Syntheses of the complete set of isomerically pure, partially fluorinated cyclopentadienyl ligands (C5F5-nHn) [n = 1-4] by flash vacuum thermolysis of (η5-Oxocyclohexadienyl)ruthenium complexes. Molecular structures of [Ru(η5-C5Me5) (η5-C5-1,2-F2H3)]

Article abstract of DOI:10.1021/om970775p

A series of partially fluorinated (pentamethylcyclopentadienyl)(η⁵-oxocyclohexadienyl)-ruthenium complexes [RuCp*(2-6-η⁵-C₅F_5-nH_nCO)] (n = 1-4; Cp* = C₅Me₅) has been prepared by treatment of the cation [RuCp*(MeCN)₃]⁺ with thallium(I) salts of the corresponding phenols. Depending upon the number and location of the fluorine substituents, the complexes hydrogen bond to one-half of a molecule, one molecule, or no molecules of water. Subjection of these compounds to flash vacuum thermolysis (FVT) results in extrusion of CO and selective formation of the corresponding complexes [RuCp*(η⁵-C₅F_5-nH_n)] containing the monofluorocyclopentadienyl, 1,2-difluorocyclopentadienyl, 1,3-difluorocyclopentadienyl, 1,2,3-trifluorocyclopentadienyl, 1,2,4-trifluorocyclopentadienyl, and tetrafluorocyclopentadienyl ligands. ¹H, ¹⁹F, and ¹³C NMR spectra are reported for the new cyclopentadienyl ligands, and the ¹³C-¹⁹F coupling constants thus obtained are used to generate a full simulation of the ¹³C NMR spectrum of the pentafluorocyclopentadienyl ligand in [Ru(C₅Me₅)(C₅F₅)]. The solid-state structures of two complexes containing partially fluorinated cyclopentadienyl complexes have been determined by X-ray crystallography: [RuCp*(η⁵-C₅-1,2-F₂H₃)] (15), and [RuCp*(η⁵-C₅FH₄)] (17).

Full text of DOI:10.1021/om970775p

Organometallics 1998, 17, 457-465
457
Syn th eses of th e Com p lete Set of Isom er ica lly P u r e,
P a r tia lly F lu or in a ted Cyclop en ta d ien yl Liga n d s
(C5F 5-n Hn ) [n ) 1-4] by F la sh Va cu u m Th er m olysis of
(η5-Oxocycloh exa d ien yl)r u th en iu m Com p lexes.
Molecu la r Str u ctu r es of [Ru (η5-C5Me5)(η5-C5-1,2-F 2H3)]
a n d [Ru (η5-C5Me5)(η5-C5H4F )]
Russell P. Hughes,* Xiaoming Zheng, Christopher A. Morse,
Owen J . Curnow, and J effrey R. Lomprey
Department of Chemistry, Burke Laboratory, Dartmouth College,
Hanover, New Hampshire 03755-3564
Arnold L. Rheingold and Glenn P. A. Yap
Department of Chemistry, University of Delaware, Newark, Delaware 19716
Received September 2, 1997
A series of partially fluorinated (pentamethylcyclopentadienyl)(η5-oxocyclohexadienyl)-
ruthenium complexes [RuCp*(2-6-η5-C5F5-nHnCO)] (n ) 1-4; Cp* ) C5Me5) has been
prepared by treatment of the cation [RuCp*(MeCN)3]+ with thallium(I) salts of the
corresponding phenols. Depending upon the number and location of the fluorine substituents,
the complexes hydrogen bond to one-half of a molecule, one molecule, or no molecules of
water. Subjection of these compounds to flash vacuum thermolysis (FVT) results in extrusion
of CO and selective formation of the corresponding complexes [RuCp*(η5-C5F5-nHn)]
containing the monofluorocyclopentadienyl, 1,2-difluorocyclopentadienyl, 1,3-difluorocyclo-
pentadienyl, 1,2,3-trifluorocyclopentadienyl, 1,2,4-trifluorocyclopentadienyl, and tetrafluo-
rocyclopentadienyl ligands. 1H, 19F, and 13C NMR spectra are reported for the new
cyclopentadienyl ligands, and the 13C-19F coupling constants thus obtained are used to
generate a full simulation of the 13C NMR spectrum of the pentafluorocyclopentadienyl ligand
in [Ru(C5Me5)(C5F5)]. The solid-state structures of two complexes containing partially
fluorinated cyclopentadienyl complexes have been determined by X-ray crystallography:
[RuCp*(η5-C5-1,2-F2H3)] (15), and [RuCp*(η5-C5FH4)] (17).
