure of [LRhBr]+·Br− (4c)
completely accommodated by movement of only two of
the carbons attached to phosphorus, and there is no
hint of disorder in the third attached carbons [C(21)
and C(41)]. There are indications of solvent molecules
in the crystal which could not be fully identified, the
strongest residual density was modelled by a single
oxygen atom, presumably of water.
Crystals of 4c suitable for single-crystal X-ray dif-
fraction study were obtained by slow evaporation of
solvent from an acetone solution of the crude product
of the reaction between [(h5-C5Me5)RhBr(m-Br)]2 and
dfppe. The crystallographic data, atomic coordinates
and selected distances and angles are given in Tables
1–3, respectively. The cation has a potential mirror
plane [through Rh(1), Br(1) and the C5 ring centroid],
but in the crystal the PCH2 methylene carbon atoms
and the terminal pentafluorophenyl rings are twisted
away from an ideal conformation, presumably to re-
duce steric interactions between the ortho-F atoms
The structure (Fig. 2) is similar to that of the te-
trafluoroborate salt of the chloride analogue 1b [1]. The
RhꢀC5(centroid) distance is the same as that of 1b. The
two RhꢀP distances are identical within experimental
error and similar to those of 1b. The RhꢀBr distance of
,
,
2.5018(10) A is ca. 0.12 A longer than the RhꢀCl
,
distance of 1b (2.380(3) A) consistent with the larger
,
F(12) and F(32). [Distance F(12)···F(32) is 2.601 A,
size of bromide relative to chloride. The PꢀRhꢀP angle
is 86.94(7)°, consistent with that of 1b. The
C5(centroid)ꢀRhꢀBr angle of 122.5(2)° is ca. 4° more
acute, and the two PꢀRhꢀBr angles of 94.43(6) and
93.81(6)° are more obtuse than the analogous angles of
1b (126.1(1), 90.6(1) and 93.0(1)°, respectively). Pre-
sumably, the differences in the angles between 4c and
1b arise because of the larger size of the bromide ligand
compared with the chloride ligand.
,
which would be ca. 2.26 A in the idealized, untwisted
conformation]. The atoms involved in this twisted con-
formation are 50% disordered between the two possible
twist directions, thus maintaining the near mirror sym-
metry of the cation. These distortions appear to be
Table 1
Crystallographic data for [LRhBr]+·Br− (4c)
Formula
C36H17Br2F18P2Rh.O
Formula weight
Crystal system
Space group
1132.17
Triclinic
P1
10.616(1)
13.904(2)
14.911(1)
66.86(1)
86.38(1)
84.72(1)
2014.4(4)
2
1096
1.867
0.63×0.54×0.41
Mo–Ka (0.71073)
Graphite
2.3. Reaction between [(p5-C5Me5)RhBr(v-Br)]2 and
dfppe in benzene
,
a (A)
,
b (A)
The reaction of dfppe with [(h5-C5Me5)RhBr(m-Br)]2
is similar to that with the chloride analogue [1,2], but
proceeded at a far slower rate. This afforded a conve-
nient opportunity to carry out investigations of the
reaction mixture at various times throughout the reac-
tion. This was achieved by stopping the reflux, allowing
the mixture to cool and removing the benzene by rotary
evaporation to allow characterization of the solids by
mass spectrometry and multinuclear NMR spectro-
scopies in CDCl3. After characterization, the mixture
was redissolved in benzene and reflux continued.
,
c (A)
h (°)
i (°)
k (°)
3
,
V (A )
Z
F(000)
Dcalc (g cm−3
)
Crystal size (mm)
,
Radiation (u (A))
Monochromator
v (Mo–Ka) (mm−1
T (K)
)
2.601
190(2)
The 19F- and 31P{1H}-NMR spectra after 6 h indicate
that dfppe was the predominant phosphorus and
fluorine containing species in the mixture. The 31P-
NMR and mass spectra indicate the presence of the
Scan method
h, k, l Ranges
2q limits (°)
Total reflections
Unique reflections
Observed reflections
[Fo=4|(Fo)]
Absorption correction
method
Max and min transmission
Restraints
ꢀ
−1 to 12, −15 to 15, −17 to 17
5.04–50.00
8045
6837 (Rint=0.0238)
5302
bromo-diphosphine cation
5 and the diactivated
product cation 4. The mass spectrum also contained
small peaks at 1057 and 1055 suggesting the presence of
the singly activated cation [L%RhBr]+ (6). After 26 h
the spectral data indicated that virtually no dfppe re-
mained and that cation 5 was the predominant species.
After 48 h the 31P{1H}-NMR spectrum displays two
more doublets of multiplets at l 83.8 and 57.9 with
1JRhꢀP 143 and 128 Hz, respectively, in equal intensity.
Comparison of these data with those of singly activated
3b [2] indicate that these can be assigned to cation 6.
