Synthesis and structural trends in pentafluorophenyl-substituted ferrocenes, 1,4-tetrafluorophenylene-linked diferrocenes, and 1,1′-ferrocenylene-1,4-tetrafluorophenylene co-oligomers

Article abstract of DOI:10.1021/om990930v

1,1′-Dilithioferrocene, or (CpLi)₂Fe (1), reacts with excess C₆F₆ in THF to afford mostly (C₆F₅Cp)₂Fe (2), accompanied by the diiron complex (C₆F₅Cp)FeCp(1,4-C₆F₄)CpFe(CpC ₆F₅) (3). The reaction of 1 with 1.0 equiv of C₆F₆ affords a mixture of sparingly soluble stepgrowth oligomers assigned to [-FeCp(1,4-C₆F₄)-]_n (4). [1-Li-2-R^NCp]FeCp (6, R^N = CH₂-NMe₂) reacts with 1 equiv of C₆F₆ to afford [1-(C₆F₅)-2-R^NCp]FeCp (7). Complex 6 reacts with 0.5 equiv of C₆F₆ to afford CpFe[2-R^NCp](1,4-C₆F₄)[2-R^NCp]FeCp as a mixture of two diastereomers (meso-8, major and dl-8, minor). The relative stereochemistry of meso-8 is established by crystallographic analysis of the corresponding bis(methiodide) (12). [1-Li-2-R^NCp]Fe(CpLi) (9) reacts with excess C₆F₆ to afford mainly the diarylated ferrocene [1-(C₆F₅)-2-R^NCp]Fe(CpC₆F ₅) (10). The reaction of 9 instead with 1.0 equiv of C₆F₆ affords a mixture of oligomeric species assigned to the formula [-(2-R^NCp)FeCp(1,4-C₆F₄)-]_n (11). In contrast to 4, 11 is sufficiently soluble for characterization by solution NMR spectroscopy (end group analysis), and under optimized conditions a number-averaged degree of oligomerization corresponding to a linear Fe₉ species is determined. The complexes 3, 7, 10, and 12 are characterized by single-crystal X-ray diffraction. Crystalline 3 exhibits infinite π-stacking interactions composed of intramolecular triple-decker C₆F₅-C₆F₄-C₆F₅ stacking and intermolecular Cp-C₆F₅ stacking. The crystal structure of 10 also shows intramolecular arene stacking, but intermolecular Cp-C₆F₅ interaction is interrupted by the dimethylaminomethyl substituent. None of these complexes show intermolecular stacking among the various C₆F₅ and C₆F₄ groups. Extensive intermolecular C-H?F-C interactions in the crystallographic packing diagrams of 3, 7, 10, 12, and CpFeCpC₆F₄C₆F₄CpFeCp (13) are characterized as either purposeful or diffuse by examining correlations of their respective H?F distances and C-H?F angles. Solution voltammetry of 3 shows two unresolved, reversible Fe^II|Fe^III couples, suggesting weak electronic communication between the two iron centers despite the fully conjugated structure.

Full text of DOI:10.1021/om990930v

Organometallics 2000, 19, 1013-1024
1013
Syn th esis a n d Str u ctu r a l Tr en d s in
P en ta flu or op h en yl-Su bstitu ted F er r ocen es,
1,4-Tetr a flu or op h en ylen e-Lin k ed Difer r ocen es, a n d
1,1′-F er r ocen ylen e-1,4-Tetr a flu or op h en ylen e
Co-oligom er s
Paul A. Deck,* Michael J . Lane,1 J ennifer L. Montgomery,1 and Carla Slebodnick
Department of Chemistry, Virginia Tech, Blacksburg, Virginia 24061-0212
Frank R. Fronczek
Department of Chemistry, Louisiana State University, Baton Rouge, Louisiana 70803-1804
Received November 26, 1999
1,1′-Dilithioferrocene, or (CpLi)2Fe (1), reacts with excess C6F6 in THF to afford mostly
(C6F5Cp)2Fe (2), accompanied by the diiron complex (C6F5Cp)FeCp(1,4-C6F4)CpFe(CpC6F5)
(3). The reaction of 1 with 1.0 equiv of C6F6 affords a mixture of sparingly soluble step-
growth oligomers assigned to [-FeCp(1,4-C6F4)-]n (4). [1-Li-2-RNCp]FeCp (6, RN ) CH2-
NMe2) reacts with 1 equiv of C6F6 to afford [1-(C6F5)-2-RNCp]FeCp (7). Complex 6 reacts
with 0.5 equiv of C6F6 to afford CpFe[2-RNCp](1,4-C6F4)[2-RNCp]FeCp as a mixture of two
diastereomers (m eso-8, major and d l-8, minor). The relative stereochemistry of m eso-8 is
established by crystallographic analysis of the corresponding bis(methiodide) (12). [1-Li-2-
RNCp]Fe(CpLi) (9) reacts with excess C6F6 to afford mainly the diarylated ferrocene [1-(C6F5)-
2-RNCp]Fe(CpC6F5) (10). The reaction of 9 instead with 1.0 equiv of C6F6 affords a mixture
of oligomeric species assigned to the formula [-(2-RNCp)FeCp(1,4-C6F4)-]n (11). In contrast
to 4, 11 is sufficiently soluble for characterization by solution NMR spectroscopy (end group
analysis), and under optimized conditions a number-averaged degree of oligomerization
corresponding to a linear Fe9 species is determined. The complexes 3, 7, 10, and 12 are
characterized by single-crystal X-ray diffraction. Crystalline 3 exhibits infinite π-stacking
interactions composed of intramolecular “triple-decker” C6F5-C6F4-C6F5 stacking and
intermolecular Cp-C6F5 stacking. The crystal structure of 10 also shows intramolecular
arene stacking, but intermolecular Cp-C6F5 interaction is interrupted by the dimethylami-
nomethyl substituent. None of these complexes show intermolecular stacking among the
various C6F5 and C6F4 groups. Extensive intermolecular C-H‚‚‚F-C interactions in the
crystallographic packing diagrams of 3, 7, 10, 12, and CpFeCpC6F4C6F4CpFeCp (13) are
characterized as either “purposeful” or “diffuse” by examining correlations of their respective
H‚‚‚F distances and C-H‚‚‚F angles. Solution voltammetry of 3 shows two unresolved,
reversible FeII|FeIII couples, suggesting weak electronic communication between the two iron
centers despite the fully conjugated structure.
In tr od u ction
molecular level is usually sought with electrochemical
or spectroscopic measurements.2,3a,4 Ring-opening po-
lymerization of heteroferrocenophanes is currently the
most general synthetic approach to high molecular
weight polymers.2a-d Other approaches lead to more
fully conjugated species with lower molecular weights.4,5
Ferrocene-containing oligomers and polymers in which
the complexed iron atoms are integral to the principal
chain are under extensive investigation in several
laboratories.2 The use of the FeII|FeIII couple as a charge
carrier in conducting organic filaments or films is often
mentioned as a potential application of these materials.3
Evidence for interactions among ferrocene groups at the
(3) (a) Pittman, C. U.; Carraher, C. E.; Zeldin, M.; Sheats, J . E.;
Culbertson, B. M. Metal-Containing Polymeric Materials; Plenum:
New York, 1996. (b) Buretea, M. A.; Tilley, T. D. Organometallics 1997,
16, 1507. (c) Yamamoto, T.; Morikita, T.; Maruyama, T.; Kubota, K.;
Katada, M. Macromolecules 1997, 30, 5390.
(4) (a) Barlow, S.; O’Hare, D. Chem. Rev. 1997, 97, 637. (b) Hradsky,
A.; Bildstein, B.; Schuler, N.; Schottenberger, H.; J aitner, P.; Ongania,
K.-H.; Wurst, K.; Launay, J .-P. Organometallics 1997, 16, 392-402.
