Transition metal cyclopentadienyl complexes bearing perfluoro-4-tolyl substituents

Article abstract of DOI:10.1016/j.jorganchem.2005.12.056

Depending on the ratio of starting materials and the reaction conditions, perfluorotoluene (C₆F₅CF₃) reacts with sodium cyclopentadienide (NaCp; Cp = C₅H₅) and excess sodium hydride to afford, after a

Full text of DOI:10.1016/j.jorganchem.2005.12.056

Journal of Organometallic Chemistry 691 (2006) 1973–1983
www.elsevier.com/locate/jorganchem
Transition metal cyclopentadienyl complexes bearing
perfluoro-4-tolyl substituents
1
*
Paul A. Deck , Benjamin D. McCauley , Carla Slebodnick
Department of Chemistry, Virginia Tech, 107 Davidson Hall, Blacksburg, VA 24061, United States
Received 28 December 2005; accepted 29 December 2005
Available online 3 March 2006
Abstract
Depending on the ratio of starting materials and the reaction conditions, perfluorotoluene (C6F5CF3) reacts with sodium cyclopen-
tadienide (NaCp; Cp = C5H5) and excess sodium hydride to afford, after acidic aqueous workup, moderate to high yields of mono-, bis-,
tris-, and tetrakis(perfluoro-4-tolyl)cyclopentadiene (1, 2, 3, and 4, respectively). Treatment of 1 with excess NaH in THF afforded
sodium (perfluoro-4-tolyl)cyclopentadienide (5) in 90% yield. Reaction of FeBr2 with 2 equiv. of 5 afforded a 68% yield of (g5-
C5H4C7F7)2Fe (6). Reaction of ZrCl4(THF)2 with 2 equiv. of 5 afforded a 58% yield of (g5-C7F7C5H4)2ZrCl2 (7). Reaction of
Mn(CO)5Br with 5 afforded a 74% yield of (g5-C7F7C5H4)Mn(CO)3 (8). Treatment of 3b with NaH and then with Mn(CO)5Br in
DME afforded a 26% yield of [g5-1,2,4-(C7F7)3C5H2]Mn(CO)3 (9). Treatment of 3b with NaH and then with FeBr2 in DME afforded
a trace yield of [g5-1,2,4-(C7F7)3C5H2]2Fe (10), which was not fully characterized. Dienes 2a, 3a, and 3b and metal complexes 7, 8,
and 9 were structurally characterized by single-crystal X-ray diffraction. Infrared spectroscopic analysis of the substituted CpMn(CO)3
complexes showed a linear increase of 5 cmꢀ1 in the A-symmteric stretching frequency for each C7F7 substituent, compared to the anal-
ogous value of 4 cmꢀ1 reported earlier for each pentafluorophenyl (C6F5) substituent. Solution voltammetric analysis of the substituted
ferrocene 6 revealed a shift in the E1/2 of 465 mV relative to ferrocene, compared to the analogous value of about 340 mV for 1,10-
bis(pentafluorophenyl)ferrocene.
ꢀ 2006 Elsevier B.V. All rights reserved.
Keywords: Cyclopentadienyl complexes; Metallocenes; Ferrocene; Fluorous; Fluoroaromatic; Electronic effects; Ligand effects
1. Introduction
tution of hexafluorobenzene [1–5]. C6F5-substituted Cp
anions are stable as sodium salts and can be used in stan-
dard synthetic procedures to prepare complexes of both
early and late transition metals. C6F5-substituted ferrocenes
and cobaltocenes exhibit reversible electrochemical MII|-
MIII couples. The C6F5 substituents attached to zircono-
cene dichloride are stable under the conditions of
methylalumoxane-cocatalyzed coordination polymeriza-
tion of ethylene [6,7]. Yet, in some reactions, such as the
conversion of (g5-C6F5C5H4)2ZrCl2 to the corresponding
zirconium dimethyl [8], we have observed side reactions in
which a para fluorine of a C6F5 group undergoes a nucleo-
philic substitution. Although these problems can often be
minimized by careful control of reaction conditions,
another approach would be to replace the para fluorine with
a less nucleophilically susceptible group. We chose CF3 to
retain the strongly electron-withdrawing effect of the aryl
Ring substituents bring about interesting and useful
changes in the chemical reactivity, spectroscopic features,
and physical properties of transition metal cyclopentadienyl
(Cp) complexes. Systematic analyses of electron-withdraw-
ing substituents on metal-centered reactivity are still rela-
tively sparse, however, mainly because ‘‘standard’’
electron-withdrawing groups such as acetyl, nitro, or triflu-
oromethyl can engage in undesirable reactions of their own.
We showed previously that up to four pentafluorophenyl
(C6F5) groups are readily attached to cyclopentadiene in
one-pot procedures involving nucleophilic aromatic substi-
*
Corresponding author.
E-mail address: pdeck@vt.edu (P.A. Deck).
Undergraduate research participant.
1
0022-328X/$ - see front matter ꢀ 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.jorganchem.2005.12.056

1974
P.A. Deck et al. / Journal of Organometallic Chemistry 691 (2006) 1973–1983
substituent. We now report the preparation of perfluoro-4-
tolyl (C7F7) substituted cyclopentadienes and Cp com-
plexes. We find that the C7F7 substituent is slightly more
electron-withdrawing than C6F5.
0.27 mm3) of 2a were obtained by cooling a concentrated
hexane solution from 60 to 25 ꢁC and allowing the super-
saturated solution to stand at 25 ꢁC overnight. Colorless
plates (0.030 · 0.182 · 0.231 mm3) of 3a were crystallized
from hexanes by cooling a concentrated hexane solution
from 60 to 25 ꢁC and allowing the supersaturated solution
to stand at 25 ꢁC overnight. Colorless blocks (0.17 · 0.26 ·
0.30 mm3) of 3b were crystallized from hexane by cooling a
concentrated hexane solution from 60 to 25 ꢁC and allow-
ing the supersaturated solution to stand at 25 ꢁC overnight.
Pale yellow rods (0.06 · 0.14 · 0.42 mm3) of 7 were crystal-
lized from toluene by cooling a warm, Celite-filtered tolu-
2. Experimental
2.1. General procedures
Standard inert-atmosphere techniques were used for all
reactions. C6F5CF3 was used as received from Matrix Sci-
entific. FeBr2 and Mn(CO)5Br were used as received from
Aldrich. ZrCl4(THF)2 was prepared by the method of
Manzer [9]. NaH was purchased as a 60% mineral oil dis-
persion from Aldrich, washed repeatedly with hexanes,
dried under vacuum, and stored in a glove box. NaCp
was prepared by the method of Roesky et al. [10]. Reaction
solvents were purified by the method of Pangborn et al.
[11], except diglyme, which was distilled from CaH2 under
reduced pressure. THF-d8 (Isotec) was vacuum-transferred
from NaK2 alloy. Infrared spectra were recorded on a
Midac M-series transmission instrument operating at
1 cmꢀ1 resolution, using dilute n-tetradecane solutions
and KBr-windowed cells. NMR experiments used Varian
Unity 400 and Varian Inova 400 instruments. 19F NMR
spectra were referenced to external C6F6 in CDCl3
(ꢀ163 ppm). Elemental analyses were performed by Desert
Analytics (Tucson, Arizona) or Oneida Research Services
(Whitesboro, New York).
ene solution to 25 ꢁC. Large yellow chunks of
8
crystallized from a concentrated hexanes solution in a free-
zer at ꢀ10 ꢁC. Cutting afforded a single-crystalline block
(0.25 · 0.25 · 0.43 mm3). Yellow rods (0.16 · 0.17 ·
0.29 mm3) of 9 were crystallized from hexanes/toluene
(10:1) by dissolving a sample at 60 ꢁC and cooling to
ꢀ10 ꢁC.