The relative proportion of 5 had diminished and no
dfppe remained. After 122 h at reflux, and addition of
Semi-empirical based on „-scans
0.9547, 0.7108
12
507
Variables
R1, wR2 [I=2|(I)]
R1, wR2 (all data)
Weighting scheme
0.0581, 0.1641
0.0794, 0.1783
w=1/[|2(Fo)2+0.09P2+11.55P]a
(D/|)max
0.25
2.472, −0.882
1.076
−3
,
Max, min Dz (eA
)
Goodness of fit on F2
a P=[max(Fo2, 0)+2F2c]/3.

M.J. Atherton et al. / Journal of Organometallic Chemistry 582 (1999) 163–172
167
Table 2 (Continued)
Table 2
Atomic coordinates (×104) and equivalent isotropic displacement
2
3
Atom
x
y
z
U(eq)
,
coefficients (A ×10 ) with estimated standard deviations in paren-
theses for the non-hydrogen atoms of [LRhBr]+·Br− (4c)
F(14A)
F(15)
F(15A)
F(16)
F(16A)
F(23)
F(24)
3278(4)
12557(2)
7060(2)
11147(3)
6102(2)
9092(2)
2702(4)
2536(4)
4295(5)
6168(4)
9858(2)
9138(2)
11914(2)
10082(2)
12616(2)
8944(2)
11260(3)
6862(2)
9203(2)
5917(2)
6370(4)
7269(5)
8676(6)
9070(4)
6640(7)
5759(2)
3061(2)
7055(3)
5044(2)
7947(2)
7703(4)
7039(4)
6037(4)
5700(4)
5581(2)
4993(2)
4841(2)
3018(1)
5845(2)
1997(1)
7591(2)
2950(2)
8332(2)
4923(2)
83(5)
29(2)
90(5)
30(3)
48(3)
50(1)
54(1)
57(1)
49(1)
44(3)
45(2)
53(3)
58(3)
55(4)
48(3)
65(4)
47(3)
59(4)
42(3)
51(1)
60(2)
71(2)
52(1)
78(2)
3974(3)
4924(3)
3681(3)
4401(3)
2791(5)
588(5)
Atom
x
y
z
U(eq)
Rh(1)
Br(1)
Br(2)
P(1)
P(2)
3696(1)
4567(1)
847(1)
6989(1)
8441(1)
5047(1)
6801(2)
8080(2)
6282(6)
7169(7)
7383(2)
9456(2)
8286(2)
7165(1)
5717(1)
1521(1)
6358(1)
7625(1)
8604(5)
7784(7)
5044(2)
7104(1)
4528(2)
6444(2)
3516(2)
5990(2)
3021(2)
6196(2)
3537(2)
6856(2)
4548(2)
7310(2)
8471(5)
6560(5)
7060(5)
7221(6)
6882(6)
6371(6)
6220(6)
7667(5)
6964(1)
4976(2)
6074(2)
4491(1)
5696(2)
3486(1)
6208(2)
2966(1)
7099(2)
3451(2)
7477(2)
4456(2)
7321(5)
8917(5)
9624(6)
10579(6)
10875(6)
10207(7)
9241(6)
7896(6)
6920(18)
7523(23)
7239(21)
6633(20)
9420(6)
9125(6)
7348(6)
6561(6)
5004(2)
6245(2)
3021(2)
5354(2)
2049(2)
24(1)
45(1)
55(1)
27(1)
28(1)
33(2)
47(2)
13(3)
18(3)
29(3)
42(4)
37(4)
50(5)
29(3)
52(5)
21(5)
47(4)
25(6)
31(4)
32(2)
30(2)
30(2)
36(2)
39(2)
38(2)
36(2)
31(2)
37(4)
32(3)
32(4)
34(3)
35(4)
36(4)
29(3)
37(4)
41(4)
46(8)
48(5)
31(6)
30(2)
30(2)
30(2)
37(2)
44(2)
46(2)
40(2)
34(2)
21(5)
41(8)
26(6)
27(6)
41(2)
38(2)
37(2)
43(2)
40(2)
57(3)
59(3)
90(5)
53(4)
2071(2)
2356(2)
4533(7)
6598(8)
2047(2)
2489(3)
1290(2)
1651(2)
1440(3)
1917(3)
2347(3)
3021(3)
3104(3)
3859(3)
2954(3)
3593(3)
3742(7)
1755(7)
2522(7)
2082(8)
950(9)
F(25)
F(26)
F(32)
F(32A)
F(33)
F(33A)
F(34)
F(34A)
F(35)
F(35A)
F(36)
F(36A)
F(43)
−904(5)
−140(5)
1653(2)
1544(2)
2201(4)
1916(3)
3705(4)
2983(4)
4660(4)
3679(3)
4111(4)
3307(3)
3658(5)
1611(5)
53(6)
C(1)
C(10)
C(11)
C(11A)
C(12)
C(12A)
C(13)
C(13A)
C(14)
C(14A)
C(15)
C(15A)
C(16)
C(16A)
C(2)
C(21)
C(22)
C(23)
C(24)