(c) Hirao, T.; Kurashina, M.; Aramaki, K.; Nishihara, H. J . Chem. Soc.,
Dalton Trans. 1996, 2929-2933. (d) Ribou, A.-C.; Launay, J .-P.;
Sachtleben, M. L.; Li, H.; Spangler, C. W. Inorg. Chem. 1996, 35, 3735-
3740. (e) Stanton, C. E.; Lee, T. R.; Grubbs, R. H.; Lewis, N. S.;
Pudelski, J . K.; Callstrom, M. R.; Erickson, M. S.; McLaughlin, M. L.
Macromolecules 1995, 28, 8713-8721. (f) Hudson, R. D. A.; Foxman,
B. M.; Rosenblum, M. Organometallics 1999, 18, 4098-4106.
(1) Undergraduate research participants.
(2) (a) Nguyen, P.; Gomez-Elipe, P.; Manners, I. Chem. Rev. 1999,
99, 1515. (b) Kingsborough, R. P.; Swager, T. M. Prog. Inorg. Chem.
1999, 48, 123. (c) Manners, I. Angew. Chem., Int. Ed. Engl. 1996, 35,
1602-1621. (d) Manners, I. Adv. Organomet. Chem. 1995, 37, 131-
168. (e) Kurosawa, M.; Nankawa, T.; Nishihara, H. Inorg. Chem. 1999,
38, 5113. (f) Plenio, H.; Hermann, J .; Leukel, J . Eur. J . Inorg. Chem.
1998, 12, 2063. (g) Heo, R. W.; Somoza, F. B.; Randall, T. J . Am. Chem.
Soc. 1998, 120, 1621. (h) Southard, G. E.; Curtis, M. D. Organometallics
1997, 16, 5618.
10.1021/om990930v CCC: $19.00 © 2000 American Chemical Society
Publication on Web 02/25/2000

1014 Organometallics, Vol. 19, No. 6, 2000
Deck et al.
Sch em e 1
gested the presence of linear, alternating 1,1′-ferroce-
nylene-1,4-tetrafluorophenylene co-oligomers (4). This
paper presents our attempts to optimize the formation
of these oligomeric, or polymeric, species (4). Ultimately,
the low solubility of 4 as well as insurmountable
difficulties in the purification and characterization of
the lithioferrocene intermediates frustrated our efforts
to obtain high polymers by this step-growth process.13
However, in the course of these studies, we isolated
several arylated ferrocenes and diferrocenylarene com-
plexes, and we found that these complexes show several
interesting and pervasive solid-state structural motifs,
including intramolecular (transannular) arene stacking,
intermolecular arene-Cp stacking, and networks of
C-H‚‚‚F-C interactions. This contribution presents the
syntheses of these new complexes and an analysis of
the trends in their structural data.
“Crystal engineering” is a rapidly growing area of
research in which diverse combinations of intermolecu-
lar forces are orchestrated to enable rational syntheses
of ordered solids,6 including organometallic compounds.7
Whereas hydrogen bonding remains the most important
molecular recognition design factor, several groups have
recently turned their attention toward weaker interac-
tions, including those characteristic of organic fluor-
ides.8-10 Detailed understanding of these weaker forces
stands to broaden the scope of crystal engineering
concepts to compounds lacking OH, NH, and other
hydrogen-bonding functional groups.
We showed earlier that reactions of NaCp and hexaflu-
orobenzene afford pentafluorophenyl-substituted cyclo-
pentadienes, which we used in a general synthesis of
C6F5-substituted metallocene complexes.11 It seemed
reasonable that 1,1-dilithioferrocene (1)12 would react
with fluoroaromatic compounds such as hexafluoroben-
zene (Scheme 1) to give fluoroarylated ferrocenes more
directly. We sought to prepare a variety of fluoroaryl-
ated ferrocene complexes by this route as part of our
ongoing investigation of the effects of fluoroaryl sub-
stituents on the structure, physical and spectroscopic
properties, and reactivity of metallocenes and other Cp
complexes.
Resu lts a n d Discu ssion
Syn th esis of Mon om eta llic Com p lexes. Dilithio-
ferrocene reacted with an excess of C6F6 to produce the
diarylated ferrocene (2) as the major product. We
prepared complex 2 previously by the reaction of FeBr2
with Na(C6F5Cp).11 Among the byproducts of this reac-
tion were species that had NMR spectra tentatively
assigned to structural motifs (A and B) in which either
n-butyllithium or 1,2-bis(dimethylamino)ethyllithium
displaced aromatic fluorides of C6F6 or of the product
2. Momentary isolation and washing of the dilithiofer-
rocene intermediate 1 minimized the formation of these
byproducts in subsequent reactions with C6F6. We found
that 1 is not sufficiently stable for long-term storage
1
(even in our glovebox freezer at -35 °C). Also, the H
NMR spectra of 1 showed varying amounts of residual
N,N,N′,N′-tetramethylethylenediamine (TMEDA) as well
as broad spectral features attributed to aggregation
processes. The presence of monolithioferrocene could be
neither confirmed nor ruled out by spectroscopic means.
In our initial experiment, we used only a small excess
of C6F6, and in addition to the desired diarylated
ferrocene (2), a complex mixture of dark orange-red,
poorly soluble byproducts was obtained (Scheme 1).
Preliminary analyses of the byproduct mixture sug-
We next attempted to generalize the route shown in
Scheme 1 to substituted ferrocenes. Treatement of 1,1′-
dimethylferrocene or 1,1′-bis(trimethylsilyl)ferrocene
using n-BuLi/TMEDA in hexanes followed by electro-
philic quenching (Me3SiCl) and subsequent 1H NMR
analysis of the product mixture showed that lithiation
was incomplete even after long reaction times with large
excesses of n-BuLi. However, the commercially available
tertiary amine 5 was previously known to undergo
highly efficient and regioselective lithiation with 1 or 2
equiv of n-BuLi/TMEDA (Scheme 2) to afford the
monolithiated (6) or dilithiated (9) intermediates, re-
spectively.14 Subsequent reactions of 6 or 9 with excess
hexafluorobenzene afforded the corresponding arylated
ferrocenes (7 and 10, respectively) in moderate yields.
(5) Hollandsworth, C. B.; Hollis, W. G., J r.; Slebodnick, C.; Deck,
P. A. Organometallics 1999, 18, 3610.
(6) (a) Nangia, A.; Desiraju G. R. Top. Curr. Chem. 1998, 198, 57.
(b) Aakeroy, C. B.; Borovik, A. S. Coord. Chem. Rev. 1999, 183, 1.
(7) (a) Braga, D.; Grepioni, F. Coord. Chem. Rev. 1999, 183, 19. (b)
Rivas, J . C. M.; Brammer, L. Coord. Chem. Rev. 1999, 183, 43.
(8) (a) Renak, M. L.; Bartholomew, G. P.; Wang, S. J .; Ricatto, P.
J .; Lachicotte, R. J .; Bazan, G. C. J . Am. Chem. Soc. 1999, 121, 7787.
(b) Row: T. N. G. Coord. Chem. Rev. 1999, 183, 81. (c) Evans, T. A.;
Seddon, K. R. Chem. Commun. 1997, 2024. (d) Williams, J . H.;
Cockcroft, J . K.; Fitch, A. N. Angew. Chem., Int. Ed. Engl. 1992, 31,
1655.
(9) Weck, M.; Dunn, A. R.; Matsumoto, K.; Coates, G. W.; Lobkovsky,
E. B.; Grubbs, R. H. Angew. Chem., Int. Ed. 1999, 38, 2741.
(10) Thalladi, V. R.; Weiss, H.-C.; Bla¨ser, D.; Boese, R.; Nangia, A.;
Desiraju, G. R. J . Am. Chem. Soc. 1998, 120, 8702.