The general crystallographic procedure was as follows.
Each chosen crystal was mounted on a nylon CryoLoope
(Hampton Research) with Krytoxꢂ Oil (DuPont) and cen-
tered on the goniometer of an Oxford Diffraction Xcali-
bure diffractometer equipped with a Sapphire 3e CCD
detector. The data collection routine, unit cell refinement,
and data processing were carried out with the program
CRYSALIS. The structure was solved by direct methods and
refined using the SHELXTL NT software package. Except
where noted in the following paragraph, the final refine-
ment model involved anisotropic displacement parameters
for non-hydrogen atoms and a riding model for all hydro-
gen atoms. Molecular graphics generation used SHELXTL
NT software.
2.2. Electrochemical measurements
Single-sweep cyclic (CV) and Osteryoung square-wave
(OSWV) voltammograms were obtained for 6 at nominal
concentrations of about 1 mM in CH2Cl2 using 0.10 M tet-
ra(n-butylammonium) hexafluorophosphate at the electro-
lyte, activated alumina as an internal desiccant, and
ferrocene as an internal standard. The apparatus was a
CH Instruments 620B digital potentiostat with a platinum
disk working electrode, a platinum wire auxiliary electrode,
and an [Ag|AgCl] reference electrode. The cell was purged
thoroughly with nitrogen prior to measurement. CV
sweeps were initialized at 0 mV, scanned to +1500 mV,
and reversed to 0 mV, at a scan rate of 100 mV sꢀ1. Wave
parameters (|Eox ꢀ Ered|, Ic/Ia) similar to internal ferrocene
suggested substantially reversible FeII|FeIII couples for 6.
Shifts in oxidation potential relative to internal ferrocene
were obtained using OSWV. The cell voltage was swept
from 0 to 1500 mV with a step resolution of 4 mV, a
square-wave amplitude of 25 mV, a frequency of 15 Hz,
and 256 samples per point. Voltammograms were analyzed
using the CH Instruments software package.
Noteworthy details of the structure determinations are
as follows. For the diene 2a, the asymmetric unit of the
structure comprises two crystallographically independent
molecules. For the diene 3a, the asymmetric unit of the
structure comprises two crystallographically independent
cyclopentadiene molecules and 0.5 hexane molecules. For
the diene 3b, the asymmetric unit of the structure comprises
two crystallographically independent molecules. Redisual
electron density suggested disorder in all three CF3 groups,
each modeled over two positions, with relative occupancies
of 71.3%/28.7%; 85.2%/14.8%, and 67.8%/32.2%. The
hydrogen atom positions and isotropic thermal parameters
were refined independently. For the zirconocene dichloride
complex 7, the Laue symmetry and systematic absences
were consistent with the monoclinic space groups C2/c
and Cc; the centric space group C2/c was chosen. The
asymmetric unit of the structure comprises 0.5 crystallo-
graphically independent molecules. There was substantial
rotational disorder in the sole crystallographically indepen-
dent CF3 group, which was modeled as a threefold disorder
with relative occupancies refining to 36.9%, 34.9%, and
28.2%. The C–F distances, Cꢁ ꢁ ꢁF (CCF3) distances, and
Fꢁ ꢁ ꢁF (F–C–F) distances were restrained to 1.32(1),
2.3. Crystallographic studies
Crystallographic data are presented in Table 1. Samples
were prepared as follows. Colorless blocks (0.16 · 0.25 ·
˚
2.35(1), and 2.09(1) A, respectively, for each of the three

Table 1
Crystallographic data
Compound
2a
3a
3b
7
8
9
Empirical formula
FW
C
498.22
19H4F14
C26H3F21 Æ 0.25(C6H14
735.83
)
C26H3F21
735.83
C24H8Cl2F14Zr
724.42
C15H4F7MnO3
420.12
C29H2F21MnO3
852.25
Crystal dimensions (mm)
0.27 · 0.25 · 0.16
Monoclinic
8.7178(13)
17.724(3)
21.561(3)
90
93.201(12)
90
3326.3(9)
P21/c
0.23 · 0.18 · 0.030
Monoclinic
12.5475(18)
19.713(3)
21.690(4)
90
99.103(14)
90
5297.2(14)
P21/n
0.30 · 0.26 · 0.17
Triclinic
10.4204(18)
11.3004(19)
21.345(3)
86.110(13)
83.263(13)
84.913(14)
2482.1(7)
0.42 · 0.14 · 0.060
Monoclinic
27.528(4)
6.7256(7)
14.347(2)
90
120.41(2)
90
2290.7(5)
C2/c
0.43 · 0.25 · 0.25
Monoclinic
10.3003(14)
17.129(2)
8.0472(10)
90
91.331(11)
90
1419.4(3)
P21/c
0.29 · 0.17 · 0.16
Monoclinic
11.9406(15)
14.6514(19)
15.6930(19)
90
94.756(10)
90
2736.0(6)
P21/n
Crystal system
˚
a (A)
˚
b (A)
˚
c (A)
a (ꢁ)
b (ꢁ)
c (ꢁ)
3
˚
V (A)
ꢀ
Space group
P1
Z
8
8
4
4
4
4
Dcalc (Mg mꢀ3
)
1.990
None
1952
1.845
None
2884
1.911
None
1392
2.101
0.843
1408
1.966
1.029
824
2.069
0.663
1656
Absorption coefficient (mmꢀ1
F(000)
)
Temperature (K)
h Collection range
100(2)
3.93 to 25.09
14536
5889
None
None
5889/0/595
0.0684
100(2)
4.12 to 25.06
23855
9336
None
None
9336/0/875
0.1080
100(2)
3.99 to 27.56
15316
11431
None
None
11431/0/955
0.0590
100(2)
3.35 to 30.08
8512
100(2)
4.38 to 30.08
9632
100(2)
4.09 to 30.10
18805
8027
None
None
8027/0/487
0.0469
Number of reflections collected
Number of independent reflections
Absorption correction method
Maximum and minimum transmission
Data/restraints/parameters
R [I > 2r(I)]
3363
4162
Analytical
0.952 and 0.799
3363/28/198
0.0658
Analytical
0.8740 and 0.7388
4162/0/235
0.0364
Rw [I > 2r(I)]
0.1545
0.961
0.828 and ꢀ0.420
0.2021
1.210
0.454 and ꢀ0.285
0.1489
0.908
0.598 and ꢀ0.347
0.1560
1.074
1.791 and ꢀ0.991
0.0978
1.119
0.807 and ꢀ0.364
0.1213
1.032
0.887 and ꢀ0.445
GoF on F2
ꢀ3
˚
Peak and hole (e A
)

1976
P.A. Deck et al. / Journal of Organometallic Chemistry 691 (2006) 1973–1983
orientations. The final refinement model involved aniso-
tropic displacement parameters for non-hydrogen atoms,
except the fluorine atoms of the disordered CF3 group; a
riding model was used for all hydrogen atoms. The manga-
nese complexes 8 and 9 each comprised a single indepen-
dent molecule per unit cell without disorder in the CF3
groups.
decomposition (over several hours) to a sticky, hexane-
insoluble gum. Satisfactory elemental analysis was not
1
obtained. H and 19F NMR spectra are provided in the
supporting information as evidence of substantial bulk
purity.