In tr od u ction
by decomposition of the BF4 salt to afford the monof-
luorocyclopentadienyl ligand.3 However, the compound
was characterized only by its IR spectrum, microanaly-
For a number of years we have been interested in
evaluating the effect of fluorine as a ligand substituent
in transition-metal organometallic chemistry.1 The
σ-electron-withdrawing and π-electron-donating sub-
stituent effects of fluorine in organic molecules have
been thoroughly examined and reviewed.2 The chemical
introduction of fluorine as a ligand substituent directly
into organometallic systems has, in most cases, been
unsuccessful, probably due to the strongly oxidizing
conditions required, although one or two examples can
be found in the literature. The first reported example,
by Cais and Narkis, of a compound containing a
fluorinated cyclopentadienyl ligand, [Mn(η5-C5H4F)-
(CO)3], appeared in 1965.3 Introduction of fluorine was
effected by a classical organic conversion of an amine
substituent on the Cp ring to a diazonium salt followed
1
sis for C, H, and Mn, and a melting point; no H NMR
and 19F NMR spectra were obtained. The relatively
harsh conditions required for the conversion of the
amino group on the Cp ring to a diazonium group
severely limited the application of this synthetic method
to the preparation of other fluorocyclopentadienyl com-
plexes. The second example, monofluoroferrocene, was
reported by Hedberg and Rosenberg,4 also by direct
functionalization of a coordinated Cp ligand. Treatment
of bromoferrocene with n-butyllithium gave lithiofer-
rocene, which was then allowed to react with perchloryl
fluoride gas in THF. The product, isolated in less than
10% yield, was identified as monofluoroferrocene by
NMR and mass spectra. Alternatively, the same com-
pound was obtained in 37% yield by Popov et al. in 1990
by reaction of trifluoroacetyl hypofluorite with mono-
(acetoxymercurio)ferrocene.5 It is worth mentioning
that while perchlorination of some metallocenes can be
(1) Hughes, R. P. Adv. Organomet. Chem. 1990, 31, 183.
(2) Smart, B. E. In The Chemistry of Functional Groups, Supplement
D Patai, S., Rappoport, Z., Eds.: J ohn Wiley & Sons: New York, 1983;
Chapter 4. Smart, B. E. Organofluorine Chemistry, Principles and
Applications; Banks, R. E., Smart, B. E., Tatlow, J . L., Eds; Plenum:
New York, 1994; Chapter 3.
(4) Hedberg, F. L.; Rosenberg, H. J . Organomet. Chem. 1971, 28,
C14
(3) Cais, M.; Narkis, N. J . Organomet. Chem. 1965, 3, 269.
(5) Popov, V. I.; Lib, M.; Hass, A. Ukr. Khim. Zh. 1990, 56, 1115.
S0276-7333(97)00775-9 CCC: $15.00 © 1998 American Chemical Society
Publication on Web 01/08/1998

458 Organometallics, Vol. 17, No. 3, 1998
Hughes et al.
similarly accomplished by mercuration followed by
halogen substitution, analogous perfluorination of Cp
ligands has not been accomplished.6-8
ficant differences, corresponding syntheses of isomeri-
cally pure partially fluorinated cyclopentadienyl rings
were required. Here, we describe the selective prepara-
tion of the complete set of complexes containing partially
fluorinated cyclopentadienyl rings and their oxocyclo-
hexadienyl precursors. Studies confirming the approxi-
mate additivity of fluorine substituent effects in these
new cyclopentadienyl compounds were recently pub-
lished.23
In view of the challenges associated with the intro-
duction of fluorine directly into organometallic mol-
ecules, we have limited our synthetic approaches to
those which have all of the necessary carbon-fluorine
bonds in place before complexation of the ligand. In the
case of the perfluorocyclopentadienyl ligand, this posed
-
a significant problem; although the C5F5 anion is
known,9 attempts to use it as a precursor to transition-
metal complexes proved to be frustratingly unsuccess-
ful.10 Our eventual synthetic approach to fluorinated
cyclopentadienyl ligands resulted from the realization
that certain six-membered organic rings underwent ring
contraction to give five-membered rings, with elimina-
tion of CO, under conditions of flash vacuum thermoly-
sis.11 For example, in this fashion, o-benzoquinones are
converted to cyclopentadienones,12-14 2-pyrone is trans-
formed to furan,15 and hexafluorocyclohexadienones
afford hexafluorocyclopentadiene.16 Coupling of this
latent knowledge with subsequent reports of the prepa-
ration of ruthenium complexes containing the η5-oxo-
cyclohexadienyl (η5-phenoxide) ligand17,18 suggested that
analogous decarbonylations of the latter complexes to
give η5-cyclopentadienyl ligands could be achieved,
especially if they afforded stable metallocenes as the
final organometallic product. Indeed, initial attempts
were successful in affording the pentafluorocyclopenta-
dienyl complexes 119 and 220 via flash vacuum ther-
molysis (FVT) of the corresponding complexes 3 and 4,
Chart 1. The effects of perfluorination on the electronic
properties of the cyclopentadienyl ligand were evaluated
by measuring the gas-phase ionization free energy of
1. It proved to be 18.5 kcal/mol higher than that of [Ru-
(C5Me5)(C5H5)], confirming that perfluorination of a Cp
ligand increases its electron-withdrawing ability, al-
though the effects of σ-electron withdrawal are, not un-
expectedly, strongly attenuated by π-donation from the
fluorines.21 Similar conclusions result from a detailed
comparisons of the photoelectron spectra of 1 and its
analogues [Ru(C5Me5)(C5H5)] and [Ru(C5Me5)(C5Cl5)].22
In order to establish whether the effects of perfluori-
nation of the cyclopentadienyl were additive and whether
the specific location of fluorines on the ring made signi-
Resu lts a n d Discu ssion
P a r tia lly F lu or in a ted (2-6-η5-Oxocycloh exa d i-
en yl)r u t h en iu m
Com p lexes
[R u Cp *(2-6-η5-
C5F 5-n Hn CO)] (n ) 1-4). The desired oxocyclohexa-
dienyl complexes, [RuCp*(2-6-η5-C5F5-nHnCO)] n )
1-4, 5-11 (see Table 2), were synthesized similarly to
their perfluorinated analogue 3.19 Refluxing the ruthe-
nium tetramer [RuCp*Cl]4 (Cp* ) η5-C5Me5) in aceto-
nitrile afforded, in situ, the [RuCp*(CH3CN)3]+ cation,
which was subsequently treated with the Tl+ salts of
the appropriate commercially available fluorophenols to
afford the corresponding partially fluorinated oxocyclo-
hexadienyl complexes 5-11 in good yields.