(11) Deck, P. A.; J ackson, W. F.; Fronczek, F. R. Organometallics
1996, 15, 5287-5291.
(12) (a) Guillaneux, D.; Kagan, H. B. J . Org. Chem. 1995, 60, 2502.
(b) Riant, O.; Argouarch, G.; Guillaneux, D.; Samuel, O.; Kagan, H. B.
J . Org. Chem. 1998, 63, 3511. (c) Dietz, S. D.; Bell, W. L.; Cook, R. L.
J . Organomet. Chem. 1997, 546, 67. (d) Rausch, M. D.; Moser, G. A.;
Meade, C. F. J . Organomet. Chem. 1973, 51, 1. (e) Slocum, D. W.;
Englemann, T. R.; Ernst, C.; J ennings, C. A.; J ones, W.; Koonsvitsky,
B.; Lewis, J .; Shenkin, P. J . Chem. Educ. 1969, 46, 145.
(13) In step-growth polymerization, a fractional monomer impurity
(p) leads to a limit in the number-averaged degree of oligomerization
according to the Carothers equation, D ) (1 - p)-1. See: Flory, P. J .
Principles of Polymer Chemistry; Cornell University Press: Ithaca,
1953; p 81.

Pentafluorophenyl-Substituted Ferrocenes
Organometallics, Vol. 19, No. 6, 2000 1015
19F NMR spectrum of this mixture showed that the ratio
of m eso-8 to d l-8 was about 3:2, determined by compar-
ing the relative intensities of the two singlets at -138.80
ppm (m eso-8) and -138.65 ppm (d l-8) in CDCl3.
Miraculously, m eso-8 and d l-8 were separated simply
by repeated extraction of the mixture with hexanes at
25 °C until the filtrate ran colorless to the eye. The filter
cake was 95% pure m eso-8, and the complex isolated
from the filtrate was 97% pure d l-8, as determined by
19F NMR analysis. The relative configuration of m eso-8
was confirmed by obtaining a crystal structure of the
corresponding bis(methiodide) derivative (see below).
These compounds are reminiscent of the “planar-chiral”
ferrocenes that have attracted attention as ligands for
stereoselective catalysts.15 Work is underway to resolve
d l-8 by fractional crystallization of the bis(hydrocam-
phorsulfonate).
Sch em e 2
Isola t ion a n d Ch a r a ct er iza t ion of Oligom er ic
Sp ecies. Our initial plan of isolating alternating fer-
rocenylene-tetrafluorophenylene copolymers met with
varied results. We conjectured that the diiron complex
3 was produced by a step-growth process (Scheme 1)
and should be accompanied by linear, oligomeric homo-
logues (4). Unfortunately, the product mixture 4 was
not soluble enough (e.g., in ether, dichloromethane,
chloroform, benzene, or ethyl acetate) for chromato-
graphic separations or fractional crystallization. Despite
the presence of perfluoroaromatic groups, the solubility
of 4 was not noticeably improved in “fluorous” solvents16
such as hexafluorobenzene or 1-(perfluorobutyl)tetrahy-
drofuran. We used Soxhlet extraction of crude 4 with a
series of solvents having increasing boiling points
(benzene, toluene, chlorobenzene, and 1,2-dichloroben-
zene) to separate 4 into several “extracts”. Analysis of
the initial benzene “extract” showed the same general
spectroscopic features as 3 (internal 1,4-C6F4 and end
group C6F5, both considerably broadened), but with
much lower relative integration of the end groups. The
1H NMR spectrum also showed a low population of
unsubstituted Cp end groups. We estimated the number-
averaged degree of oligomerization to be in the range
10 < Dn < 20. End group analysis alone cannot rule
out macrocyclic species. Moreover, the “extract” was only
partially dissolved in the NMR solvent (CDCl3), so the
solid substance was not sampled representatively. Tri-
iron, tetrairon, and pentairon homologues of 3 were
detected as a progression of signals separated by about
330 amu in their mass spectra (EI, positive ion). The
repeat unit of the oligomer has a mass of 332 amu. In
contrast to our recent report of octafluorobiphenylene-
linked oligoferrocenes,5 the triiron and tetrairon oligo-
mers of 4 could not be separated chromatographically
on either silica gel or alumina. Surprisingly, direct probe
CI (positive and negative ion), FAB, and MALDI-TOF
mass spectrometric techniques all failed to give useful
spectra.17 The “extracts” obtained using toluene, chlo-
Conformational rotation about the C6F5-Cp axis in
these compounds is relatively free. For example, using
19F NMR analysis, the diastereotopic ortho C-F signals
in 7 showed a coalescence temperature of about -60(5)
°C in CD2Cl2, and the corresponding ortho C-F signals
of 10 (for C6F5 attached to RNCp) and of 2 were still in
pseudo-fast exchange at -75 °C.
Isola t ion a n d Ch a r a ct er iza t ion of Dim et a llic
Com p lexes. As shown in Scheme 1, the reaction of
(LiCp)2Fe and C6F6 (1:1 ratio) afforded a mixture of
oligomeric species. The shortest “oligomer”, the diiron
complex 3, was isolated from the product mixture and
characterized. The 19F NMR spectrum of 3 was instruc-
tive. A sharp singlet (4 F) for the bridging 1,4-C6F4
group confirmed the regioselectivity of aromatic substi-
tution in the “repeat unit”, whereas three additional
signals (total of 10 F) indicated two chemically equiva-
lent C6F5 “end groups”. The crystal structure and the
solution voltammetric analysis of 3 are discussed below.
The reaction of the monolithiated intermediate 6 with
C6F6 (2:1 ratio) afforded the diiron complex 8 as a
mixture of meso and dl diastereomers (Scheme 2). The
(15) (a) Houdus, B. L.; Ruble, J . C.; Fu, G. C. J . Am. Chem. Soc.
1999, 121, 2637-2638. (b) Widhalm, M.; Mereiter, K.; Bourghida, M.
Tetrahedron: Asymmetry 1998, 9, 2983. (c) Riant, O.; Samuel, O.;
Flessner, T.; Taudien, S.; Kagan, H. B. J . Org. Chem. 1997, 62, 6733.
(16) (a) Fish, R. H.; Chem. Eur. J . 1999, 5, 1677. (b) J uliette, J . J .
J .; Rutherford, D.; Horvath, I. T.; Gladysz, J . A. J . Am. Chem. Soc.
1999, 121, 2696. (c) Smith, D. C.; Stevens, E. D.; Nolan, S. P. Inorg.
Chem. 1999, 38, 5277. (d) Barthel-Rosa, L. P.; Gladysz, J . A. Coord.
Chem. Rev. 1999, 192, 587. (e) de Wolf, E.; van Koten, G.; Deelman,
B. J . Chem. Soc. Rev. 1999, 28, 37. (f) Horvath, I. T. Acc. Chem. Res.
1998, 31, 641.
(14) (a) Xiao, L.; Mereiter, K.; Weissensteiner, W.; Widhalm, M.
Synthesis 1999, 8, 1354. (b) Hitchcock, P. B.; Hughes, D. L.; Leigh, G.
J .; Sanders, J . R.; De Souza, J . S. J . Chem. Soc., Dalton Trans. 1999,
7, 1161. (c) Nishibayashi, Y.; Yasuyoshi, A.; Ohe, K.; Uemura, S. J .
Org. Chem. 1996, 61, 1172. (d) Slocum, D. W.; J ennings, C. A.;
Engelmann, T. R.; Rockett, B. W.; Hauser, C. R. J . Org. Chem. 1971,
36, 377. (e) Marr, G. J . Organomet. Chem. 1967, 9, 147.

1016 Organometallics, Vol. 19, No. 6, 2000
Deck et al.