C(25)
C(26)
C(3)
C(31)
C(31A)
C(32)
C(32A)
C(33)
C(33A)
C(34)
C(34A)
C(35)
C(35A)
C(36)
C(36A)
C(4)
C(41)
C(42)
C(43)
C(44)
C(45)
C(46)
C(5)
10174(2)
8774(2)
11220(2)
8361(2)
11549(2)
7458(2)
10830(2)
6969(2)
9784(2)
5563(6)
5462(6)
4564(6)
3596(6)
3490(6)
4367(7)
5332(6)
5316(6)
9510(2)
7519(2)
10201(2)
8579(2)
11250(2)
9060(2)
11608(2)
8480(2)
10916(2)
7420(2)
9867(2)
6939(2)
5883(6)
7849(6)
7163(6)
7021(6)
7498(8)
8195(8)
8366(7)
6488(6)
7181(18)
7714(27)
8076(21)
7459(21)
6659(7)
5022(6)
4512(6)
5811(7)
8683(2)
9857(2)
9641(2)
11912(2)
8830(2)
11281(3)
11840(4)
10477(4)
8594(4)
9512(6)
F(44)
F(45)
F(46)
495(5)
O(1)
−582(8)
210(8)
629(8)
NH4BF4, the 19F- and H-NMR spectra recorded in-
dicate that the diactivated product 4b was the major
species, and that only small quantities of other com-
pounds, including 5b and 6b, were also present. The
19F-NMR spectrum displays more resonances than
can be accounted for by the presence of 4b, 5b and
6b alone, and it is therefore inferred that the reaction
is not as clean as between [(h5-C5Me5)RhCl(m-Cl)]2
and dfppe. The identities of the by-products are, as
yet, not known.
The data clearly establish that the reaction between
[(h5-C5Me5)RhBr(m-Br)]2 and dfppe in benzene pro-
ceeds via formation of the bromo-diphosphine cation
5, followed by sequential CꢀF and CꢀH bond activa-
tion with concomitant CꢀC bond formation to form
the singly activated 6 and finally the di-activated
product 4. This reaction scheme is analogous to the
reaction between [(h5-C5Me5)RhCl(m-Cl)]2 and dfppe
in ethanol (Scheme 1). Although there is no direct
evidence, it is reasonable to assume that the reaction
between [(h5-C5Me5)RhCl(m-Cl)]2 and dfppe in ben-
zene proceeds via cations 2 and 3. The lack of a
reaction on thermolysis of 2b in benzene may, there-
fore, be inferred to be due to the effect of the anion.
To test this hypothesis it was proposed to synthesize
the chloride, tetraphenylborate and hexafluorophos-
phate salts of cation 2 and investigate any reaction
occurring on thermolysis in benzene.
1
4311(7)
2876(3)
2422(2)
2389(2)
2068(2)
2669(3)
2257(3)
3436(3)
2801(3)
3923(3)
3155(3)
3643(3)
2965(3)
5404(7)
2279(7)
3134(7)
2872(8)
1852(9)
1064(9)
1291(8)
5540(7)
567(26)
767(36)
727(37)
667(30)
4378(8)
2661(8)
3888(8)
6335(8)
419(2)
C(51)
C(51A)
C(52)
C(52A)
C(6)
C(7)
C(8)
C(9)
F(12)
F(12A)
F(13)
F(13A)
F(14)
586(2)
713(3)
1109(3)
2491(4)

168
M.J. Atherton et al. / Journal of Organometallic Chemistry 582 (1999) 163–172
CꢀC bond forming reactions have also occurred to a
small extent. That these reactions have occurred in a
solvent in which 2b undergoes no reaction, even under
prolonged reflux [2], strongly suggests that substitution
of the tetrafluoroborate for chloride facilitates the reac-
tion under mild conditions.