All oxocyclohexadienyl products were characterized
by IR, 1H and 19F NMR spectra, and microanalysis; data
1
are presented in Tables 1 and 2. The H NMR spectra
of 5-11 all exhibit a sharp singlet in the region 1.79-
1.90 ppm for the Cp* protons, with multiplets in the
region 3.9-5.7 ppm corresponding to the protons of the
fluorooxocyclohexadienyl ligands. Their 19F NMR spec-
tra displayed the expected number of resonances in the
appropriate ratio in the region from -160 to -194 ppm
the significant upfield shift of the fluorine resonances
from those of the corresponding free fluorophenols being
consistent with η5 coordination of the phenoxide ligands.
1
The 19F-19F, 19F-1H, H-1H coupling constants listed
in Table 2 were obtained directly from the H and 19F
1
NMR spectra.
In confirmation of the 2-6-η5 coordination mode of
the phenoxide ligands, the IR spectra of 5-11 showed
a strong CdO stretch. The position of this stretch is
dependent upon the number and relative location of the
fluorine atoms on the ligand. For the fully fluorinated
ring in 3, the observed value of υCO is 1620 cm-1 19
as
;
expected, all values observed for 5-11 lie below this
value and are consistent with a predominantly inductive
electron-withdrawing effect of fluorine, exercised more
potently at locations closer to the oxygen atom. The
presence of two ortho-fluorines, in 5 and 9, give the
highest values of υCO (1582-1574 cm-1; Table 1), a
single ortho-fluorine (in 6, 7, 8, and 10) affords values
between 1541 and 1555 cm-1), and no ortho-fluorines
(in 11) give the lowest value of 1534 cm-1. In addition
to CdO stretches, broad absorption bands around 3450
cm-1 attributable to water were also observed in the IR
spectra of 6-11. Consistent with this observation, the
microanalysis data of the compounds (except for 7 and
8) did not correlate to the calculated percentages of
(6) Hedberg, F. L.; Rosenberg, H. J . Am. Chem. Soc. 1970, 92, 3239.
(7) Hedberg, F. L.; Rosenberg, H. J . Am. Chem. Soc. 1973, 95, 870.
(8) Winter, C. H.; Han, Y.; Ostrander, R. L.; Rheingold, A. L. Angew.
Chem., Int. Engl. 1993, 32, 1161.
(9) Paprott, G; Seppelt, K. J . Am. Chem. Soc. 1984, 106, 4060.
(10) Paprott, G.; Lehmann, S.; Seppelt, K. Chem. Ber. 1988, 121,
727.
(11) For a discussion of this technique, see: Magrath, J .; Fowler,
F. W. Tetrahedron Lett. 1988, 29, 2171 and references cited therein.
(12) DeJ ongh, D. C.; Van Fossen, R. Y.; Bourgeois, C. F. Tetrahedron
Lett. 1967, 271.
(13) DeJ ongh, D. C.; Brent, D. A.; Van Fossen, R. Y. J . Org. Chem.
1971, 36, 1469.
(14) Chapman, O. L.; McIntosh, C. L. J . Chem. Soc., Chem. Com-
mun. 1971, 770.
(15) Brent, D. A.; Hribar, J . D.; DeJ ongh, D. C. J . Org. Chem. 1970,
35, 135.
(16) Soelch, R. R.; Mauer, G. W.; Lemal, D. M. J . Org. Chem. 1985,
50, 5845.
(17) Bu¨cken, K.; Koelle, U.; Pasch, R.; Ganter, B. Organometallics
1996, 15, 3095.
(18) Loren, S. D.; Campion, B. K.; Heyn, R. H.; Tilley, T. D.; Bursten,
B. E.; Luth, K. W. J . Am. Chem. Soc. 1989, 111, 4712.
(19) Curnow, O. J .; Hughes, R. P. J . Am. Chem. Soc. 1992, 114, 5895.
(20) Hughes, R. P.; Zheng, X.; Rheingold, A. L.; Ostrander, R. L.
Organometallics 1994, 13, 1567.
(21) Richardson, D. E.; Ryan, M. F.; Geiger, W. E.; Chin, T. T.;
Hughes, R. P.; Curnow, O. J . Organometallics 1993, 12, 613.
(22) Lichtenberger, D. L.; Elkadi, Y.; Gruhn, N.; Hughes, R. P.;
Curnow, O. J .; Zheng, X. Organometallics 1997, 16, 5209.
(23) Richardson, D. E.; Lang, L.; Eyler, J . R.; Zheng, X.; Morse, C.
A.; Hughes, R. P. Organometallics 1997, 16, 149.

Synthesis of C5F5-nHn Ligands
Organometallics, Vol. 17, No. 3, 1998 459
Ch a r t 1
Ta ble 1. IR a n d Micr oa n a lysis Da ta for 5-11
IR spectra and microanalyses of 6-11 must have been
incorporated after the samples had been prepared.