Ta ble 1. Cr ysta llogr a p h ic Da ta
7
3
10
12‚3CHCl3
empirical formula
fw
C38H16F14Fe2
850.21
C19H16F5FeN
409.18
C25H15F10FeN
575.23
C37H41Cl9F4Fe2I2N2
1274.27
diffractometer
cryst dimens (mm)
cryst syst
a (Å)
b (Å)
c (Å)
Siemens P4
0.1 × 0.2 × 0.08
monoclinic
11.330(3)
6.565(2)
20.192(5)
94.04(3)
1498.2(8)
P2(1)/c
Siemens P4
0.28 × 0.34 × 0.41
monoclinic
17.0820(11)
6.7453(5)
15.8893(10)
111.234(5)
1706.5(2)
P2(1)/c
Siemens P4
0.5 × 0.5 × 0.1
monoclinic
7.1150(10)
32.954(3)
10.1340(10)
104.640(10)
2299.0(4)
P2(1)/c
Enraf-Nonius CAD4
0.50 × 0.40 × 0.37
monoclinic
12.897(1)
27.419(3)
14.091(3)
110.29(1)
4673(1)
C2/m
4
1.760
2.490
â (deg)
V (Å)3
space group
Z
2
4
4
D
calc (Mg m-3
)
1.885
1.088
844
1.593
0.937
832
1.662
0.752
1152
abs coeff (mm-1
F000
)
2428
λ(Mo KR) (Å)
temp (K)
0.71073
298(2)
0.71073
298(2)
0.71073
293(2)
0.71073
100(2)
θ collection range
no. of reflns colld
no. of indep reflns
abs corr method
max, min transm
no. of data/restrs/params
R [I > 2σ(I)]
1.80-25.01
2.56-25.00
2.17-25.00
2.75-29.98
3641
3937
5203
7188
2640
2994
4010
ψ scans
0.8929, 0.7085
3639/0/335
0.0511
0.1306
1.071
0.37, -0.414
6938
ψ scans
0.398, 0.328
6938/0/260
0.0387
0.1023
1.051
3.043, -1.722
integration
0.8414, 0.8030
2399/0/245
0.0506
0.1273
1.015
integration
0.7828, 0.6767
2994/0/299
0.0316
0.0686
0.876
Rw [I > 2σ(I)]
GoF on F2
diff peak, hole (e Å-3
)
0.782, -0.685
0.201, -0.287
robenzene, and dichlorobenzene yielded only intractable
red to red-brown solids.
The reaction of the substituted dilithiated ferrocene
derivative 9 with 1 equiv of C6F6 (Scheme 2) afforded
an orange-red solid (11) that was freely soluble in
benzene, dichloromethane, chloroform, and ether, but
only partly soluble in hexanes and methanol. Crude 11
was purified by chromatography on alumina mainly to
remove small amounts of mononuclear complexes such
as 7 and 10. The isolation of the monoarylated complex
7 suggesed that the dilithiated intermediate 9 either
contained some 6 or had abstracted protons from the
solvent, neither of which can be tolerated in the
1
synthesis of a high polymer. H and 19F NMR spectra
of the “purified” oligomer 11 were consistent with the
assigned structure. Using mononuclear complexes 7 and
10 as models for chemical shift assignments, the number-
average degree of polymerization (n) of 11 was esti-
mated to be approximately 9 ( 1, corresponding to a
molecular weight (Mn) of 3500 ( 500. Details of this “end
group” analysis (spectral assignments and algebraic
workup of integration data) are provided in the Sup-
porting Information. Published step-growth routes to
ferrocene-containing polymers gave similar degrees of
oligomerization.4
Cr ysta llogr a p h ic An a lyses. Collection and refine-
ment data for the X-ray diffraction analysis of 3, 7, 10,
and 12 are compiled in Table 1. Complete tables of
crystallographic data are deposited in the Supporting
Information.
F igu r e 1. Thermal ellipsoid plot of 7 shown at 50%
probability.
electrons on nitrogen is directed away from the C6F5
substituent (compare 10, below). The Cp-C6F5 torsion
angle is 48.4°.
The ferrocene moiety of 10 (Figure 2) is relatively
undistorted, with equal Fe-Cp(centroid) distances (1.653
and 1.645 Å), parallel Cp ligands (2.9° interplanar
angle), and Fe-C distances within 0.027 Å of their
mean. The two ipso carbons, C(14) and C(20), of the C6F5
groups appear slightly bent out of the C5 least-squares
planes of the Cp ligands, away from the iron atom, by
0.027 and 0.032 Å, respectively. Fluorine atom positions
deviated no more than 0.051 Å from their respective C6
least-squares planes. The Cp-C6F5 interplanar torsion
angles were 40.7° and 32.4°. The dimethylaminomethyl
substituent of 10 adopts a conformation (N-CH2-C(2)-
C(1) ) -54.5°) in which the pair of nitrogen nonbonding
electrons is directed toward the center of the C6F5 group
attached to the same Cp ligand (compare 7, above). The
lack of a clear trend in the disposition of the CH2N-
(CH3)2 group suggests that this substituent will not
The structure of 7 (Figure 1) is highly regular, with
no significant difference between the two Fe-Cp(cen-
troid) distances (1.651 and 1.659 Å) and a Cp(centroid)-
Fe-Cp(centroid) angle of 177.6°. The dimethylamino-
methyl substituent adopts a conformation (N-CH2-
C(2)-C(1) ) -123.7°) in which the non-bonding pair of
(17) For
a discussion of MALDI applications in oligoferrocene
characterization, see: J uhasz, P.; Costello, C. E. Rapid Commun. Mass
Spectrom. 1993, 7, 343.

Pentafluorophenyl-Substituted Ferrocenes
Organometallics, Vol. 19, No. 6, 2000 1017
F igu r e 2. Thermal ellipsoid plot of 10 shown at 50%
probability.
F igu r e 4. Thermal ellipsoid plot of 3 shown at 50%
probability.
C6 planes. Note that no points may appear above this
“boundary” line, because a centroid-centroid distance
must be at least as long as the corresponding centroid-
plane distance. As centroid-centroid distances become
shorter, interactions among the anisotropic charge
distributions responsible for stacking become stronger,
and alignment of the C6F5 groups increasingly adheres
to an empirical “rule” defined by the clustering of points
toward the left of Figure 3. Note that this “rule” does
not enforce the “idealized” (coplanar, cofacial) geometry
represented by the boundary line. We describe the
stacking interaction in 10 as “purposeful”, because its
location among the points clustered toward the left of
Figure 3 suggest that it follows the “rule”.
F igu r e 3. Scatterplot of parameters for C6F5-C6F5 arene
stacking interactions from the Cambridge Crystallographic
Database. All interactions having centroid-centroid dis-
tances less than 4.9 Å were included.
In the crystalline lattice of the diiron complex 3, each
molecule is situated on a crystallographic inversion
center, so that half of the molecule defines the asym-
metric unit. The complex adopts an S-shaped conforma-
tion (Figure 4a), in which the three fluoroaryl groups
are stacked nearly parallel to one another (Figure 4b).
The interaction is defined by a C6F5-C6F4 interplanar
(C6-C6) angle of 5.8°, a C6(centroid)-C6(centroid) dis-
tance of 3.671 Å, and two centroid-to-plane distances
of 3.383 and 3.366 Å, nearly identical to the correspond-
ing parameters in 10, again suggesting a “purposeful”
intramolecular arene stacking interaction. The Cp-C6F5
and Cp-C6F4 torsion angles (21.4° and 23.6°, respec-
tively) are relatively shallow compared to the corre-
sponding parameters in 7 and 10 (see above). The ipso
carbons, C(11) and C(17), of the C6 rings are bent
slightly out of the C5 least-squares planes of the Cp
ligands, away from the iron atom, by 0.027 and 0.032
Å, respectively. Otherwise, the ferrocene portion of the
serve as reliable design element for ordered solids
without the directing effects of hydrogen bond donors.