Table 3
Selected interatomic distances (A) and angles (°) for [LRhBr]+·Br−
(4c)
,
Bond lengths
Rh(1)ꢀP(1)
Rh(1)ꢀBr(1)
2.256(2) Rh(1)ꢀP(2)
2.5018(10)Rh(1)ꢀCp*a
2.257(2)
1.829(7)
The syntheses of [(h5-C5Me5)RhCl(dfppe)]+·X−
(X− =BPh4− 2d, X− =PF6−, 2e) were attempted using
a similar route to the preparation of the tetrafluorobo-
rate salt 2b [2] (Scheme 3). For 2d, [(h5-C5Me5)RhCl(m-
Cl)]2 was treated with two equivalents of NaBPh4,
followed by addition of dfppe. For 2e, an excess of
KPF6 was used. The NMR spectroscopic and mass
spectrometric data indicate that 2d and 2e had been
formed, but were components of mixtures also contain-
ing activated cations 1 and 3. The tetraphenylborate
salt 2d constituted ca. 90% and the hexafluorophos-
phate salt 2e considerably less of the respective mix-
tures. The positive ion FAB mass spectra of both
mixtures show the presence of 2 and [2−Cl], and the
negative ion FAB mass spectra confirm the presence of
BPh4− in 2d and PF6− in 2e. The 1H-, 19F- and 31P{1H}-
NMR spectra, recorded in (CD3)2CO, were similar to
those of 2b [2], but the 19F-NMR spectrum lacked
P(1)ꢀC(52A)
P(1)ꢀC(21)
P(1)ꢀC(31A)
P(2)ꢀC(41)
P(2)ꢀC(52)
C(51)ꢀC(52)
Rh(1)ꢀC(1)
C(1)ꢀC(6)
1.79(3) P(1)ꢀC(11)
1.824(8) P(1)ꢀC(51)
1.933(3) P(2)ꢀC(11A)
1.823(8) P(2)ꢀC(51A)
1.86(4) P(2)ꢀC(31)
1.52(5) C(51A)ꢀC(52A)
2.187(7) Rh(1)ꢀC(3)
1.497(11) C(3)ꢀC(8)
1.508(11) C(8)ꢀC(22)
1.419(10) C(21)ꢀC(22)
1.803(3)
1.89(3)
1.775(3)
1.84(4)
1.955(3)
1.51(5)
2.192(7)
1.488(11)
1.526(11)
1.391(10)
C(6)ꢀC(42)
C(41)ꢀC(42)
Bond angles
P(1)ꢀRh(1)ꢀP(2)
P(2)ꢀRh(1)ꢀBr(1)
Cp*ꢀRh(1)ꢀP(1)
C(11)ꢀP(1)ꢀRh(1)
C(31A)ꢀP(1)ꢀRh(1)
C(52A)ꢀP(1)ꢀRh(1)
C(41)ꢀP(2)ꢀRh(1)
C(52)ꢀP(2)ꢀRh(1)
C(52A)ꢀP(1)ꢀC(21)
C(11)ꢀP(1)ꢀC(51)
86.94(7) P(1)ꢀRh(1)ꢀBr(1)
93.81(6) Cp*ꢀRh(1)ꢀBr(1)
94.43(6)
122.5(2)
124.7(2)
116.7(2)
107.2(9)
121.07(13)
104.9(13)
108.96(12)
99.5(2)
124.9(2)
Cp*ꢀRh(1)ꢀP(2)
121.03(13) C(21)ꢀP(1)ꢀRh(1)
107.97(12) C(51)ꢀP(1)ꢀRh(1)
109.1(11) C(11A)ꢀP(2)ꢀRh(1)
117.0(2)
C(51A)ꢀP(2)ꢀRh(1)
110.2(11) C(31)ꢀP(2)ꢀRh(1)
109.4(10) C(11)ꢀP(1)ꢀC(21)
111.4(8)
C(21)ꢀP(1)ꢀC(51)
99.0(8)
106.9(2)
1
resonances due to BF4−. The H-NMR spectrum of 2d
C(52A)ꢀP(1)ꢀC(31A) 106.3(8)
C(21)ꢀP(1)ꢀC(31A)
C(11A)ꢀP(2)ꢀC(51A) 113.0(10)
also contained three resonances between l 6.5 and 7.5
assigned to BPh4− and the 19F- and 31P{1H}-NMR
spectra of 2e also included doublet and septet reso-
nances, respectively, assigned to PF6−. The elemental
analyses of both salts could not be obtained due to
their contamination by significant amounts of 1 and 3,
from which they could not be separated. Preparations
of 2b, if not performed with care, occasionally produce
small quantities of activated cations 1 and 3, but very
much less than produced in the syntheses of 2d and 2e.
These observations are consistent with the anion exert-
ing a large influence on the reactions. Hexafluorophos-
phate and tetraphenylborate appear to render the CꢀF
and CꢀH activation of cation 2 more facile relative to
tetrafluoroborate.
Thermolysis of the mixture containing 2a in benzene
for 11 h gave a yellow solution. The 1H-, 19F- and
31P{1H}-NMR spectra indicated that the product of
thermolysis was the diactivated cation 1, contaminated
by traces of uncomplexed dfppe and cations 2 and 3.
The 19F-NMR spectrum showed that BF4− was present
in an almost stoichiometric quantity. The increase in
the relative amount of this anion on thermolysis is
indicative of CꢀF and CꢀH bond activation occurring
to produce HF, which reacts with the borosilicate glass
reaction vessel to form BF4− [2].