However, on the basis of the assumption that they were
as air stable in the solid state as the perfluorinated
analogue 3, compounds 5, 6, and 9-11 were exposed to
the air in the subsequent purification (sublimation,
chromatography) and characterization processes. The
water affinity of 6-11 likely originates from the hydro-
gen-bonding tendency of the oxygen atom in the oxocy-
clohexadienyl ligand. Indeed, the hydrogen-bonding
ability of the oxygen atom in analogous oxocyclohexa-
dienyl ligands is well-established. X-ray crystallo-
graphic studies have demonstrated that complexes 18
and 19 show this hydrogen-bonding ability to phenol,24,25
and NMR studies have also illustrated a 1:1 solution
interaction between D2O and a series of oxocyclohexa-
dienyliron complexes like 20.26 The number of water
microanalysis (%)
calcd
found
IR (cm-1
)
C
H
C
H
νCdO
νO-H
5
6
7
8
9
47.89
4.02
48.00
48.92
50.09
52.60
51.71
50.19
52.82
3.98
4.90
4.40
5.24
4.94
5.06
5.84
1582
1548
1555
1543
1574
1541
1534
48.97a
50.13
4.63a
4.47
3400-3500
3400-3500
3400-3500
3400-3500
3400-3500
3300-3500
52.59
4.97
51.33a
50.12b
52.59b
5.12a
5.26b
5.89b
10
11
a
b
Calculated using 0.5 H2O per formula unit. Calculated using
1.0 H2O per formula unit.
carbon and hydrogen unless one molecule (for 10, 11)
or one-half (for 6, 9) of a molecule of water was added
to their molecular formula. In the cases of 7 and 8,
special efforts were taken to exclude moisture during
the purification process (see below). The 1H NMR
spectra of some complexes revealed broad resonances
at about 3 ppm, confirmed as being due to hydroxylic
protons by their disappearance upon addition of D2O.
Since the compounds were synthesized in predried
solvents under dry nitrogen, the water evident from the
(24) Chaudret, B.; He, X.; Huang, Y. J . Chem. Soc., Chem. Commun.
1989, 1844. Koelle, U.; Wang, M. H.; Raabe, G. Organometallics 1991,
10, 2573.
(25) Cole-Hamilton, D. J .; Young, R. J .; Wilkinson, G. J . Chem. Soc.,
Dalton Trans. 1976, 1995.
(26) Piorko, A.; Zhang, C. H.; Reid, R. S.; Lee, C. C.; Sutherland, R.
G. J . Organomet. Chem. 1990, 395, 293.

460 Organometallics, Vol. 17, No. 3, 1998
Ta ble 2. 1H a n d 19F NMR Sp ectr a l Da ta for 5-11a
δH
Hughes et al.
δF
Cp*(s) other
OC6F5-nHn
J FF
J HF
J HH
-170.5 (m, 2F)
-188.0 (m, 2F)
1.90
5.46 (bs)
-180.1 (dd, F2)
-185.8 (td, F1)
-191.9 (ddd, F3)
1.85
2.7 (s, H2O)
4.67 (ddd, H1)
5.22 (dd, H2)
J F ) 37.4
J H ) 4.3
J H ) 6.3
1H2
1F2
1F3
J F ) 8.3
1F3
J H ) 2.1
2F1
J F ) 38.7
2F3
J H or J H ) 1.1
1F2
1F1
-168.2 (ddd, F2)
-172.4 (d, F1)
-180.4 (d, F3)
1.83
5.15 (bs, H2)
5.69 (t, H1)
J F ) 10
J H ) 2.2
1F2
1F2
J F ) 0
1F3
J F ) 35
2F3
-170.3 (d, F2)
-193.6 (d, F1)
1.88
2.8 (s, H2O)
4.8 (m, 2H)
5.11 (m, 1H)
J F ) 35.2
1F2
-169.7 (bs)
1.87
4.69 (tt, H2)
5.25 (dt, H1)
(J H
J H ) 1.8
2F1
)
) 2.7
1F1 app
J H ) 5.4
1H2
-162.4 (bs, F1)
-171.3 (bs, F2)
1.79
3.0 (bs, H2O)
4.12 (dt, H2)
5.12 (dq, H3)
5.58 (dd, H1)
J H ) 6
J H ) 2
2F2
1H2
J H ) 1
1H3
J H ) 1.5
3F2
J H ) 1.5
3F1
J H ) 6
2H3
-162.9 (bs)
1.86
3.0 (bs, H2O)
4.64 (m, 2H)
5.26 (m, 2H)
a
Solvent: CDCl3; Units: δ in ppm; J in hertz. For other general information on the NMR experiments, see the Experimental Section.
Subscript app stands for apparent bs for broad singlet.
molecules associated with 6, 9, 10, and 11, as indicated
by microanalysis data, correlates inversely to υCdO and
also to the number and relative position of F substitu-
ents on the fluorooxocyclohexadienyl ligand. If the
hydrogen-bonding capability of the oxocyclohexadienyl
ligand reflects the amount of negative charge residing
on the oxygen atom, F substituents will clearly compete
for the negative charge of the ligand and reduce the
hydrogen-bonding ability of the oxygen. More F sub-
stituents closer to oxygen lead to a less anionic oxygen
atom, a stronger CdO bond, and a lower tendency for
the oxygen to form hydrogen bonds.
P a r tia lly F lu or in a ted Cyclop en ta d ien ylr u th e-
n iu m Com p lexes [Ru Cp *(η5-C5F 5-n Hn )] (n ) 1-4).