2.5. Synthesis of bis(perfluoro-4-tolyl)cyclopentadiene
isomers (2a, 2b, 2c, and 2d)
2.4. Synthesis of (perfluoro-4-tolyl)cyclopentadiene (1)
To a stirred mixture of NaCp (0.44 g, 5.0 mmol), NaH
(0.60 g, 25 mmol), and THF (50 mL) maintained under
nitrogen at 0 ꢁC was added perfluorotoluene (2.6 g,
11 mmol) slowly using a syringe. The mixture turned yel-
low and then brown. After stirring at 0 ꢁC for 1 h, the
cooling bath was removed and the mixture was stirred
at room temperature for 24 h. TLC analysis of the reac-
tion mixture showed a faint spot (Rf ꢂ 0.55, brown
under UV irradiation) assigned to the monoarylated
cyclopentadiene 1, a pair of intense spots (Rf ꢂ 0.38
and 0.32, respectively, brown and blue under UV irradi-
ation), faint spots (Rf ꢂ 0.20 and 0.1) assigned to the tri-
arylated cyclopentadienes 3a and 3b, and a dark brown
spot on the baseline. The THF was removed in vacuo,
and ether (50 mL) was added, followed by water
(10 mL, cautiously at first because of the vigorous H2
evolution) and aqueous sulfuric acid (10%, 5 mL). The
biphasic mixture was stirred for 15 min to ensure com-
plete hydrolysis of the reaction mixture. The aqueous
phase was separated and extracted with ether (10 mL).
The combined organic layers were washed with water
(3 · 10 mL), dried over anhydrous MgSO4, filtered, and
evaporated to afford 2.63 g of a dark, sticky residue.
TLC analysis of the residue was consistent with the
product mixture. The residue was subjected to flash chro-
matography on silica gel (6 cm · 20 cm column), eluting
with hexanes. After a forerun of 300 mL, 50-mL frac-
tions were collected. Fractions 1–5 afforded about
100 mg of 1. Fractions 7–20 afforded 1.61 g (3.2 mmol,
64%) of a colorless oil, which solidified to a pale yellow
waxy solid containing a few well-formed crystals. The
crystals were found by NMR spectroscopy to be nearly
pure 2a. One of the crystals was selected for single-crys-
tal X-ray diffraction. The waxy bulk solid was found by
1H NMR spectroscopic analysis to contain a mixture of
diarylated cyclopentadiene isomers 2a (24%), 2b (4%), 2c
(63%), and 2d (9%). 1H NMR data (CDCl3): for 2a, d
7.54 (m, 2H), 4.18 (m, 2H); for 2b, d 7.65 (m, 1H),
7.27 (m, 1H), 3.87 (br s, 2H); for 2c, d 6.90 (m, 1H),
6.74 (m, 1H), 3.70 (br s, 2H); and for 2d, d 6.93 (m,
2H), 3.54 (m, 2H). The 19F NMR spectrum of the iso-
meric mixture showed signals for the CF3 groups (rela-
tive integration 3F) centered around ꢀ56.7 ppm and
signals for the aromatic CF groups (relative integration
4 F) between ꢀ137 and ꢀ143 ppm. An analytical sample
of 2a/2b was obtained as a byproduct of the synthesis of
3a/3b. Analysis calcd. (Found) for C19H4F14 C, 45.80
(45.54); H, 0.81 (0.90).
To a mixture of NaCp (2.2 g, 25 mmol), NaH (1.2 g,
50 mmol), and THF (75 mL) maintained at about ꢀ10 ꢁC
(salt-ice bath) was added C6F5CF3 (4.72 g, 20 mmol) in
small portions over 15 min using a syringe. With the first
few drops of C6F5CF3, the mixture turned bright yellow,
and with each addition gas (H2) was evolved. After stirring
for 1 h, the mixture had turned a tan color. The cooling
bath was removed and the mixture was stirred at room
temperature for 15 h; during this interval the reaction
turned a tarry brown-black color, whether shielded from
light or not. TLC analysis (silica with fluorescent indicator,
developed with hexanes) showed, under UV irradiation, a
brown spot at Rf = 0.65 and a poorly-resolved pair of
smaller spots (one brown and one blue) at Rf = 0.40. The
solvent was then thoroughly evaporated at 25 ꢁC. Pentane
(100 mL, not hexanes) was added, and a spatula was
inserted under a nitrogen counterstream to break up the
residue. After restarting the stirrer, ice-water (50 mL) was
added (cautiously at first – hydrogen is evolved) followed
by 10 mL of ice-cold concentrated hydrochloric acid. The
biphasic mixture was capped again with a septum and stir-
red for 30 min with ice cooling to ensure complete hydroly-
sis of the numerous small black lumps. The aqueous layer
was separated and extracted with 100 mL of pentane. The
combined organic layers (orange to red) were washed with
water (2 · 50 mL), dried over MgSO4, filtered, and evapo-
rated using a rotary evaporator with equipped with aspira-
tor suction, an ice-water bath, and a solvent collection bulb
chilled with liquid nitrogen. The latter conditions ensure
efficient solvent evaporation without significant loss or
dimerization of the product. About 5.4 g (95% crude yield)
of an orange residue was obtained. The product was puri-
fied by flash chromatography on a 6 cm · 20 cm column of
silica, eluting with pentane. After the forerun (about
300 mL), a colorless band (about 250 mL of eluent) was
collected and evaporated, using the same rotary evaporator
setup as before, to afford 4.2 g (75%) of a white solid.
Yields ranged from 4.0 to 4.5 g in triplicate experiments.
1H NMR (400 MHz, CDCl3): Major isomer (1-arylcyclo-
pentadiene, ca. 80% of product) d 7.43 (m, 1H), 6.78 (m,
1H), 6.73 (m, 1H), 3.59 (m, 2H); minor isomer (2-arylcyclo-
pentadiene, ca. 20% of product) d 7.02 (m, 1H), 6.85 (m,
1H), 6.63 (m, 1H), 3.27 (m, 2H). 19F NMR (376 MHz,
CDCl3): Major isomer d ꢀ56.5 (4J = 21 Hz, 3F), ꢀ139.6
(m, 2F), ꢀ142.7 (m, 2F); minor isomer
d
ꢀ56.7
(4J = 21 Hz, 3F), ꢀ139.8 (m, 2F); ꢀ142.0 (m, 2F). Storage
of this compound at room temperature results in rapid

P.A. Deck et al. / Journal of Organometallic Chemistry 691 (2006) 1973–1983
1977
2.6. Synthesis of tris(perfluoro-4-tolyl)cyclopentadiene
isomers (3a and 3b)
the mixture turned a dark yellow-brown color. After the
initial exotherm subsided (about 15 min), TLC analysis
showed that the reaction mixture contained di- and triary-
lated cyclopentadienes (2 and 3). The mixture was lowered
into an oil bath maintained at 115 ꢁC. After 18 h, a small
aliquot was worked up and analyzed by 1H NMR spectros-
copy (the methylene region at 4.0–4.5 ppm is diagnostic
and well resolved from the diglyme signals in the region
3.3–3.7 ppm), which revealed a roughly 50:50 mixture of
3b and 4 (30:70, respectively, after 36 h; 15:87 after 60 h,
5:95 after 84 h). The reaction was allowed to proceed (typ-
ically for 3 or 4 days at 115 ꢁC) until 3b was no longer
observed. After cooling, water (10 mL) was added under
a nitrogen counterstream to quench the unreacted sodium
hydride. Then 10 mL of 10% aqueous sulfuric acid was
added to acidify the mixture. The mixture was then diluted
with ether (300 mL). The layers were separated, and the
organic layer was washed with water (15 · 25 mL) and
brine (3 · 25 mL), dried over anhydrous MgSO4, filtered,
and evaporated. The crude yield at this point was 5.0 g,
To a stirred mixture of NaCp (0.44 g, 5.0 mmol), NaH
(0.60 g, 25 mmol), and THF (50 mL) maintained under
nitrogen was added perfluorotoluene (5.0 g, 21 mmol).