Further inspection of the structure of 10 reveals
intramolecular stacking of the two C6F5 groups char-
acterized by an interplanar (C6-C6) angle of 10.1°, a
centroid-to-centroid distance of 3.66 Å, and two centroid-
to-plane distances of 3.33 and 3.46 Å. We infer that a
“purposeful” stacking interaction is the reason for the
eclipsed Cp-Fe-Cp conformation observed in 10. To
better characterize this stacking interaction, we com-
pared the C6F5-C6F5 centroid-centroid distances with
data obtained from the Cambridge Crystallographic
Database (364 structures displayed C6F5-C6F5 cen-
troid-centroid distances less than 4.9 Å). The results
are shown in Figure 3 (darkened squares represent data
from 10). The line drawn in Figure 3 represents an
“idealized” stacking arrangement in which the vector
connecting the two centroids is perpendicular to both

1018 Organometallics, Vol. 19, No. 6, 2000
Deck et al.
F igu r e 5. Stereoview of packing diagram of 3.
F igu r e 6. Scatterplot of parameters for Cp-C6F5 π-stack-
ing interactions from the Cambridge Crystallographic
Database. All interactions having centroid-centroid dis-
tances less than 4.9 Å were included.
asymmetric unit exhibits a relatively undistored struc-
ture, with Fe-Cp(centroid) distances of 1.641 and 1.648
Å, parallel Cp ligands (3.9° interplanar angle), and
uniform Fe-C distances (within 0.012 Å of their mean).
The pentafluorophenyl group is planar (mean deviation
of 0.006 Å from the C6 least-squares plane), as is the
central tetrafluorophenylene group (mean deviation of
0.002 Å from the least-squares plane).
The packing diagram of 3 (Figure 5) reveals further
stacking interactions of aromatic groups in the crystal-
line lattice. The two Cp ligands of each ferrocene group
alternate with the three fluoroaryl groups of a neighbor-
ing molecule in nearly linear, infinite stacking chains.
This result contrasts both the structure of 10, in which
the dimethylaminomethyl group appears to impede
analogous stacking, and the structure of 1,1′-metallo-
cenylene-1,8-naphthalenylene co-oligomers,18 which
showed only Cp-Cp stacking. The intermolecular Cp-
(centroid)-C6F5(centroid) distance in 3 is 3.707 Å, while
the Cp(centroid)-C6F5(least-squares plane) distance is
3.474 Å. The Cp and C6F5 least-squares planes intersect
in a 21.4° angle. These data are placed into the context
of 90 other Cp-C6F5 interactions (centroid-centroid
distance less than 4.9 Å) extracted from the Cambridge
Crystallographic Data Centre (Figure 6). In accordance
with the discussion of C6F5-C6F5 interactions in 10
(above), the decreased scatter toward the left of Figure
F igu r e 7. Thermal ellipsoid plot of the dicationic portion
of m eso-12 shown at 50% probability.
6 reflects increasingly well-defined interactions at closer
Cp-C6F5 distances. Thus, the position of the filled
square (representing 3) in this portion of the scatterplot
suggests a “purposeful” Cp-C6F5 contact in crystalline
3.
The structure of 12 is shown in Figure 7. Our objective
was to assign the structure of the corresponding free
base (m eso-8), which did not form useful single crystals
from several solvent combinations that we tried. The
molecule 12 rests on a crystallographic mirror plane,
with half of the molecule as the asymmetric unit. The
structure has disordered lattice solvent molecules (chlo-
roform and perhaps water; please see the Experimental
Section), which did not interfere with relatively precise
estimation of bond lengths and angles in the molecular
structure of the dimetallic dication. One iodide ion, I(1),
rests on a 2-fold position, while I(2) and I(3) rest on 2/m
positions. Noteworthy are the conformational disposition
of the trimethylammoniomethyl substituent (N-CH2-
C(1)-C(2) ) -96.0°) directing the nitrogen anti to the
iron atom, possibly to minimize partial charge repul-
sions. The large Cp-C6F4 interplanar torsion angle
(54.8°) avoids steric interactions between the fluorine
atoms and the trimethylammonio group. The metal-
locene structure is relatively undistorted, with average
deviations from the C(1)-C(2)-C(3)-C(4)-C(5) and the
C(13)-C(14)-C(15)-C(16)-C(17) planes of 0.003 and
(18) Hudson, R. D. A.; Foxman, B. M.; Rosenblum, M. Organome-
tallics 1999, 18, 4098.

Pentafluorophenyl-Substituted Ferrocenes
Organometallics, Vol. 19, No. 6, 2000 1019
Ta ble 2. Un iqu e In ter m olecu la r C-H···F -C
In ter a ction s in th e Cr ysta l Str u ctu r es of 3, 7, 10,
12, a n d 13
compound
interaction
H‚‚‚F (Å)
C-H‚‚‚F (deg)
3
C3-H3‚‚‚F13
C4-H4‚‚‚F13
C5-H5‚‚‚F16
C8-H8‚‚‚F15
C8-H8‚‚‚F19
C9-H9‚‚‚F14
C5-H5‚‚‚F7
C3-H3‚‚‚F22
C4-H4‚‚‚F23
C5-H5‚‚‚F19
C8-H8‚‚‚F22
C8-H8‚‚‚F23
C9-H9‚‚‚F25
C10-H10‚‚‚F24
C10-H10‚‚‚F19
C3-H3‚‚‚F2
C2-H2‚‚‚F28
C3-H3‚‚‚F13
C3-H3‚‚‚F29
C8-H8‚‚‚F13
C9-H9‚‚‚F16
C18-H18‚‚‚F15
C18-H18‚‚‚F29
C18-H13‚‚‚F31
C20-H20‚‚‚F16
C21-H21‚‚‚F12
C23-H23‚‚‚F12
C25-H25‚‚‚F16
C26-H26‚‚‚F28
2.84
2.99
2.55
2.97
2.99
2.48
2.58
2.89
2.56
2.60
2.59
2.75
2.70
2.57
2.60
2.89
2.88
2.64
2.96
2.77
2.73
2.72
2.85
2.95
2.97
2.49
2.77
2.75
2.72
117.1
111.5
153.2
174.1
113.8
156.4
149.5
137.9
148.3
152.5
144.7
140.7
144.6
115.0
151.7
148.5
153.3
146.7
141.5
145.4
165.6
139.1
102.3
161.0
110.3
146.7
136.9
152.6
174.0
7
10
F igu r e 8. Structure of 13. (a) Numbering scheme with
hydrogen atoms removed for clarity. Hydrogen atoms are
numbered the same as the carbon atoms to which they are
attached. (b) Packing diagram showing intermolecular
C-H‚‚‚F-C interactions.
12
13
0.002 Å, respectively. The interplanar angle of the two
Cp ligands is 3.5°. The displacement of the C6F4 ipso
carbon from the attached Cp plane is 0.142 Å in the
direction distal to iron as in 7 and 10, above, but greater
in magnitude. This results in a “wishbone” distortion
in which the unsubstituted Cp ligands are pulled apart
from one another. The angle between the least-squares
Cp planes C(1)-C(2)-C(3)-C(4)-C(5) and C(1A)-
C(2A)-C(3A)-C(4A)-C(5A) is 14.5°.
H‚‚‚F distances and C-H‚‚‚F angles. However, their
study was limited to a very narrow range of compounds
(fluorinated benzenes), wherein rigid structure, high
molecular symmetry, relatively constant molecular
“shape”, and a lack of other competing structural
influences (such as intramolecular C-H‚‚‚F interac-
tions) contribute to a much cleaner correlation among
selected metric parameters. They concluded that only
compounds consisting of C, H, and F atoms are “ap-
propriate” for the study of C-H‚‚‚F interactions and
that the C-H groups must be “acidic”. The latter
criterion is ostensibly satisfied in ferrocenes, which have
relatively acidic Cp C-H bonds, evidenced by their facile
metalation. We assert that the presence of coordina-
tively saturated iron atoms embedded in complexes such
as 3 or 13 should not make these species “inappropriate”
for the study of C-H‚‚‚F interactions.