In a manner analogous to the FVT conversion of the
perfluorinated precursors 3 and 4 to their η5-C5F5
complexes 119 and 2,20 each of the partially fluorinated
precursors 5-11 afforded the partially fluorinated cy-
clopentadienylruthenium complexes 12-17 (see Table
3). As observed in the FVT of 3 and 4, thin films of a
black material were formed in the quartz pyrolysis
tubes. This material was characterized by qualitative
X-ray microanalysis and found to be composed of mostly
Ru and C and traces of O and Si. Yields and conversions
were variable and difficult to control in a reproducible
fashion, making this a less than optimum synthetic
method. For example, FVT of the tetrafluoro complex
5 cleanly afforded 12 in 42% yield, whereas FVT of
monofluorinated 11 afforded 17 in only 9% yield. Yields
and conversions were not optimized, and details of the
furnace temperatures, yields, and conversions are pre-
sented in the Experimental Section. All thermolysis
reactions proceeded cleanly to give the fluorine sub-
stituent pattern on the resultant cyclopentadienyl ring,
consistent with that expected from the precursor com-
plex(es). In some cases, traces of isomeric impurities
and complexes containing fewer fluorines than expected
could be observed by 19F NMR spectroscopy at levels
between 1-2%; in all cases, these impurities could be
traced back to impurities in the starting phenols, which
were purchased commercially and not purified before
use. We conclude that in the actual thermolysis experi-
ments no rearrangements or elimination reaction occur
to any significant extent.
The partially fluorinated cyclopentadienyl products
12-17 were characterized by 1H, 19F, and 13C{1H} NMR
spectra (Table 3). The molecular formulas of 12-14
were confirmed by microanalysis. The 1H NMR spectra
all show a Cp* proton resonance in the region from 1.88
to 1.74 ppm; the corresponding resonances in 1 and
[RuCp*(C5H5)] appear at 1.68 and 1.90 ppm, respec-
tively. The resonances corresponding to the protons on
the fluorocyclopentadienyl ligands appear in the region
3.51-4.43 ppm. All of the 19F NMR spectra of 12-17

Synthesis of C5F5-nHn Ligands
Organometallics, Vol. 17, No. 3, 1998 461

462 Organometallics, Vol. 17, No. 3, 1998
Hughes et al.
recognizing that the presence of a single 13C nucleus in
the η5-C5F5 ring should lead to different chemical
shift perturbations for the different fluorines in the
ring.27 The successful simulation data are presented
in Figure 1.
The solid-state structures of two of these partially
fluorinated metallocenes, 15 and 17, were confirmed by
X-ray crystallographic studies. Details of the crystal-
lographic determinations are provided in Table 4. The
ORTEP diagram of 15 is shown in Figure 2, and a
schematic diagram showing relevant bond lengths
within the cyclopentadienyl rings is shown in Figure 3.
The two C5 rings are eclipsed, consistent with the
structures of ruthenocene28 and [Ru(C5Me5)(C5H5)].29
Unlike ruthenocene itself,28 the two rings in 15 are not
parallel. Examination of the Ru-C bond distances for
the difluorocyclopentadienyl ligand in 15 shows the
three Ru-C(H) distances (2.186(9), 2.189(10), and 2.197-
(11) Å) to be identical to the average Ru-C(H) distance
(2.191(7) (Å)) in ruthenocene or in [Ru(C5Me5)(C5H5)]
(2.190(5) Å). However, one of the two Ru-C(F) dis-
tances in 15 is considerably shorter (2.133(11) Å) and
the other slightly shorter (2.164(10) Å). Correspond-
ingly, the five Ru-C(Me) distances show a significant
inequivalence, with the two carbon atoms eclipsing the
fluorine bearing carbon atoms being significantly closer
(2.122(13) and 2.123(13)Å) to the metal center. The
overall result is that the metallocene rings become
slightly canted. The shorter Ru-C(F) distances are
consistent with, although less significant than, the
shortening of M-C distances in unsaturated fluorocar-
bon ligands such as η2-C2F430 and η5-C5F520 as compared
to their hydrocarbon analogues.1
F igu r e 1. (a) Experimental and (b) simulated 13C NMR
resonance for the pentafluorocyclopentadienyl ring carbon
of 1. Simulated as an AXYY′ZZ′ spin system (see labeling
scheme) with J AX ) 295.00, J AY ) J AY′ ) -16.00, J AZ
J AZ′ ) 2.00, J XY ) J XY′ ) 23.00, J XZ ) J XZ′ ) 19.00, J YY′
)
)
19.00, J ZZ′ ) 23.00, J YZ ) J Y′Z′ ) 23.00, J YZ′ ) J Y′Z ) 19.00;
δ(13C) ) 7777.50, δFX ) -59896.00, δFY ) δFY′ ) -59826.00
Hz, δFZ ) δFZ′ ) -59820.00 (all values in hertz).
display the expected number of resonances in the right
ratios in the region from -189 to -216 ppm. By way
of comparison, the fluorine chemical shifts of 1 and 2
are -213.2 and -206.9 ppm, respectively. The chemical
shifts of the fluorine substituents in 12-17 and in their
Cp* analogue 1 are almost independent of the total
number of fluorines but do depend on their relative
positions: a fluorine vicinally adjacent to no other
fluorines has a chemical shift of about -190 ppm; a
fluorine adjacent to one other fluorine has a chemical
shift of around -204 ppm; and a fluorine sandwiched
between two other fluorines has a chemical shift of
approximately -214 ppm. The 13C{1H} NMR spectra
of 12-17 all exhibit two singlets in the region 11-12
and 86-91 ppm corresponding, respectively, to the ring
carbon and methyl substituents of the Cp* ligand.