C(11A)ꢀP(2)ꢀC(41)
C(41)ꢀP(2)ꢀC(51A)
C(41)ꢀP(2)ꢀC(31)
C(52)ꢀC(51)ꢀP(1)
100.2(3)
98.7(11) C(41)ꢀP(2)ꢀC(52)
108.1(11)
107.2(8)
113(2)
104.8(3)
112(2)
C(52)ꢀP(2)ꢀC(31)
C(51)ꢀC(52)ꢀP(2)
C(52A)ꢀC(51A)ꢀP(2) 112(3)
C(51A)ꢀC(52A)ꢀP(1) 114(2)
C(3)ꢀRh(1)ꢀP(1)
C(22)ꢀC(21)ꢀP(1)
C(21)ꢀC(22)ꢀC(8)
C(3)ꢀC(8)ꢀC(22)
C(8)ꢀC(3)ꢀRh(1)
92.0(2)
C(1)ꢀRh(1)ꢀP(2)
C(42)ꢀC(41)ꢀP(2)
C(41)ꢀC(42)ꢀC(6)
C(1)ꢀC(6)ꢀC(42)
C(6)ꢀC(1)ꢀRh(1)
92.0(2)
125.7(6)
127.0(7)
120.5(6)
128.8(5)
125.2(6)
126.3(7)
121.3(7)
128.2(5)
a Cp* denotes the centroid of the cyclopentadienyl ring.
2.4. Syntheses and thermolyses of [(p5-C5Me5)RhCl(df-
ppe)]+·X− (X− =Cl− 2a, BPh−4 2d, PF−6 2e)
In an attempt to prepare [(h5-C5Me5)RhCl-
(dfppe)]+·Cl− (2a) by anion metathesis, 2b was twice
treated with an excess of NH4Cl in acetone at room
temperature (Scheme 3). The H-, 19F- and 31P{1H}-
1
NMR spectra showed the orange product of the at-
tempted metathesis to be
a mixture comprising
predominantly chloro-diphosphine cation 2 together
with activated cations 3, 1 and some uncomplexed
dfppe. The 19F-NMR spectrum reveals the presence of
BF4−, but in significantly less than stoichiometric quan-
tity. The mixture was found to be almost completely
soluble in benzene, which 2b is not, suggesting that the
desired chloride salt, 2a, was the major product. Thus,
it appears that the metathesis has been at least partially
successful, but that CꢀF and CꢀH bond activating and
Thermolysis of 2d in benzene for 5 h gave a purple
solution, which yielded a purple oil on removal of the
solvent. The NMR spectra indicated that CꢀF and CꢀH

M.J. Atherton et al. / Journal of Organometallic Chemistry 582 (1999) 163–172
169
Fig. 2. Molecular structure of [LRhBr]+·Br− (4c). Displacement ellipsoids are shown at the 35% probability level. All hydrogen atoms are
omitted for clarity. The disordered site names are derived from their pseudo-mirror-related partners, e.g. C(12A) is opposite C(12) and close to
C(32).
bond activation/CꢀC bond formation had taken place
presumed that the HF formed in this reaction cleaves
one or more of the BꢀC bonds of BPh4−. Since 2e was
very impure the thermolysis was not attempted.
The results of these reactions, together with the evi-
dence presented in Section 2.3, confirms the hypothesis
1
to form cation 1. The H-NMR spectrum showed the
presence of BPh4−, but was complicated by many other
signals due to aromatic hydrogen atoms. No attempt
was made to identify the additional species, but it is
Scheme 3. (i) NH4Cl, (CH3)2CO; (ii) NaBPh4 (2d) or KPF6 (2e), MeOH; (iii) dfppe, CH2Cl2; (iv) C6H6, reflux.

170
M.J. Atherton et al. / Journal of Organometallic Chemistry 582 (1999) 163–172
that the anion is having a large effect in determining
whether the CꢀF and CꢀH bond activating reaction
occurs. Reaction (1), which involves Cl− as the anion,
occurs in aprotic, non-polar solvents, such as benzene,
toluene and dichloromethane (mr 2.274, 2.379 at 25°C
and 9.08 at 20°C, respectively [6]), as well as the
protic, polar solvent ethanol (mr 24.30 at 25°C [6]).
Reaction (2) does not occur in aprotic, non-polar sol-
vents, such as benzene, nor in the aprotic, polar sol-
vent dry acetone (mr 20.7 at 25°C [6]), to only a small
extent in the more polar acetontirile (mr 37.5 at 20°C
[6]), but occurs readily in ethanol [2]. However, on
substitution of the anion from BF4− to Cl− or BPh4−
the reaction occurs readily in benzene. These observa-
tions suggest that the Coulombic interaction between
the cation and anion is not the predominant factor in
inhibiting reaction (1) since the ion pairs should be
dissociated to a greater extent in acetonitrile than in
ethanol. It is clear that BF4− exerts a different influ-
ence than the other anions on the reaction, and that a
protic solvent is necessary to facilitate reaction (1).