The solution turned bright yellow and then brown in color,
initially with vigorous gas (H2) evolution. The mixture was
stirred at 60 ꢁC for 2 days. TLC analysis (silica, hexanes)
showed a spot corresponding to 2a/2b and two intense spots
corresponding to 3a and 3b as well as a dark spot on the
baseline of the chromatogram. (In other experiments the
reaction temperature was increased to a gentle reflux and
extended for one more day, or until TLC analysis showed
that 2a/2b/2c/2d was no longer present, however, no signif-
icant improvements in the isolated yields of 3a and 3b were
realized.) The THF was removed in vacuo, and ether
(60 mL) was added, followed by water (10 mL, cautiously
at first because of the vigorous H2 evolution) and aqueous
sulfuric acid (10%, 5 mL). The biphasic mixture was stirred
for 15 min to ensure complete hydrolysis of the reaction
mixture. The aqueous phase was separated and extracted
with ether (20 mL). The combined organic layers were
washed with water (3 · 10 mL), dried over anhydrous
MgSO4, filtered, and evaporated to afford 4.2 g of a dark,
sticky residue. The residue was subjected to flash chroma-
tography on silica, loading with 5 mL of toluene and eluting
with hexanes. A 6 cm · 25 cm column clearly resolved the
major products. After a long forerun (about 1.1 L), 50-
mL fractions were collected. Based on TLC analysis, frac-
tions 1–10 were combined and evaporated to obtain
200 mg (0.40 mmol, 8%) of 2a/2b; fractions 12–32 were
combined to afford 1.3 g (1.8 mmol, 36%) of 3b; and frac-
tions 36–75 were combined to afford 1.2 g (1.7 mmol,
34%) of 3a. Analytical samples of both 3a and 3b (including
crystals suitable for single-crystal X-ray diffraction) were
obtained by recrystallization from n-hexane. Data for 3a:
1
but H NMR spectroscopic analysis revealed the presence
of residual diglyme. A 1.0-g sample of the residue was
taken up in 10 mL of ether. The insoluble portion was trit-
urated with cold methanol (5 mL) and collected on a filter
1
to afford 140 mg of 4 (98+% pure by H NMR spectro-
scopic analysis). The soluble portion was subjected to flash
chromatography on silica gel (4 cm · 10 cm). A forerun
was eluted with ether and evaporated to afford a yellow-
orange residue (100 mg) that was mostly diglyme according
1
to H NMR spectroscopic analysis. A second band was
eluted with 10% acetic acid in ether and evaporated to
afford 380 mg of a brown solid. A third band was eluted
with 20% acetic acid in methanol and evaporated to afford
about 150 mg of a violet solid. The brown and violet resi-
dues were found by 1H NMR spectroscopy to contain
nearly pure 4, ostensibly contaminated by small amounts
of highly colored impurities. Thus, the total amount of 4
(95% purity) isolated from the ether trituration and chro-
matographic separation of the 1.0-g sample was 670 mg
(67%). The remaining 4.0 g of crude product was triturated
with 10 mL of methanol at 0 ꢁC for 1 h, collected on a fil-
ter, washed with cold methanol (2 · 5 mL) and pentane
(2 · 5 mL) and dried to afford 2.2 g (55%) of a gray solid
that was found to contain 0.5% by weight of diglyme. Sub-
sequent trituration with methanol and pentane affords ana-
lytical (rigorously diglyme-free) samples at a recovery level
of about 50%. 1H NMR (CDCl3) d 4.33 (m, 2H). 19F NMR
1
1H NMR (CDCl3) d H NMR (CDCl3) d 7.16 (s, 1H),
3.93 (s, 2H). 19F NMR (CDCl3) d ꢀ56.30 (t, 4JFF = 20 Hz,
4
3F, CF3), ꢀ56.33 (t, JFF = 22 Hz, 3F, CF3), ꢀ56.36 (t,
4JFF = 22 Hz, 3F, CF3), ꢀ138.3 (m, 6F), ꢀ138.9 (m, 4F),
ꢀ139.2 (m, 2F). Anal calcd. (Found) for 3a C, 43.72
(43.58); H, 0.42 (0.73). Data for 3b 7.53 (s, 1H), 4.26 (s,
4
2H). 19F NMR (CDCl3) d ꢀ56.26 (t, 3F, JFF = 22 Hz,
4
CF3), ꢀ56.35 (t, 3F, JFF = 21 Hz, 3F, CF3), ꢀ56.37 (t,
4JFF = 22 Hz, 3F, CF3) ꢀ137.6 (m, 2F), ꢀ137.9 (m, 2F),
ꢀ138.1 (m, 2F), ꢀ138.8 (m, 4F), ꢀ140.4 (m, 2F). Anal
calcd. (Found) for 3b C, 43.72 (44.06); H, 0.42 (0.73).
4
(CDCl3) d ꢀ56.78 (t, JFF = 22 Hz, 6F), ꢀ56.84 (t,
4JFF = 22 Hz, 6F), ꢀ138.0 (m, 4F), ꢀ138.1 (m, 4F),
ꢀ138.3 (m, 4F), ꢀ138.5 (m, 4F). Analysis calcd. (Found)
for C33H2F28 C, 42.60 (42.36); H, 0.22 (0.33).
2.7. Synthesis of tetrakis(perfluoro-4-tolyl)cyclopentadiene
(4)
Under a nitrogen atmosphere, perfluorotoluene (7.1 g,
0.030 mol) was added over 5 min to a mixture of NaCp
(0.44 g, 5.0 mmol), NaH (0.72 g, 0.030 mol), and anhy-
drous diglyme (40 mL). An exothermic reaction occurred
(Trxn ꢂ 70 ꢁC), during which gas (H2) was evolved and
2.8. Synthesis of sodium (perfluoro-4-tolyl)
cyclopentadienide (5)
A suspension of NaH (600 mg, 25 mmol) in THF
(50 mL) was prepared and cooled to 0 ꢁC using an ice bath.

1978
P.A. Deck et al. / Journal of Organometallic Chemistry 691 (2006) 1973–1983
With continued cooling and stirring, (perfluoro-4-
tolyl)cyclopentadiene (1, 4.10 g, 14.5 mmol) dissolved in
10 mL of THF was added using a cannula. Gas (H2)
evolved, and the solution turned yellow and then a drab
green color with continued stirring. After stirring 30 min,
the mixture was filtered through a medium fritte. The fil-
trate was evaporated, and the residue was heated at
60 ꢁC under vacuum for 15 h. The resulting purple solid
was triturated with pentane (2 · 50 mL), collected on a fil-
ter, and dried under vacuum to afford 4.10 g (13.5 mmol,
93%) of a purple solid. 1H NMR (THF-d8) d 6.62 (m,
2H), 6.00 (m, 2H). 19F NMR (THF-d8) d ꢀ55.3 (t,
4J = 20 Hz, 3F, CF3), ꢀ147.1 (m, 2F), ꢀ149.0 (m, 2F).