H‚‚‚F distances and C-H‚‚‚F angles for complexes 3,
7, 10, 12, and 13 are listed in Table 2 and presented in
a scatterplot (Figure 9). The small, unfilled circles
(Figure 9) represent the data of Desiraju and co-workers
for fluorinated benzenes.10 Indeed, their data show an
impressive inverse correlation with very little scatter.
This correlation can be considered a “rule” for C-H‚‚‚
F-C interactions. Next, consider the data for complex
3 (filled circles in Figure 9), which follows the same
trend except for a single interaction in the upper right
quadrant of the scatterplot. We infer that the C-H‚‚‚
F-C interactions in 3 (dotted lines in Figure 5) repre-
sent “purposeful” interactions in the sense that they
follow the “rule” established by the fluorinated ben-
zenes, with one additional, “diffuse” C-H‚‚‚F-C inter-
action accompanying the otherwise optimal packing
arrangement. In contrast, the packing in complex 13 is
dominated by a large number of “diffuse” interactions,
represented by highly scattered squares in the upper
Inter- and intramolecular C-H‚‚‚F interactions are
prevasive in the crystal structures of 3, 7, 10, 12, and
13. The synthesis and crystallographic characterization
of complex 13 are reported elsewhere;5 please refer to
the numbering scheme in Figure 8. Several investiga-
tors have addressed the issue of whether such C-H‚‚‚
F interactions should be considered “hydrogen bonds”.19
Dunitz and Taylor19b suggested that an H‚‚‚F distance
less than 2.3 Å is an appropriate criterion for hydrogen
bonding, because the sum of the van der Waals radii of
hydrogen and fluorine is in the range 2.5-2.7 Å20 and
because H‚‚‚N and H‚‚‚O hydrogen bonds generally
exhibit H‚‚‚X distances that are significantly shorter
than the sum of their respective van der Waals radii.
They concluded that C-F bonds are poor hydrogen-bond
acceptors and that donors such as O-H, N-H, and C-H
will interact with stronger acceptors if available. In
general, interactions among organic C-F and C-H
groups must often coexist with various other intermo-
lecular packing motifs, which leads to a broad range of
observed H‚‚‚F distances and C-H‚‚‚F angles.10 In the
absence of stronger acceptors, however, “close” C-H‚‚
‚F-C contacts are relatively common: 39 compounds
having H‚‚‚F distances of less than 2.35 Å were found
among 360 compounds in the Cambridge Crystal-
lographic Database containing sp2-hybridized C-F
bonds.21 Desiraju and co-workers10 advanced the alter-
native criterion of a strict, inverse correlation between
(19) (a) Grepioni, F.; Cojazzi, G.; Draper, S. M.; Scully, N.; Braga,
D. Organometallics 1988, 7, 296. (b) Dunitz, J .; Taylor, R. Chem. Eur.
J . 1997, 3, 89. (c) Plenio, H. Chem. Rev. 1997, 97, 3363.
(20) (a) Bondi, A. J . Phys. Chem. 1964, 68, 441. (b) Rowland, R. S.;
Taylor, R. J . Phys. Chem. 1996, 100, 7384.
(21) Howard, J . A. K.; Hoy, V. J .; O’Hagan, D.; Smith, G. T.
Tetrahedron 1996, 52, 12613.

1020 Organometallics, Vol. 19, No. 6, 2000
Deck et al.
F igu r e 9. Scatterplot of H‚‚‚F and C-H‚‚‚F angles: (b)
data for 3; (/) data for 7; (4) data for 10; (0) data for 13;
(O) data for fluorinated benzenes from ref 10.
F igu r e 11. Packing diagram of 12 (dication) showing
intermolecular C-H‚‚‚F-C interactions as dotted lines.
Ta ble 3. Un iqu e In tr a m olecu la r C-H···F -C
In ter a ction s in th e Cr ysta l Str u ctu r es of 3, 7, 10,
12, a n d 13
compound
interaction
H‚‚‚F (Å)
C-H‚‚‚F (deg)
3
C2-H2‚‚‚F16
C5-H5‚‚‚F12
C7-H7‚‚‚F19
C10-H10‚‚‚F18
C5-H5‚‚‚F11
C17-H17‚‚‚F7
C5-H5‚‚‚F21
C5-H5‚‚‚F15
C7-H7‚‚‚F19
C7-H7‚‚‚F25
C10-H10‚‚‚F21
C3-H3‚‚‚F2
C13-H13‚‚‚F1
C17-H17‚‚‚F1
C2-H2‚‚‚F12
C5-H5‚‚‚F16
C6-H6‚‚‚F12
C18-H18‚‚‚F32
C21-H21‚‚‚F28
C26-H26‚‚‚F32
2.37
2.33
2.38
2.38
2.74
2.63
2.61
2.53
2.72
2.52
2.49
2.87
2.62
2.72
2.55
2.53
2.87
2.67
2.59
2.75
108.2
108.3
106.9
107.4
92.2
109.0
114.3
97.9
108.7
101.8
103.9
86.3
104.3
100.7
104.8
100.6
111.1
106.0
98.6
7
10
F igu r e 10. Packing diagram of 13 showing intermolecular
C-H‚‚‚F-C interactions as dotted lines.
12
13
right quadrant of Figure 9. The lack of specific, “pur-
poseful” H‚‚‚F contacts in 13 is also reflected in the
extensive, unsymmetrical bifurcation of these interac-
tions, as illustrated in Figure 10. Next, consider the data
for complex 10 (triangles in Figure 9), which shows a
weak distance-angle correlation, excluding the single
point in the lower left quadrant. We propose that the
six weaker C-H‚‚‚F-C interactions and the irregular
shape of the molecule constrain the last C-H‚‚‚F-C
interaction into an “awkward” geometry in order to
optimize the overall packing arrangement. The monoaryl-
ated complex 7 exhibited only one C-H‚‚‚F-C interac-
tion less than 3.0 Å (asterisk in Figure 9), while the
single C-H‚‚‚F-C interaction in 12 exhibited two,
symmetry-related interactions (dotted lines in Figure
11, plus symbol in Figure 9). It should be noted that
the closest N‚‚‚H contacts in 7 are 2.87 and 3.02 Å, while
there are no N‚‚‚H interactions evident in 10, suggesting
that F-C acceptors compete effectively with the nitro-
gen acceptors for C-H donors in these compounds.
Considering the relatively narrow range of C6F5-Cp
torsion angles in our compounds (see above), it is not
surprising that the metric data for intramolecular C-H‚
‚‚F-C interactions (Table 3) are also narrowly distrib-
uted. Figure 12 illustrates these intramolecular C-H‚
‚‚F-C contacts. Two polymorphs of 2,3,4,5,6-penta-
fluorobiphenyl have been characterized crystallo-
110.9
graphically,22 and these show analogous ortho-ortho H‚
‚‚F distances of 2.68 and 2.80 Å, similar to the values
shown in Table 3.
Electr och em ica l An a lysis of 3. Cyclic voltammetric
analysis of 3 showed a single, reversible oxidation/
reduction wave having an E1/2 of 340(10) mV relative
to Fc+|Fc in dichloromethane. The Osteryoung square
wave voltammogram (Figure 13) showed a slightly
broadened signal (fwhm ) 132 mV), indicating a small,
unresolved splitting of two FeIII|FeII couples. The dif-
ference in line widths between the diiron complex (3,
fwhm ) 132 mV) and the internal standard (ferrocene,
fwhm ) 106) suggests a difference of only about 26(10)
mV between the first and second oxidations of 3, which
would then be characterized by E1/2 values of 323(10)
and 349(10) mV. These oxidation potentials are close
to the value of 345(5) mV observed for (C6F5Cp)2Fe,11
(22) Brock, C. P.; Naae, D. G.; Goodhand, N.; Hamor, T. A. Acta
Crystallogr. B 1978, 34, 3691.