Anion effects in the reactions of cationic [(h5-
C5R5)Rh] species have been observed previously, but
are thought to arise due to differences in coordination
of the anions to the metal. For example, in CH3NO2
at 80°C benzene can be displaced from [(h5-
C5Me4Et)Rh(h6-C6H6)]2+·2PF6− by methanol, whereas
[(h5-C5Me4Et)Rh(h6-C6H6)]2+·2BF4− shows no reac-
tion after 48 h [7]. However, it is improbable that
the differences in reactivity of [(h5-C5Me5)-
RhCl(dfppe)]+·X− are due to coordination of the an-
ion to the metal. Even if an intermediate such as
[(h5-C5Me5)Rh(dfppe)]2+ is formed the metal will co-
ordinate chloride preferentially to the bulkier and
more weakly coordinating anions. We are currently
carrying out further investigations to understand more
fully the mechanism of these intriguing reactions and
why tetrafluoroborate exerts such an effect.
4. Experimental
4.1. General procedures
1H-, 19F- and 31P-NMR spectra were recorded on
Bruker AM300, DPX300 or ARX250 spectrometers.
1H-NMR spectra were referenced using the residual
protio solvent resonance relative to tetramethylsilane
(l 0), 19F- and 31P-NMR spectra externally to CFCl3
(l 0) and 85% H3PO4 (l 0), respectively, using the
high frequency positive convention. 19F-NMR spectra
of salts containing BF4−, the fluorine atoms of which
have a relatively high value of T1, were recorded with
a relaxation delay of 10 s. All chemical shifts (l) are
quoted in ppm and coupling constants in Hz. Abbre-
viations used in multiplicities are: s, singlet; d, dou-
blet; t, triplet; m, multiplet; br denotes a signal
broadened due to a fluxional process. Elemental
analyses were performed by Butterworths Ltd. Positive
and negative ion FAB mass spectra were obtained on
a Kratos Concept 1H mass spectrometer using m-ni-
trobenzyl alcohol as matrix.
4.2. Reagents
(C6F5)2PCH2CH2P(C6F5)2 (Fluorochem) and [(h5-
C5Me5)RhCl(m-Cl)]2 (Aldrich) were used as supplied.
[(h5-C5Me5)RhX(m-X)]2 (X=Br or I) was prepared as
described [8].
4.3. Preparations
4.3.1. Preparation of [LRhBr]+·BF−4 (4b)
A slurry of [(h5-C5Me5)RhBr(m-Br)]2 (0.080 g, 0.10
mmol) and dfppe (0.150 g, 0.20 mmol) in benzene (60
cm3) was refluxed under nitrogen for 122 h. The sol-
vent was removed by rotary evaporation to afford an
orange–yellow solid, which was extracted into
methanol (70 cm3) and filtered. NH4BF4 (1.00 g, 9.54
mmol) was added and the solution left for 2 h. The
solvent was removed by rotary evaporation and the
product extracted into dichloromethane (100 cm3) and
filtered. The solvent was removed by rotary evapora-
tion to yield an orange solid, which was washed with
chloroform (60 cm3) to yield 0.035 g of the yellow
product 4b, which was washed with light petroleum
and dried. A further 0.025 g was obtained on concen-
tration of the chloroform washings and addition of
light petroleum. Yield, 0.060 g, 27%. 1H [CD2Cl2,
3. Conclusion
In conclusion we have demonstrated that formation
of the cation [LRhCl]+, 4, from [(h5-C5Me5)RhBr(m-
Br)]2 and dfppe proceeds via the bromo-diphosphine
cation
5
and then the singly activated cation
[L%RhCl]+, 6, with concomitant formation of HF in
both ethanol and benzene. This corroborates mecha-
nistic data obtained for reaction (1) in ethanol and
suggests that the mechanism is identical in benzene.
We have also determined that the lack of reaction
observed on thermolysis of 2b in both non-polar and
polar aprotic solvents can ascribed to the influence of
the tetrafluoroborate anion, since substitution for
chloride or tetraphenylborate leads to the facile for-
mation of cation 1 in benzene.
2
300.14 MHz]: 4.47 (dm, JHꢀH% 18.3, 2H, CHH%C6F4),
3.76 (m, 2H, PCH2), 3.65 (d, JHꢀH% 18.3, 2H,
2
4
CHH%C6F4), 3.03 (m, 2H, PCH2), 2.17 (d, JPꢀH 7.6,
6H, 4- and 5-CH3), 1.04 (s, 3H, 2-CH3). 19F [CD2Cl2,
282.41 MHz]: −117.1 (s, 2F, C6F4), −134.65 (s, 2F),
−127.60 (br s, 2F, Fortho), −132.80 (br s, 2F, Fortho),
3
−143.44 (dt, JFꢀF 20.4, 10.4, 2F), −144.72 (t, JFꢀF

M.J. Atherton et al. / Journal of Organometallic Chemistry 582 (1999) 163–172
171
3
21.6, 2F), −152.26 (t, JFꢀF 21.6, 2F), −152.99 and
−153.04 (1:4) (4F, BF4−), −158.43 (br s, 4F, Fmeta).