1H NMR spectroscopic analysis revealed the presence of
about 5% of residual THF.
gel, eluting with additional hexanes. The yellow filtrate
was evaporated to afford 310 mg (74%) of the cryde prod-
uct as a dark yellow solid. An analytical sample (also suit-
able for single-crystal X-ray diffraction) was obtained by
preparing a warm, concentrated hexanes solution and cool-
ing it overnight in a freezer at ꢀ10 ꢁC. 1H NMR (CDCl3) d
5.53 (m, 2H), 4.93 (m, 2H). 19F NMR (CDCl3) d ꢀ56.7 (t,
4JFF = 21 Hz, 3F, CF3), ꢀ137.3 (m, 2F), ꢀ140.7 (m, 2F).
IR (tetradecane) mCO = 2034, 1957 cmꢀ1. Analysis calcd.
(Found) for C15H4F7MnO3 C, 42.88 (42.56); H, 0.96 (0.96).
2.12. Synthesis of [g5-1,2,4-(C7F7)3C5H2]Mn(CO)3 (9)
To a suspension of NaH (72 mg, 3.0 mmol) in anhy-
drous DME (25 mL), the diene 3b (715 mg, 1.00 mmol)
was added as a solid under a nitrogen counterstream.
The mixture turned a deep yellow color and gas (H2)
evolved. After stirring several hours at room temperature,
the mixture was filtered to remove unreacted NaH. To
the filtrate was added Mn(CO)5Br (250 mg, 0.91 mmol)
as a solid under a nitrogen counterstream. The mixture
was stirred at 75 ꢁC for 15 h. After evaporating the solvent,
the dark residue was triturated for 15 min with 20 mL of
5% ethyl acetate in hexanes. The yellow supernatant was
transferred to a column of silica (4 cm · 10 cm). A yellow
band eluted with 5% ethyl acetate in hexanes. Evaporation
afforded 205 mg (0.24 mmol, 26% based on Mn(CO)5Br) of
a yellow solid. An analytical sample (also suitable for sin-
gle-crystal X-ray diffraction) of yellow crystals was
obtained by recrystallization from a warm 10:1 hex-
anes:toluene solution at ꢀ10 ꢁC. 1H NMR (CDCl3) d
6.01 (m, 2H). 19F NMR (CDCl3) d ꢀ56.7 (t, 4JFF = 21 Hz,
2.9. Synthesis of [g5-(C7F7)C5H4]2Fe (6)
A mixture of sodium (perfluoro-4-tolyl)-cyclopentadie-
nide (5, 670 mg, 2.2 mmol), FeBr2 (216 mg, 1.0 mmol)
and THF (15 mL) was stirred under reflux for 15 h. After
cooling, the solvent was evaporated, and the residue was
transferred to a column of silica gel (5 cm · 10 cm). A dark
red band eluted with dichloromethane. Evaporation fol-
lowed by recrystallization of the residue from hexane/chlo-
roform afforded 420 mg (0.68 mmol, 68%) of a brick-red
1
micro-crystalline solid. H NMR (CDCl3) d 4.93 (m, 2H),
4.54 (m, 2H). 19F NMR (CDCl3) d ꢀ56.6 (t, 4JFF = 22 Hz,
6F), ꢀ138.6 (m, 4F), ꢀ142.3 (m, 4F). Analysis calcd.
(Found) for C24H8F14Fe C, 46.63 (46.83); H, 1.30 (1.48).
2.10. Synthesis of [g5-(C7F7)C5H4]2ZrCl2 (7)
4
3F, CF3), ꢀ56.8 (t, JFF = 21 Hz, 6F, 2 CF3), ꢀ134.1 (m,
A mixture of sodium (perfluoro-4-tolyl)cyclo-pentadie-
nide (5, 640 mg, 2.1 mmol), ZrCl4(THF)2 (380 mg,
1.0 mmol), and toluene (50 mL) was stirred at 100 ꢁC for
1 h. The initial purple color gradually turned to a deep
orange. Using a heat gun, the mixture was boiled and then
quickly filtered through a bed of Celite. Fine orange nee-
dles separated as the filtrate cooled to room temperature.
The crystals were collected on a filter, washed with pentane
(20 mL), and dried in vacuo to obtain 410 mg (5.8 mmol,
58%) of orange crystals subsequently found suitable for
single-crystal X-ray diffraction. Cooling the mother liquor
4F), ꢀ137.0 (m, 2F), ꢀ138.6 (m, 4F), ꢀ139.2 (m 2F). IR
(tetradecane) mCO = 2044, 1979, 1972 cmꢀ1. Analysis calcd.
(Found) for C15H4F7MnO3 C, 40.87 (40.79); H, 0.24 (0.52).
3. Results and discussion
3.1. Cyclopentadiene arylation
As shown in Scheme 1, perfluorotoluene (C6F5CF3)
reacts with an excess of sodium cyclopentadienide (NaCp)
and excess sodium hydride (NaH) to afford, after aqueous
workup, a 75% yield of (perfluoro-4-tolyl)cyclopentadiene
(1). The use of limiting C6F5CF3 in the synthesis of 1 is
required to prevent diarylation (2), in contrast to the syn-
thesis of (pentafluorophenyl)cyclopentadiene, where NaCp
was treated with excess C6F6 at 0 to 25 ꢁC. The CF3 group
has an activating effect on the nucleophilic susceptibility of
the arene. The enhanced reactivity of perfluorotoluene rel-
ative to hexafluorobenzene has been noted elsewhere [12].
The diene 1 is formed as a mixture of tautomers (1a and
1b) which are not separated on silica gel. The monoarylated
diene 1 also undergoes significant decomposition over sev-
eral hours in solution at 25 ꢁC. Like other nucleophilic sub-
stitution reactions of C6F5CF3, these reactions are
1
in a freezer (ꢀ10 ꢁC) afforded no additional crystals. H
NMR (THF-d8) d 7.11 (m, 4H), 6.86 (m, 4H). 19F NMR
(THF-d8) d ꢀ57.6 (t, J = 21 Hz, 6F), ꢀ138.6 (m, 4F),
ꢀ143.8 (m, 4F). Analysis calcd. (Found) for C24H8Cl2-
F14Zr C, 39.79 (39.89); H, 1.11 (0.92).
2.11. Synthesis of [g5-(C7F7)C5H4]Mn(CO)3 (8)
A mixture of sodium (perfluoro-4-tolyl)-cyclopentadie-
nide (5, 310 mg, 1.02 mmol), Mn(CO)5Br, and DME
(8 mL) was stirred at 75 ꢁC for 15 h. The resulting dark res-
idue was evaporated, the residue was triturated with hex-
anes (3 · 20 mL) and filtered through a 5-cm pad of silica

P.A. Deck et al. / Journal of Organometallic Chemistry 691 (2006) 1973–1983
1979
Under more severe conditions we obtain the tetraary-
lated diene 4. Crude yields of 4 are high (about 90%) but
removal of residual reaction solvent (diglyme) from the
reaction mixture is difficult and inefficient, leading to some-
what lower yields (about 50%) of the purified diene 4. Fol-
lowing the reaction leading to diene 4 by 1H NMR
spectroscopic analysis of small, worked-up aliquots reveal
that the 1,2,3-triarylated diene 3a is depleted from the reac-
tion mixture relatively quickly (in less than 24 h) but the
isomeric 1,2,4-triarylated diene 3b is consumed more slowly
(after about 100 h) at a reaction temperature of 115 ꢁC. We
speculate that this difference in reactivity arises because 3a
has two ring carbons having only one C7F7 substituent in a
vicinal position, while both reactive ring carbons in 3b have
two C7F7 groups in vicinal positions. Although consistent
with the reactivity trends observed for the diarylated cyclo-
pentadienes 2a/2b versus 2c/2d, is not clear whether these
effects are steric or electronic in origin. Attempts to prepare
the pentaarylated homologue under forcing conditions
(e.g., diglyme, reflux) led only to intractable black tar.