Pentafluorophenyl-Substituted Ferrocenes
Organometallics, Vol. 19, No. 6, 2000 1021
F igu r e 13. Square wave voltammogram of 3 showing
estimation of full-width at half-maximum (fwhm) for 3 and
internal ferrocene.
C6H4)(3,4-Me2Cp)Fe(Me5Cp), showed significant split-
ting (124 mV) in the solution voltammogram.23 This
comparison suggests that either the ancillary substit-
uents (one C6F5 vs seven methyl groups) or the fluorina-
tion of the phenylene spacer (C6F4 vs C6H4) weakens
the electronic interaction of the two iron centers. Work
is underway in our laboratories to examine each of these
possibilities. Neither 7, 8, nor 12 showed reversible
solution voltammetric behavior.
Con clu sion s. Nucleophilic aromatic substitution re-
actions of lithioferrocenes with hexafluorobenzene is a
versatile approach to organometallic complexes in which
ferrocenes are linked either to pentafluorophenyl sub-
stituents or to one another via 1,4-tetrafluorophenylene
spacers. This approach enables the construction of
oligomeric species via a step-growth process, although
the formation of high polymers by this approach is
prevented by the many side-reactions attending both the
lithiation and aromatic substitution processes, as indi-
cated by the number and quantity of byproducts ob-
served in model reactions leading to monometallic and
dimetallic species. A single dimethylaminomethyl sub-
stituent on the starting ferrocene offers improved oli-
gomer solubility as well as planar-chirality leading to
diastereomerism in 1,4-tetrafluorophenylene-linked di-
iron complexes. We are interested in complexes such as
8 as bidentate ligands for transition metals. Crystal
structures of C6F5-substituted and 1,4-C6F4-linked fer-
rocenes exhibit “purposeful” arene stacking interactions
as well as both “purposeful” and “diffuse” C-H‚‚‚F-C
interactions. We are presently examining, by experi-
mental and computational methods, a hypothesis that
the Cp-C6F5 interplanar torsion angles in pentafluo-
rophenyl-substituted Cp complexes reflect electron de-
mand by the coordinated metal fragment.
F igu r e 12. Diagrams showing intramolecular C-H‚‚‚F
contacts. N(CH3)2 groups were omitted for clarity: (a) 7;
(b) 10; (c) asymmetric unit of 12, cation only.
suggesting that the 1,4-C6F4 group exerts a strong
electron-withdrawing influence on both attached ferro-
cenyl groups.
This electrochemical data suggest that electronic
communication between the two ferrocenyl groups
through the tetrafluorophenylene spacer is weak and
that complexes such as 3 and their higher oligomeric
homologues do not hold promise as “molecular wires”.
This result was somewhat surprising, as ferrocenes
linked by the corresponding unsubstituted 1,4-phen-
ylene group, for example, (Me5Cp)Fe(3,4-Me2Cp)(1,4-
(23) (a) Bunel, E.; Campos, P.; Ruz, J .; Valle, L.; Chadwick, I.; Santa
Ana, M.; Gonazlez, G.; Manriquez, J . M. Organometallics 1988, 7, 474.
(b) Iyoda, M.; Kondo, T.; Okabe, T.; Matsuyama, H.; Sasaki, S.;
Kuwatani, Y. Chem. Lett. 1997, 35.

1022 Organometallics, Vol. 19, No. 6, 2000
Deck et al.
precipitate on a filter and washing with pentane. The dilithio-
ferrocene and TMEDA remaining in this product mixture were
assayed by 1H NMR in THF-d8 solution. Triturating these
mixtures with THF to remove TMEDA prior to the fluoroaryl-
ation reactions did not result in significantly different results.
Syn th esis of 1,1′-Bis(p en ta flu or op h en yl)fer r ocen e (2).3
A solution of ferrocene (9.3 g, 0.050 mol), n-butyllithium (69
mL, 1.6 M in hexanes, 0.11 mol), and TMEDA (12 g, 0.10 mol)
in hexanes (250 mL) was stirred at 25 °C for 30 h. The mixture
was allowed to settle, and the supernatant was decanted with
a canulla. The remaining solid was washed with an additional
100 mL of hexanes. The resulting orange solid was taken up
as a slurry in 100 mL of hexanes and added, via canulla, to a
stirred soltuion of hexafluorobenzene (75 g, 0.40 mol) in THF
(300 mL) maintained at 0 °C using an ice bath. After the
addition was complete, the reaction was stirred at 25 °C for
12 h. The volatile components were evaporated, and the
resulting red residue was recrystallized from toluene/hexanes
(with filtration to remove LiF and other insoluble substances).
A yield of about 60% was obtained in three crops of orange-
red crystals. NMR spectra were consistent with those previ-
ously published.
Exp er im en ta l Section
Gen er a l Con sid er a tion s. All reactions were carried out
using standard inert-atmosphere techniques.24 NMR spectra
were recorded on either a Varian U-400 or a Bruker AM-360
instrument. Elemental microanalysis was performed by Oneida
Research Services (Whitesboro, NY) or Desert Analytics
(Tucson, AZ). Melting points were obtained in open capillaries
and are uncorrected. Ferrocene, (dimethylaminomethyl)fer-
rocene, and hexafluorobenzene were used as received from
Aldrich, Strem, and PCR (now Lancaster), respectively. TME-
DA was distilled from CaH2 under nitrogen and stored over
molecular sieves.
Cr ysta llogr a p h ic Stu d ies. Crystals of 7 and of 10 were
obtained by cooling concentrated hexanes solutions of the
respective complexes to -5 °C. Crystals of 4 were obtained by
allowing a warm, concentrated 1,2-dichloroethane solution of
the complex to stand at 25 °C. Crystals of 12‚3(CHCl3) were
obtained by dropwise addition of water to a warm, stirred
suspension of 12 in methanol/chloroform until the solution
became clear; subsequent filtration and cooling to -5 °C for 2
days afforded small yellow-orange prisms. Data-collection and
structure-refinement information for all four complexes com-
plexes is summarized in Table 1. Complete details of the
crystallographic data collection, structure solution, and refine-
ment are included in the Supporting Information.
Of the four complexes, the bis(methiodide) (12‚3CHCl3)
warrants further discussion of crystallographic analysis. Suc-
cessful refinement of this structure was plagued by residual
electron density in the solvent region that could not be modeled
in a satisfactory way. Attempts were made to model this
disorder as both methanol and water. The residual peaks in
no way resembled methanol, so this model was quickly
rejected. Assignment of the residual peaks as two partially
occupied oxygens of water improved the R-factor, but this
model was deemed unreasonable, as the oxygen positions were
not located within hydrogen-bonding distance of any polar
residues. In the end, the three strongest Fourier peaks (3.07,
2.07, and 2.02 e-) were not modeled in the final refinement.
Electr och em ical Measu r em en ts. Single-sweep cyclic (CV)
and Osteryoung square-wave (OSWV) voltammetric data were
obtained for the dimeric complex 3 at a nominal concentration
of 2 µM in CH2Cl2 using 0.10 M [n-Bu4N][PF6] as the
electrolyte and activated alumina as an internal desiccant. The
apparatus was a BAS 100B potentiostat with a glassy Pt
working electrode, a Pt wire auxiliary electrode, and a Ag/AgCl
reference electrode. The CV sweep was initialized at 0 mV,
scanned to 1200 mV, and returned to 0 mV, at a scan rate of
100 mV s-1. Reported E1/2 values represent the average of two
experiments. |Eox - Ered| ranged from 70 to 80 mV, and Ic/Ia
was within 85% of unity, both indicators of substantially
reversible oxidation. Increases in scan rate did not improve
reversibility. The oxidation potential of 3 relative to internal
Cp2Fe was obtained by Osteryoung square wave voltammetry
(OSWV). The cell voltage was swept from 0 mV to +1200 mV
with a step resolution of 2 mV, a square wave amplitude of 25
mV, and frequency of 15 Hz. Internal resistance compensation
was applied to all voltammetric experiments.