Resonances due to the fluorinated cyclopentadienyl
ligands appear in two separate regions: 108-135 ppm
for carbon atoms bearing a F substituent and 44-66
ppm for the unsubstituted carbon atoms. All 13C-19F,
The structure of 17 is similar to that of ruthenocene,
although disorder problems thwarted a more detailed
comparison. The fluorine atom in 17 was located at
three disordered positions with a 56-28-16% distribu-
tion, the Cp* ligand was treated as a rigid body
disordered in two rotational positions with a 70-30%
distribution. An ORTEP diagram illustrating the prin-
cipal occupancy fluorine site is provided in Figure 4. In
view of this disorder, a detailed structural analysis is
not presented.
Con clu sion s
The FVT reaction of a series of partially fluorinated
oxocyclohexadienyl complexes does provide a selective
method for the synthesis of partially fluorinated cyclo-
pentadienyl complexes, and the complete set of partially
fluorinated cyclopentadienyl rings has thereby been
obtained. While the method is not as high yield as could
be desired, it appears to be the only method currently
available for synthesis of selectively fluorinated ligands
of this type. The selectively fluorinated complexes thus
prepared have been subjected to a study of their gas-
phase ionization energetics, the results of which have
been published separately.23 The series of 13C-19F
coupling constants for fluorocyclopentadienyl ligands
1
19F-19F, 19F-1H, and H-1H coupling constants listed
in Table 3 were obtained directly from the spectra.
The long range 19F-13C coupling constants for the
partially fluorinated cyclopentadienyl ligands provided
useful reference data to help simulate the 13C NMR
spectrum of the originally reported pentafluorocyclo-
pentadienyl complex 1. In the initial report,19 the
carbon resonance of the fluorinated ring was reported
1
as a broad doublet, with only the value of J CF being
(27) We are grateful to Prof. David M. Lemal for this suggestion.
(28) Seiler, P.; Dunitz, J . Acta Crystallogr. 1980, B36, 2946.
(29) Zanin, I. E.; Antipin, M. Y.; Struchkov, Y. T. Kristallografiya
1991, 36, 420.
(30) Guggenberger, L. J .; Cramer, R. J . Am. Chem. Soc. 1972, 94,
3779.
readily apparent. Under higher resolution conditions,
a much more complex pattern emerged, as shown in
Figure 1. This was eventually simulated as an AX-
YY′ZZ′ spin system (labeling as shown in Figure 1) by

Synthesis of C5F5-nHn Ligands
Organometallics, Vol. 17, No. 3, 1998 463
Ta ble 4. Cr ysta llogr a p h ic Da ta for 15 a n d 17
15
17
(a) Crystal Parameters
formula
fw
C
337.4
15H18F2Ru
C15H19FRu
319.4
cryst syst
space group
a, Å
b, Å
c, Å
monoclinic
P21/n
monoclinic
P21/n
8.376(2)
14.191(4)
12.013(5)
93.20(3)
1425.8(8)
4
8.320(4)
13.976(5)
11.948(6)
92.37(3)
1388.1(11)
4
b, deg
V, Å3
Z
cryst dimens, mm
cryst color
D(calcd), g cm3
µ(Mo KR), cm-1
temp, K
0.24 × 0.38 × 0.60
0.40 × 0.40 × 0.40
yellow
yellow
1.572
11.02
296
1.528
11.17
293
(b) Data Collection
diffractometer
monochromator
radiation
2θ scan range, deg
data collected (h,k,l)
no. of rflns collected
no. of indep rflns
Siemens P4
graphite
Mo KR (λ ) 0.710 73 Å)
4-52
4-50
(10, +17, +14
(9, +16, +14
2551
2935
2802
2433
no. of indep obsd rflns, Fo > nσ(Fo)
std rflns std/rfln
1874 (n ) 5)
3/197
1941 (n ) 4)
3/197
variance in stds, %
4
3
(c) Refinement
R(F), %
R(wF), %
∆/σ(max)
∆(F), e Å-3
No/Nv
6.71a
9.54a
0.123
2.14
4.11b
11.94b
0.328
1.025
11.75
1.12
11.5
1.74
GOF
Quantity minimized: ∑w(Fo - Fc)2. Quantity minimized: ∑[w(Fo - Fc2)2]/∑[w(Fo2)2]1/2
.
a
b
2
F igu r e 2. ORTEP diagram with labeling scheme for 15.
Thermal ellipsoids are drawn at 30% probability.
proved to be invaluable reference data in facilitating a
simulation of the 13C NMR resonance of the pentafluo-
rocyclopentadienyl ring in 1.
F igu r e 3. Key bond distances (Å) in complex 15.
Schlenk techniques, under an atmosphere of nitrogen which
had been deoxygenated over BASF catalyst and dried over
Aquasorb. All solvents were obtained from Fisher Scientific
and distilled under nitrogen over one of the following drying
Exp er im en ta l Section
Gen er a l Con sid er a tion s. All reactions except FVT experi-
ments were performed in oven-dried glassware, using standard

464 Organometallics, Vol. 17, No. 3, 1998
Hughes et al.
petroleum ether. Yields were typically 90-100%. These
complexes were used for the preparation of 5-11 without being
further characterized.