Whereas 1a/1b, 2a/2b, 3a, and 3b are colorless solids, 4 is
obtained as an off-white or tan solid. Hexane solutions of all
of these dienes are essentially colorless, whereas aqueous
acetone or aqueous THF solutions go from a very pale straw
color (1a/1b) to a deep yellow color (4). We ascribe these col-
ors to autoionization of the increasingly acidic dienes in the
more polar solvents. Efforts are underway in our laborato-
ries to measure the acidity constants of these dienes on
Streitwieser’s lithium ion-pair acidity-indicator scale [13].
Scheme 1.
completely regioselective for attack at the 4-position [12],
although the formation of black, tarry substances during
the reactions suggests the possibility of various oligomeri-
zation or coupling pathways.
The diarylated product 2 was isolated as a mixture of
regioisomers, 2a–d. Isomers 2a and 2b are tautomers of
the vicinal skeleton, whereas isomers 2c and 2d are tautom-
ers of the distal skeleton. The two skeletal isomers are
obtained in approximately equal amounts, in contrast to
our observations with C6F5-substituted cyclopentadienes,
where the distal isomer predominates and the vicinal iso-
mer is only isolated from reaction mixtures in small
amounts and even then not reproducibly. The skeletal iso-
mers of 2 could not be separated efficiently by silica gel
3.2. Synthesis of organometallic complexes
Scheme 2 presents the synthesis of several organometal-
lic complexes containing cyclopentadienyl ligands bearing
perfluoro-4-tolyl substituents. As the monoarylated diene
1 was readily prepared on a large scale but rather unstable
toward decomposition, we found it convenient to prepare
and store the corresponding sodium salt (5). Simple substi-
tution reactions of 5 and transition metal halides gave the
ferrocene 6, the zirconocene dichloride 7, and the manga-
nese tricarbonyl complex 8 in moderate to high yields.
Yields of analogous reactions of the triarylated diene 3b
were somewhat lower, perhaps due to the low nucleophilic-
ity of the corresponding anion. The triarylated manganese
tricarbonyl complex 9 was obtained in 26% yield. Attempts
to synthesize the hexaarylated ferrocene 10 largely failed.
One of three reactions afforded a trace yield, and the iden-
tity of the product was confirmed only by a relatively poor-
quality X-ray crystal structure determination (R = 0.14
with intractible disorder in the CF3 groups). Transannular
steric strain may impede the formation of this hexaarylated
ferrocene.
1
chromatography. However, the H NMR signals for the
four isomers were well resolved and readily assigned on
the basis of chemical shift and integration data. As
described below, a small sample of 2a/2b was obtained
another way. Unequivocal assignment of the skeletal iso-
mers (2a/2b versus 2c/2d) was accomplished by the crystal-
lographic analysis of 2a.
Under more vigorous reaction conditions, the arylation
proceeds to the triarylated products 3a and 3b, which are
formed in roughly equal amounts. Using TLC analysis,
we found that 2c/2d was depleted from the reaction faster
than 2a/2b. We speculate that 2c/2d is more reactive as a
nucleophile because it has one ring carbon that does not
have a C7F7 group in a vicinal position. This observation
allowed us to stop the reaction after 2c/2d was depleted
but while 2a/2b was still present. Subsequent workup and
chromatographic separation on silica gel afforded a small
yield (8%) of ‘‘skeletally pure’’ 2a/2b. The separation of
the two triarylated cyclopentadienes 3a and 3b was much
easier than with their C6F5-substituted counterparts, largely
because the C7F7-substituted species are more soluble in the
preferred chromatography solvent (hexanes). Thus, a rela-
tively short silica column resolves the two isomers easily.
3.3. Analysis of substituent effects
Table 2 presents data obtained from infrared spectra of
the two manganese tricarbonyl complexes (8 and 9). of

1980
P.A. Deck et al. / Journal of Organometallic Chemistry 691 (2006) 1973–1983
Fig. 1. Thermal ellipsoid plot of 2a shown at 50% probability. One of two
nearly identical independent molecules is shown. Unlabeled fluorine atoms
are numbered the same as the carbon atoms to which they are attached.
˚
Selected bond lengths (A): C(1)–C(2), 1.363(5); C(2)–C(3), 1.442(5); C(3)–
C(4), 1.357(5); C(4)–C(5), 1.492(5); C(1)–C(5), 1.486(5). Selected bond
angles (ꢁ): C(2)–C(1)–C(5), 108.1(3); C(1)–C(2)–C(3), 109.4(3); C(4)–C(3)–
C(2), 109.9(3); C(3)–C(4)–C(5), 107.9(3); C(1)–C(5)–C(4), 104.6(3).
Selected torsional angles (ꢁ): C(2)–C(1)–C(11)–C(12), 172.8(3); C(3)–
C(4)–C(41)–C(42), ꢀ6.7(5).
identical molecules. The cyclopentadiene and aryl groups
are nearly coplanar (torsional angles are 6–8ꢁ among both
independent molecules). The molecular structure of the tri-
arylated diene 3a is shown in Fig. 2. The asymmetric unit
contains two nearly identical diene molecules and one mol-
ecule of hexane. Vicinal aryl–aryl interactions give rise to a
propellar arrangement with cyclopentadiene–aryl torsion
angles in the range 44–55ꢁ. The molecular structure of
the isomeric triarylated diene 3b is shown in Fig. 3. The
asymmetric unit contains two nearly identical diene mole-
cules. The cyclopentadiene–aryl torsion angles for the vic-
inal aryl groups are in the range 45–60ꢁ, while the
isolated aryl groups lie more in the cyclopentadiene plane
(torsion angles of 19.0ꢁ and 35.6ꢁ for the two independent
molecules). The variations in these angles suggest that,
aside from the obvious trend for vicinal interactions to
drive the torsional angles away from 0ꢁ, deformation of
these angles are relatively soft and probably influenced
Scheme 2.
Table 2
Carbonyl stretching wavenumbers (cmꢀ1) of Piano Stool complexesa
Entry
Complex
mCO (A)
mCO (E)
b,c
1
2
3
4
5
a
CpMn(CO)3
2028
2032
2040
2034
2044
1944
1954
1973, 1965
1957
1979, 1972
(g5-C6F5C5H4)Mn(CO)3
b
[g5-1,2,4-(C6F5)3C5H2]Mn(CO)3
[g5-(4-F3CC6F4)C5H4]Mn(CO)3
b
[g5-1,2,4-(4-F3CC6F4)3C5H2]Mn(CO)3
Recorded in n-alkane solution. Reproducibility is within 1 cmꢀ1
Prior results from our laboratories [2].
Lit.: 2030 and 1946 cmꢀ1 for CpMn(CO)3 [18].
.
b
c
results. Comparing entries 1–3, the, the value for mCO(A)
increases by 4 cmꢀ1 per C6F5 substituent [2]. Entries 1, 4,
and 5 show that the effect of each perfluoro-4-tolyl substi-
tuent is an increase of about 5–6 cmꢀ1 in mCO(A). We also
used solution voltammetry to evaluate the electronic effect
of the C7F7 groups in the substituted ferrocene 6. Square-
wave voltammetric measurements gave a shift in the E1/
2(ox) of 465(5) mV relative to ferrocene. The corresponding
shift for 1,10-bis(pentafluorophenyl)ferrocene is 345(5) mV.