4.3.4. Preparation of [(p5-C5Me5)RhCl(dfppe)]+·BPh−4
(2d)
Salt 2d was prepared as for 5b from [(h5-
31P{1H} [CD2Cl2, 121.50 MHz]: 68.5 (dm, JRhꢀP 142
1
Hz). FAB MS m/z: 1038, 1036 ([M−BF4+H]+), 956
([M−BF4−Br]+).
C5Me5)RhCl(m-Cl)]2 (0.125 g, 0.20 mmol), two equiva-
lents of NaBPh4 (0.165 g, 0.48 mmol) and dfppe (0.32
g, 0.42 mmol). The product was obtained as an orange
solid, contaminated by cations 1 and 3 from which it
could not be separated. Characterization is based on
the spectroscopic data and by comparison with 2b.
4.3.2. Preparation of [(p5-C5Me5)RhBr(dfppe)]+·BF−4
(5b)
The salt NH4BF4 (0.600 g, 5.72 mmol) was added to
[(h5-C5Me5)RhBr(m-Br)]2 (0.084 g, 0.11 mmol) in
methanol (50 cm3) and the mixture stirred for 10 min.
Dfppe (0.16 g, 0.21 mmol) in dichloromethane (10 cm3)
was added and the mixture stirred for 3 h. The solvents
were removed by rotary evaporation and the product
extracted into dichloromethane (80 cm3). The solution
was filtered and concentrated by rotary evaporation to
40 cm3. Addition of light petroleum precipitated 5b as a
yellow microcrystalline solid, which was filtered off and
washed with light petroleum and dried. Yield 0.185 g
(62%). 1H [(CD3)2CO, 300.14 MHz]: 3.50 (m, 4H, CH2),
1
Yield 0.30 g (ca. 55%). H [(CD3)2CO, 300.14 MHz]:
7.33 (m, 8H, BPh4), 6.91 (m, 8H, BPh4), 6.76 (t, J 7.2,
4
4H, para-BPh4), 3.40 (m, 4H, CH2), 1.72 (t, JPꢀH 4.8,
CH3). 19F [(CD3)2CO, 282.41 MHz]: −124.48 (br s, 4F,
3
Fortho), −127.47 (dm, JFꢀF 13.9, 4F, Fortho), −144.77
(m, 2F, Fpara), −146.03 (m, 2F, Fpara), −158.20 (m,
4F, Fmeta), −160.28 (m, 4F, Fmeta). 31P{1H}
1
[(CD3)2CO, 121.50 MHz]: 35.3 (dm, JRhꢀP 151 Hz).
FAB MS m/z: 1031 ([M−BPh4]+), 996 ([M−BPh4−
Cl]+), 319 (BPh4−).
4.3.5. Preparation of [(p5-C5Me5)RhCl(dfppe)]+·PF−6
(2e)
1.82 (t, JPꢀH 4.85, CH3). 19F [(CD3)2CO, 282.41 MHz]:
−124.20 (br s, 4F, Fortho), −127.43 (dm, JFꢀF 12.4,
4
3
The salt 2e was prepared as for 5b from [(h5-
C5Me5)RhCl(m-Cl)]2 (0.09 g, 0.15 mmol), KPF6 (0.91 g,
4.9 mmol) and dfppe (0.23 g, 0.30 mmol). The product
was obtained as a yellow powder, highly contaminated
by cations 1 and 3 from which it could not be sepa-
rated. Characterization is based on the spectroscopic
data and by comparison with 2b. Yield 0.21 g (ca.
4F, Fortho), −144.95 (m, 2F, Fpara), −146.25 (m, 2F,
Fpara), −150.87 and −150.93 (1:4) (4F, BF4−),
−158.32 (m, 4F, Fmeta), −160.32 (m, 4F, Fmeta).
31P{1H} [(CD3)2CO, 121.50 MHz]: 34.0 (dm, JRhꢀP 151
1
Hz). FAB MS m/z: 1077, 1075 ([M−BF4]+), 996
([M−BF4−Br]+). Anal. Found: C, 38.3; H, 1.6.
C36H19BBrF24P2Rh. Calc.: C, 38.65; H, 1.7%.