Syn th eses of [Ru Cp *(2-6-η5-C5F 5-n Hn CO)] (n ) 1-4).
Compounds 5-11 were prepared following a modified proce-
dure for the preparation of pentafluorooxocyclohexadienyl
complex 3.19 Tables 1 and 2 in the main text contain the
1
microanalysis, IR, and H and 19F NMR spectra data of 5-11.
Microanalysis samples were purified by sublimation. The
following general procedure is outlined for 5:
An deep orange acetonitrile solution of [RuCp*(CH3CN)3]-
Cl was obtained by refluxing an acetonitrile (100 mL) solution
of [RuCp*Cl]4 (0.75 g, 2.76 mmol Ru) for 3 h. Previously
prepared [Tl(2,3,5,6-OC6F4H)] (1.02 g, 2.75 mmol) was added,
and the resultant mixture was refluxed overnight to give a
light yellow solution with a white precipitate (TlCl). The
mixture was filtered, the solvent was removed in vacuo, and
the remaining residue was extracted with methylene chloride
(65 mL). Evaporation of the extract afforded a tan solid, which
was subsequently sublimed (160 °C, 10-2 Torr) to yield a white
solid 5 (0.64 g, 1.59 mmol; 58%).
Similarly prepared were 6 (87%), 7 (74%), 8 (75%), 9 (46%),
10 (49%), and 11 (39%).
Syn th eses of [Ru Cp *(η5-C5F 5-n Hn )] (n ) 1-4). All
percentage yields of the products are based on the amounts of
the precursors consumed in the cases of less than 100%
conversion. 1H, 19F, and 13C{1H} NMR spectral data of 12 are
provided in Table 3. A typical procedure for the conversion of
5 to 12 follows:
F igu r e 4. ORTEP diagram with labeling scheme for 17.
Thermal ellipsoids are drawn at 30% probability. The
fluorine atom was located at three disordered positions
with 56-28-16% distribution; only the 56% occupancy site
is shown.
Compound 5 (0.13 g, 0.31 mmol) was evaporated under
dynamic vacuum (10-4 Torr) and mild heating (120 °C) at the
sealed end of a quartz tube (4 ft in length × 0.95 cm i.d.) and
was allowed to pass through the tube with all three zones set
at 710 °C. A shiny black film was deposited on the inside wall
of the highly heated part of the tube (see text). Solids
condensed at ambient temperature in the exit tube and were
rinsed from the tube with methylene chloride. 19F NMR
spectra of the crude extracts in CDCl3 showed the resonances
of 12 and traces of 5. Purification by sublimation (105-110
˚C, 10-2 Torr) afforded a pale yellow solid 12 (0.05 g, 0.13
mmol) in 42% yield. Anal. Calcd for C15H16F4Ru: C, 50.50;
H, 4.82. Found: C, 50.40; H, 4.91.
agents: THF and diethyl ether over K/benzophenone; meth-
ylene chloride, petroleum ether and acetonitrile over CaH2.
Aromatic and unsaturated components of 35-65 °C petroleum
ether were removed before distillation by prolonged stirring
over concentrated H2SO4 followed by washing with 10%
aqueous solution of KOH. Infrared spectra were recorded on
a Perkin-Elmer model 1600 FT-IR spectrophotometer. Mi-
croanalyses were performed by Schwarzkopf Microanalytical
Laboratory (Woodside, NY). 1H (300 MHz), 13C{1H} (75 MHz),
and 19F (282 MHz) NMR spectra were recorded on a Varian
XL-300 spectrometer or a Varian Unity Plus 300 system in
the solvent indicated. All 19F NMR chemical shifts are
reported as ppm upfield from CFCl3, while 1H and 13C{1H}
shifts are reported as ppm downfield of internal TMS and are
referenced to the solvent peak. Coupling constants are
reported in hertz and obtained either directly from the first-
order spectra or from computer simulation of the second-order
spectra.
When the temperature of the furnace was set at 690 °C (all
three zones), conversion was 97%; at 690, 680, and 650 °C,
conversion was 96%; at 680 °C (all three zones), conversion
was 74%; at 650, 650, and 630 °C, conversion was 78%; at 610,
570, and 487 °C, conversion was about 10%.
In a similar fashion, the vapor of compound 6 (0.10 g, 0.25
mmol) was allowed to pass through the hot zone (670 ( 10
°C, 0.95 cm i.d. × 35 cm in length). The crude product was
purified by sublimation (70-90 °C, 10-2 Torr) to give a white
solid, 13 (0.03 g, 0.08 mmol, 32%). Anal. Calcd for C15H17F3-
Ru: C, 50.70; H, 4.82. Found: C, 50.81; H, 4.98. Under the
same reaction conditions used for the preparation of 13, except
for a decreased furnace temperature (630 °C), 7 (0.11 g, 0.29
mmol) was converted to a yellowish product mixture from
which 14 (0.06 g, 0.17 mmol) was isolated in 59% yield by
chromatography (1 cm i.d. × 1 cm in length silica gel column,
eluting with a mixture of 1 mL of methylene chloride and 7
mL of petroleum ether). In a second experiment, when the
furnace temperature was maintained at 605 °C, the conversion
was 85%. Anal. Calcd for C15H17F3Ru: C, 50.70; H, 4.82.
Found: C, 50.40; H, 4.91.