From these results we conclude that the perfluoro-4-tolyl
group is about 1.2–1.4 times as electron-withdrawing as
the perfluorophenyl (C6F5) group when attached to cyclo-
pentadienyl ligands.
Fig. 2. Thermal ellipsoid plot of 3a shown at 50% probability. One of
two nearly identical independent molecules is shown. The solvent of
crystallization (0.5 hexane) is omitted. Fluorine atoms are numbered the
same as the carbon atoms to which they are attached. Selected bond
˚
lengths (A): C(1)–C(2), 1.362(11); C(2)–C(3), 1.429(11); C(3)–C(4),
1.399(11); C(4)–C(5), 1.466(11); C(1)–C(5), 1.465(11). Selected bond
angles (ꢁ): C(2)–C(1)–C(5), 107.6(7); C(1)–C(2)–C(3), 110.1(8); C(4)–
C(3)–C(2), 109.0(7); C(3)–C(4)–C(5), 106.5(7); C(1)–C(5)–C(4), 106.7(7).
Selected torsional angles (ꢁ): C(2)–C(1)–C(11)–C(12), 49.0(13); C(3)–
C(2)–C(21)–C(22), 51.4(11); C(4)–C(3)–C(31)–C(32); 48.0(12).
3.4. Structures of dienes
The molecular structure of the diarylated diene 2a is
shown in Fig. 1. The asymmetric unit contains two nearly

P.A. Deck et al. / Journal of Organometallic Chemistry 691 (2006) 1973–1983
1981
Fig. 3. Thermal ellipsoid plot of 3b shown at 50% probability. One of two
nearly identical independent molecules is shown. Fluorine atoms are
numbered the same as the carbon atoms to which they are attached.
˚
Selected bond lengths (A): C(1)–C(2), 1.356(4); C(2)–C(3), 1.468(4); C(3)–
C(4), 1.382(4); C(4)–C(5), 1.462(4); C(1)–C(5), 1.495(4). Selected bond
angles (ꢁ): C(2)–C(1)–C(5), 108.6(2); C(1)–C(2)–C(3), 109.3(2); C(4)–C(3)–
C(2), 108.1(2); C(3)–C(4)–C(5), 109.1(2); C(4)–C(5)–C(1); 104.9(2).
Selected torsional angles (ꢁ): C(2)–C(1)–C(11)–C(12), ꢀ120.7(3); C(1)–
C(2)–C(21)–C(22); 45.0(4); C(3)–C(4)–C(41)–C(42), 35.6(4).
significantly by crystal packing forces. Packing diagrams of
2a, 3a, and 3b do not reveal any arene stacking interactions
and no compelling evidence for CH–FC hydrogen
bonding.
Fig. 4. Thermal ellipsoid plot of 7 shown at 50% probability. Unlabeled
fluorine and hydrogen atoms are numbered the same as the carbon atoms
˚
to which they are attached. Selected bond lengths (A): Zr–Cl(1),
2.4260(12); Zr–Cp(centroid), 2.211(5). Zr–C(1), 2.549(5); Zr–C(2),
2.527(4); Zr–C(3), 2.487(5); Zr–C(4), 2.491(5); Zr–C(5), 2.537(5); C(1)–
C(2), 1.421(7); C(2)–C(3), 1.415(7); C(3)–C(4), 1.405(9); C(4)–C(5),
1.411(9); C(1)–C(5), 1.430(7). Selected bond angles (ꢁ): Cl(1)–Zr–Cl(1A),
94.56(6); Cp(centroid)–Zr–Cp(centroid), 107.1(1). Selected torsional
angles (ꢁ): Cp(least-squares plane)–aryl(least-squares plane), 7.1.
3.5. Structures of organometallic complexes
The molecular structure of the substituted zirconocene
dichloride 6 is shown in Fig. 4. The overall molecular struc-
ture (with the two aryl groups occupying positions directly
above and below the metallocene wedge) as well as key
aspects of the packing arrangement (molecules lined up
head-to-tail) are analogous to those reported for 1,10-
diphenylzirconocene dichloride [7]. The perfluoro-4-tolyl
substituent effects subtle but interesting changes in certain
bond lengths of the complex. First, the Zr–Cl bonds of 7
framework and not the result of a substituent effect. How-
ever, it is interesting that in 1,10-diphenylzirconocene
dichloride this distortion is much more pronounced: the
˚
C(1) carbon is nearly 0.16 A farther from Zr than C(3) [7].
The molecular structures of the manganese tricarbonyl
complexes 8 and 9 are shown in Figs. 5 and 6, respec-
tively. The isolated aryl substituents adopt Cp–aryl tor-
sional angles close to 0ꢁ, while the two vicinal aryl
groups adopt Cp–aryl torsional angles of about 40ꢁ. The
nearly coplanar arrangement of the Cp and isolated aryl
groups was somewhat unexpected based on our earlier
work with C6F5-substituted CpMn(CO)3 and CpRe(CO)3
complexes; isolated C6F5 substituents adopt Cp–aryl tor-
sional angles of 20–40ꢁ. Cp–aryl torsional deformations
are fairly soft, probably allowing this parameter some flex-
ibility as the molecule adapts to crystal packing forces.
˚
are somewhat contracted at 2.426(1) A, compared to
˚
2.445(2) A for Cp2ZrCl2 [14], consistent with the electron-
withdrawing character of the perfluoro-4-tolyl group. The
Cl–Zr–Cl bond angle of 7 is also contracted by about
2.5ꢁ compared to Cp2ZrCl2. Second, the Zr–Cp(centroid)
distances of both 7 and Cp2ZrCl2 are essentially identical
˚
at 2.21 A, but the individual Zr–C bond lengths of 7 are
˚
slightly (about 0.02 A) longer. This finding is consistent
with the observed lengthening of the Cp C–C bonds in 7
˚
˚
(1.41–1.43 A) compared to Cp2ZrCl2 (1.37–1.39 A). Elec-
tron-withdrawl from the pi system of the Cp ligand should
increase the C–C bond lengths. The individual Zr–C bond
lengths reveals a slight ring-slip distortion in which the C(1)
˚
The C–C bond lengths in 8 are lengthened (1.41–1.44 A)
compared to the unsubstituted CpMn(CO)3, and they also
show a slight Mills-Nixon type alternation [15], with the
C(1)–C(2) and C(1)–C(5) distances adjacent to the aryl
˚
(Cp ipso) carbon is about 0.06 A farther from the Zr atom
than the C(3) and C(4) carbons. A similar distortion of
similar magnitude is observed in Cp2ZrCl2 [14], suggesting
that the distortion is intrinsic to the bent-metallocene
˚
group the longest at 1.442(2) and 1.443(2) A, the C(3)–
C(4) bond opposite the aryl group slightly shorter at

1982
P.A. Deck et al. / Journal of Organometallic Chemistry 691 (2006) 1973–1983
stituent was the cationic cycloheptatrienyl (C7H6) moiety
[16], and an even more severe distortion was observed
when the substituent was the cationic Ph2C group [17].