1
60%). H [(CD3)2CO, 300.14 MHz]: 3.40 (m, 4H, CH2),
4
1.74 (t, JPꢀH 4.8, CH3). 19F [(CD3)2CO, 282.41 MHz]:
4.3.3. Preparation of [(p5-C5Me5)RhCl(dfppe)]+·Cl−
(2a)
−71.78 (d, JPꢀF 706.6, PF6−), −124.47 (br s, 4F,
1
3
Fortho), −127.52 (dm, JFꢀF 13.9, 4F, Fortho), −144.82
3
The salt NH4Cl (1.0 g, 9.54 mmol) was added to 2b
(0.100 g, 0.11 mmol) in acetone (150 cm3) and water (20
cm3). The mixture was left at room temperature for 30
min, then filtered. The filtrate was concentrated by
rotary evaporation to yield a yellow emulsion, which
was extracted with CH2Cl2. After drying over anhy-
drous MgSO4, the CH2Cl2 extract was filtered and the
solvent removed by rotary evaporation to give a yellow
solid, which contained significant amounts of te-
trafluoroborate. The solid was treated with NH4Cl as
above and the process repeated. The resultant yellow
solid was a mixture comprising predominantly cation 2,
and also 1 and 3 with dfppe and small amounts of
BF4−. Yield ca. 0.08 g. The impure nature of the
product precluded characterization by elemental analy-
sis and characterization is based on NMR spectroscopic
(tm, JFꢀF 21.0, 2F, Fpara), −146.21 (m, 2F, Fpara),
−158.24 (m, 4F, Fmeta), −160.29 (m, 4F, Fmeta).
1
31P{1H} [(CD3)2CO, 121.50 MHz]: 35.1 (dm, JRhꢀP 155
1
Hz, dfppe), −138.8 (sept, JPꢀF 706.6, PF6−). FAB MS
m/z: 1031 ([M−PF6]+), 996 ([M−PF6−Cl]+), 145
(PF6−).
4.3.6. X-ray crystal structure determination
Crystals of salt 4c suitable for X-ray structure deter-
mination were grown from acetone. The structure was
determined at low temperature on a Siemens P4 diffrac-
tometer. Crystallographic data are presented in Table 1
and atomic coordinates and isotropic displacement
parameters are given in Table 2. Accurate unit cell
parameters were determined by least-squares refinement
of the optimised setting angles of 39 reflections in the
range 55q512.5°. Three standard check reflections,
monitored every 97 reflections, indicated no crystal
decay. The data were corrected for absorption, Lorentz
and polarization effects.
1
data by comparison with 2b. H [(CDCl3, 300.01 MHz]:
2.56 (m, 4H, CH2), 1.56 (t, 4JPꢀH 4.6, CH3). 19F [CDCl3,
282.26 MHz]: −125.68 (br s, 4F, Fortho), −128.51
3
3
(dm, JFꢀF 10.2, 4F, Fortho), −141.87 (t, JFꢀF 20.2, 2F,
3
Fpara), −143.85 (t, JFꢀF 20.5, 2F, Fpara), −158.28 (m,
The structure was solved by direct methods using the
4F, Fmeta), −159.68 (m, 4F, Fmeta). 31P{1H} [CDCl3,
PATT option of SHELXTL
-
PC [9] and refined by full
1
matrix least squares on F2. The disordered C6F5 rings
121.45 MHz]: 30.5 (dm, JRhꢀP 152 Hz).

172
M.J. Atherton et al. / Journal of Organometallic Chemistry 582 (1999) 163–172
Karac¸ar, D.R. Russell, G.C. Saunders, Chem. Commun. (1995)
191.
[2] M.J. Atherton, J. Fawcett, J.H. Holloway, E.G. Hope, A.
Karac¸ar, D.R. Russell, G.C. Saunders, J. Chem. Soc. Dalton
Trans. (1996) 3215.
were modelled as idealized planar rings (CꢀC 1.39, CꢀF
,
1.34 A) and refined as rigid groups with individual atom
anisotropic displacement parameters. Hydrogen atoms
,
were included in calculated positions (CꢀH 0.99 A) and
refined by a riding model. All non-hydrogen atoms were
refined with anisotropic displacement parameters. An
analysis of the weighting scheme over ꢀF0ꢀ and (sin q)/u
was satisfactory.
[3] M.J. Atherton, J.H. Holloway, E.G. Hope, G.C. Saunders, J.
Organomet. Chem. 558 (1998) 209.
[4] M.J. Atherton, J. Fawcett, J.H. Holloway, E.G. Hope, S.M.
Martin, D.R. Russell, G.C. Saunders, J. Organomet. Chem. 555
(1998) 67.
[5] P.W. Atkins, Physical Chemistry, Oxford University Press, Ox-
ford, 1978.
[6] J.A. Riddick, W.B. Bunger, in: A. Weissberger (Ed.), Tech-
niques of Chemistry, vol. II, Wiley-Interscience, New York,
1970.
[7] R.P. Houghton, M. Voyle, R. Price, Chem. Commun. (1980)
884; J. Chem. Soc. Perkin Trans. I (1984) 925.
[8] D.S. Gill, P. Maitlis, J. Organomet. Chem. 87 (1975) 359.
Acknowledgements
The authors thank F2 Chemicals (G.C.S.) and the
Royal Society (E.G.H.) for support.
[9] G.M. Sheldrick, SHELXTL
-PC, Release 5.03, Siemens Analytical
References
X-ray Instruments, Madison, WI, 1996.
[1] M.J. Atherton, J. Fawcett, J.H. Holloway, E.G. Hope, A.
.