In similar fashion, compound 8 (0.18 g, 0.50 mmol) was
evaporated under a dynamic vacuum (10-4 Torr) at the sealed
end of a quartz tube (4 ft in length × 0.95 cm i.d.), and was
passed through the three hot zones of the furnace set at 750,
750, and 740 °C, respectively. The crude product was purified
by gradient sublimation (vacuum ) 10-3 Torr; Pyrex tubing
) 0.5 cm i.d. × 65 cm, gradient ) 60 °C to room temperature)
to give pure 15 (0.011 g, 0.03 mmol, 6%). Similarly, precursor
RuCl3‚3H2O was obtained from J ohnson-Matthey Æsar/Alfa.
Thallium(I) ethoxide was purchased from Strem. Fluorophe-
nols were purchased from Aldrich. [RuCp*Cl]4 was prepared
following the literature procedure.31
The FVT experiments were performed using a model 55347
moldatherm hinged tube furnace (three continuous hot zones,
maximum heating length ) 24 in., maximum usable o.d. tube
) 3 in.) connected with a model 58434 control console, both
purchased from Lindberg/Blue M, a unit of General Signal.
Qualitative X-ray microanalysis was performed at the Rippel
Electron Microscope Facility of Dartmouth College, on a Zeiss
DSM 962 scanning electron microscope at a 15 KV accelerating
voltage, using a link eXL analysis system and Pentafet UTW
Si(Li) X-ray detector, with a spectral resolution of 133 eV at
5.9 KeV.
Syn th eses of Th a lliu m P h en oxid es. All of the thallium
phenoxide salts were prepared by the following general
procedure. Thallium(I) ethoxide (0.4 mL, 5.6 mmol) was
syringed into a solution of the appropriate fluorophenol (5.7
mmol) in petroleum ether (45-50 mL) under nitrogen. The
resultant white precipitate was filtered in air and rinsed with
(31) Fagan, P. J .; Ward, M. D.; Calabrese, J . C. J . Am. Chem. Soc.
1989, 111, 1698.

Synthesis of C5F5-nHn Ligands
Organometallics, Vol. 17, No. 3, 1998 465
9 (0.20 g, 0.56 mmol) was subjected to FVT (710 °C, 10-4 Torr,
0.48 cm i.d. × 60 cm in length) under conditions of 88%
conversion to afford 15 (0.081 g, 0.24 mmol, 38%).
In an analogous fashion, complex 10 (0.19 g, 0.53 mmol) was
pyrolyzed at 10-4 Torr, zone temperatures of 770, 770, and
760 °C. Purification by gradient sublimation (4 × 10-4 Torr,
0.5 cm i.d. × 65 cm in length, Pyrex tubing, 30-60 °C) and
subsequent chromatography (1.5 cm in length × 0.5 cm i.d.
Al2O3 column, 50 mL of petroleum ether as eluate) afforded
16 (0.02 g, 0.06 mmol, 10%).
ment coefficients. Hydrogen atoms were treated as idealized
contributions in 15 and ignored in 17. The pentamethylcy-
clopentadienyl ring in 17 was treated as a rigid body disor-
dered in two rotational positions with a 70-30% distribution.
The fluorine atom in 17 was located at three disordered
positions with a 56-28-16% distribution.
All software and sources of the scattering factors are
contained in either the SHELXTL-92 (Gamma Test), SHELX-
TL (5.1), or SHELXTL PLUS (4.2) program libraries (G.
Sheldrick, Siemens XRD, Madison, WI).
The FVT reaction of 11 (0.25 g, 0.69 mmol) was conducted
in an analogous fashion, except that the temperature of the
furnace had to be raised to 795, 795, and 785 °C to achieve
100% conversion. Purification of the crude product by chro-
matography (10 cm in length × 1.5 cm i.d. Al2O3 column,
petroleum ether as eluate) followed by gradient sublimation
(4 × 10-4 Torr, 0.5 cm i.d. × 65 cm in length, Pyrex tubing,
60-20 °C gradient) afforded 17 (0.018 g, 0.06 mmol, 9%).
Cr ysta llogr a p h ic Stu d ies of 15 a n d 17. Details of the
crystallographic determinations are given in Table 4. Suitable
crystals were selected and mounted with epoxy cement to glass
fibers. The unit-cell parameters were obtained by the least
squares refinement of the angular settings of 24 reflections
(20° e 2θ e 25°). The systematic absence in the diffraction
data for the isomorphous 15 and 17 are uniquely consistent
for space group P21/n. The structures were solved using direct
methods, completed by subsequent difference Fourier synthe-
ses, and refined by full-matrix least-squares procedures. All
non-hydrogen atoms were refined with anisotropic displace-
Ack n ow led gm en t. R.P.H. acknowledges support of
this research by the National Science Foundation and
the Petroleum Research Fund, administered by the
American Chemical Society. A generous loan of RuCl3‚-
3H2O from J ohnson-Matthey Æsar/Alfa is also grate-
fully acknowledged. The assistance of Charles P.
Daghlian in obtaining X-ray microanalyses at the Rippel
Electron Microscope Facility at Dartmouth College is
also acknowledged.
Su p p or tin g In for m a tion Ava ila ble: Tables of the X-ray
structure determinations, atomic coordinates and anisotropic
thermal parameters, bond lengths and bond angles, anisotropic
displacement parameters, and hydrogen atom coordinates for
15 and 16 (19 pages). Ordering information is given on any
current masthead page.
OM970775P