In the latter case, Helmchen, Gleiter, and co-workers
rationalized the observed distortions (including a signifi-
cant displacement of the a-carbon toward the metal) by
designating the ligand as an g5-(6,6-diphenylfulvene) with
the formal positive charge shifted to Mn. In complex 8,
however, the a-carbon is actually deflected slightly away
from the Mn center; a fulvene formulation is clearly inap-
propriate in our case. There is no further lengthening of
the Cp ring C–C distances upon going from the monoary-
lated complex 8 to the triarylated complex 9, and in fact
the ring C–C bond lengths in 9 are somewhat more uni-
form. In other respects both 8 and 9 have structural
parameters very similar to CpMn(CO)3.
Fig. 5. Thermal ellipsoid plot of 8 shown at 50% probability. Fluorine and
hydrogen atoms are numbered the same as the carbon atoms to which they
˚
are attached. Selected bond lengths (A): Mn–Cp(centroid), 1.764(2), Mn–
C(6), 1.798(2); Mn–C(8), 1.800(2); Mn–C(7), 1.808(2); Mn–C(2), 2.134(2);
Mn–C(3), 2.141(2); Mn–C(1), 2.141(2); Mn–C(5), 2.141(2); Mn–C(4),
2.147(2); C(1)–C(2), 1.442(2); C(2)–C(3), 1.411(2); C(3)–C(4), 1.420(3);
C(4)–C(5), 1.415(2); C(5)–C(1), 1.443(2); C(6)–O(6), 1.151(2); C(7)–O(7),
1.139(2); C(8)–O(8), 1.141(2). Selected bond angles (ꢁ): C(6)–Mn–C(8),
92.00(8); C(6)–Mn–C(7), 92.57(7); C(8)–Mn–C(7), 92.58(8); O(6)–C(6)–
Mn, 177.61(15); O(7)–C(7)–Mn, 178.51(17); O(8)–C(8)–Mn, 177.11(17).
Selected torsional angles (ꢁ): C(2)–C(1)–C(11)–C(12), 7.6(2).
4. Conclusions
Nucleophilic arylation is an efficient approach to the
synthesis of cyclopentadienes bearing one, two, three, or
four perfluoro-4-tolyl substituents. These compounds,
when deprotonated by sodium hydride, are varied in their
utility to serve as ligands for early and late transition met-
als. In general the more highly arylated homologues give
lower yields of the desired metal complexes. Perfluoro-4-
tolyl substituents impart an electron-withdrawing effect
slightly stronger than that of pentafluorophenyl
substituents.
Acknowledgements
We thank the Virginia Tech Chemistry Department for
funding.
Appendix A. Supplementary data
Crystallographic data for the structural analysis has
been deposited with the Cambridge Crystallographic Data
Centre, CCDC No. 294056–294061 for compounds 2a, 3a,
3b, 8, 9, and 7, respectively. Copies of this information may
be obtained free of charge by writing The Director, CCDC,
12 Union Road, Cambridge, CB2 1EZ, UK. Alternatively
one may contact CCDC by telefax on +44 1223 336 033, by
e-mail at deposit@ccdc.cam.ac.uk, or on the internet at
www: http://www.ccdc.cam.ac.uk. Supplementary data
associated with this article can be found, in the online ver-
sion, at doi:10.1016/j.jorganchem.2005.12.056.
Fig. 6. Thermal ellipsoid plot of 9 shown at 50% probability. Fluorine and
hydrogen atoms are numbered the same as the carbon atoms to which they
˚
are attached. Selected bond lengths (A): Mn–Cp(centroid) 1.782(2); Mn–
C(1), 2.157(2); Mn–C(2) 2.173(2); Mn–C(3), 2.146(2); Mn–C(4), 2.157(2);
Mn–C(5), 2.138(2); Mn–C(6), 1.802(2); Mn–C(7), 1.800(3); Mn–C(8),
1.807(2); C(1)–C(5), 1.421(3); C(1)–C(2), 1.432(3); C(2)–C(3), 1.417(3);
C(3)–C(4), 1.428(3); C(4)–C(5), 1.421(3); C(6)–O(6), 1.136(3); C(7)–O(7),
1.144(3); C(8)–O(8), 1.143(3). Selected bond angles (ꢁ): C(7)–Mn–C(6),
90.75(11); C(7)–Mn–C(8), 91.36(11); C(6)–Mn–C(8), 89.03(11); O(6)–
C(6)–Mn, 179.0(2); O(7)–C(7)–Mn, 177.0(2); O(8)–C(8)–Mn, 178.3(2).
Selected torsional angles (ꢁ): C(2)–C(1)–C(11)–C(16), 43.0(3); C(1)–C(2)–
C(21)–C(22), 40.9(3); C(5)–C(4)–C(41)–C(46), 3.0(3).
References
˚
1.420(3) A, and the C(2)–C(3) and C(4)–C(5) bonds the
˚
shortest at 1.411(2) and 1.415(2) A, respectively. A search
[1] P.A. Deck, W.F. Jackson, F.R. Fronczek, Organometallics 15 (1996)
5287.
[2] M.P. Thornberry, C. Slebodnick, P.A. Deck, F.R. Fronczek,
Organometallics 19 (2000) 5352.
[3] M.P. Thornberry, C. Slebodnick, P.A. Deck, F.R. Fronczek,
Organometallics 20 (2001) 920.
[4] P.A. Deck, F.R. Fronczek, Organometallics 19 (2000) 327.
of the Cambridge Structural Database (CSD) for mono-
metallic, monosubstituted CpMn(CO)3 complexes shows
that such a pronounced structural effect occurs only when
the substituent is clearly electron-withdrawing. For exam-
ple, a comparable distortion was observed when the sub-

P.A. Deck et al. / Journal of Organometallic Chemistry 691 (2006) 1973–1983
1983
[5] S. Becke, U. Denninger, S. Kahlert, W. Obrecht, C. Schmid, H.
Windisch, European patent application, 2002, (PCT Appl. WO 02/
12154 (2002), to Bayer, AG).
[6] R.J. Maldanis, J.C.W. Chien, M.D. Rausch, J. Organomet. Chem.
599 (2000) 107.
[12] J.H. James, M.E. Peach, C.R. Williams, J. Fluorine Chem. 27 (1985)
91.
[13] S. Gronert, A. Streitwieser, J. Am. Chem. Soc. 108 (1986) 7016.
[14] T. Repo, M. Klinga, I. Mutikainen, Y.C. Su, M. Leskela, M.
Polamo, Acta Chem. Scand. 50 (1996) 1116.
[7] M.P. Thornberry, N.T. Reynolds, P.A. Deck, F.R. Fronczek, A.L.
Rheingold, L.M. Liable-Sands, Organometallics 23 (2004) 1333.
[8] E.J. Hawrelak, P.A. Deck, Organometallics 22 (2003) 3558.
[9] L.E. Manzer, Inorg. Synth. 21 (1987) 135.
[15] S. Shaik, A. Shurki, D. Danovich, P.C. Hiberty, Chem. Rev. 101
(2001) 1501.
[16] M. Tamm, A. Grzegorzewski, I. Brudgam, J. Organomet. Chem. 519
¨
(1996) 217.
[10] T.K. Panda, M.T. Garner, P.W. Roesky, Organometallics 22 (2003)
877.
[17] M.A.O. Volland, G. Helmchen, I. Hyla-Kryspin, F. Rominger, R.
Gleiter, Organometallics 20 (2001) 227.
[11] A.B. Pangborn, M.A. Giardello, R.H. Grubbs, R.K. Rosen, F.J.
Timmers, Organometallics 15 (1996) 1518.
[18] I.B. Benson, S.A.R. Knox, R.F.D. Stansfield, P. Woodward, J.
Chem. Soc., Dalton Trans. (1981) 51.