Syn th esis of 1,4-Bis[1′-(p en ta flu or op h en yl)fer r ocen -
1-yl]tetr a flu or oben zen e (3). C6F6 (0.010 mol) was added in
one portion to a solution of 1 (0.010 mol) in THF (50 mL) at
25 °C. An exothermic reaction occurred with the formation of
a dark precipitate. After stirring the mixture for 1 day at 25
°C, water (50 mL) was added. The precipitate was collected
on a filter, washed with water, and dried under vacuum to
afford 3.1 g of a red-orange solid, which was transferred to a
cellulose thimble and extracted in a Soxhlet apparatus first
with hexanes for 12 h. The extract was evaporated to afford
1.27 g of an orange solid. This solid was washed with hexanes
(3 × 100 mL) at 25 °C to give a red residue and an orange
solution. The orange solution was evaporated to afford about
20% of 2. The red residue was recrystallized from benzene to
afford about 10% of 3 as a red microcrystalline solid. An
analytical sample of 3 was obtained by sublimation at 150 °C
1
(2 × 10-5 mmHg): mp 240 °C (dec); H NMR (CDCl3) δ 4.83
(m, 4 H), 4.77 (m, 4 H), 4.56 (t, J ) 2.0 Hz, 4 H), 4.43 (t, J )
2.0 Hz, 4 H); 19F NMR (CDCl3) δ -140.7 (d, J ) 22.6 Hz, 4 F),
-142.5 (s, 4 F), -159.1 (t, J ) 21.4 Hz, 2 F), -164.3 (s, 4 F).
Anal. Calcd for C38H16F14Fe2: C, 53.86; H, 1.90. Found: C,
53.71; H, 1.91.
Extr a ction of 1,1′-F er r ocen ylen e-1,4-tetr a flu or op h e-
n ylen e Co-oligom er s (4). The thimble residue remaining
after hexanes extraction in the preparation of 3 (above) was
subjected to successive 12-h extractions with other solvents.
The extractor barrel was heated with an external Nichrome
wire to maintain the collecting liquids near their respective
boiling points. Benzene extracted 1.21 g of a red-orange solid;
toluene extracted 190 mg of a red solid; chlorobenzene
extracted 210 mg of a deep red solid; and 1,2-dichlorobenzene
extracted 190 mg of a shiny red-brown solid. A black residue
(200 mg) remained in the thimble.
Syn th esis of 1-(P en ta flu or op h en yl)-2-(N,N-d im eth y-
la m in om eth yl)fer r ocen e (7). A solution of N,N-(dimethy-
laminomethyl)ferrocene (1.22 g, 5.00 mmol), n-BuLi (3.5 mL
of a 1.6 M solution in hexanes, 5.8 mmol), and TMEDA (580
mg, 5.00 mmol) in pentane (50 mL) was stirred for 36 h. To
the resulting dark turbid mixture was added hexafluoroben-
zene (2.0 g, 11 mmol) in one portion, and the mixture was
stirred at 25 °C for 2 h. After removal of the solvent under
vacuum, the dark residue was subjected to chromatography
on neutral alumina (2 × 10 cm). The first orange band was
eluted with benzene and evaporated to afford 300 mg (0.733
mmol, 60%) of an orange oil, which crystallized upon standing.
An analytical sample was obtained by recrystallization from
hexanes at -20 °C: 1H NMR (CDCl3) δ 4.35 (m, 1 H), 4.28 (m,
2 H), 4.07 (s, 5 H), 3.29 (d, 2J HH ) 13.6 Hz, 1 H), 3.09 (dt, 2J HH
) 13.6 Hz, J HF ) 1.2 Hz, 1 H), 1.87 (s, 6 H); 19F NMR (CDCl3)
Ma ss Sp ect r om et r y. The direct probe EI, CI, and FAB
mass spectra were obtained using a Fisons VG Quattro
instrument. MALDI-TOF spectra were obtained by the Wash-
ington University Mass Spectrometry Resource.
Isola tion of Dilith iofer r ocen e (1).12 In this context,
“dilithioferrocene” (1) refers to the mixture of hydrocarbon-
insoluble complexes obtained by treatment of ferrocene with
2.2 equiv of butyllithium and 2.2 equiv of TMEDA in hexane
or pentane at 25 °C for 36 h, followed by collection of the
(24) Solvent purification: (a) Pangborn, A. B.; Giardello, M. A.;
Grubbs, R. H.; Rosen, R. K.; Timmers, F. J . Organometallics 1996,
15, 1518. General methods: (b) Shriver, D. F.; Drezdzon, M. A. The
Manipulation of Air-Sensitive Compounds; Wiley: New York, 1986.

Pentafluorophenyl-Substituted Ferrocenes
Organometallics, Vol. 19, No. 6, 2000 1023
3
3
δ -136.78 (d, J ) 21 Hz, 2 F), -157.95 (t, J ) 21 Hz, 1 F),
tion of freshly prepared and hexanes-washed 9 (500 mg, 1.35
mmol) in 10 mL of THF was added to a stirred solution of
hexafluorobenzene (4.0 g, 21.5 mmol) in 50 mL of THF at 25
°C. The resulting red solution was stirred for 0.5 h at 25 °C,
heated under reflux for 5 min, and then cooled. After removing
the solvent under vacuum, the resulting dark residue was
subjected to chromatography on neutral alumina (2 × 10 cm).
A colorless forerun (200 mL of hexanes) was discarded. The
first orange band was eluted with 1:1 CH2Cl2/hexanes and
evaporated to afford 440 mg (57%) of an orange-red viscous
oil. Crystallization from hexanes at -20 °C affored 340 mg
(44%) of red crystals: 1H NMR (CDCl3) 4.78 (m, 1 H), 4.74
(m, 1 H), 4.44 (m, 2 H), 4.40 (m, 1 H), 4.35 (t, J ) 2.0 Hz, 1 H),
3.26 (d, 2J ) 13.6 Hz, 1 H), 3.14 (dt, 2J HH ) 13.6 Hz, J HF ) 1.2
Hz, 1 H), 1.83 (s, 6 H); 19F NMR (CDCl3) δ -136.75 (d, 3J ) 20
Hz, 2 F), -140.70 (m, 2 F), -157.27 (t, 3J ) 21 Hz, 1 F),
-164.10 (m, 2 F); 13C NMR (CDCl3) δ 144.5 (d, 1J CF ) 250 Hz,
1
1
CF), 139.6 (d, J CF ) 251 Hz, CF), 137.7 (d, J CF ) 248 Hz,
2
4
CF), 113.5 (td, J CF ) 15 Hz, J CF ) 3 Hz, C6F5 ipso C), 85.2
(C), 71.6 (m, C), 71.2 (s, CH), 70.1 (s, C5H5), 69.4 (t, J CF ) 4
Hz, CH), 68.0 (s, CH), 57.5 (t, J CF ) 3 Hz, CH2), 44.8 (s, CH3).
Anal. Calcd for C19H16F5FeN: C, 55.77; H, 3.94; N, 3.42.
Found: C, 55.61; H, 4.02; N, 3.41.
Syn th esis of 1,4-Bis[2-(d im eth yla m in om eth yl)fer r o-
cen -1-yl]tetr a flu or oben zen e (8). A solution of N,N-(dim-
ethylaminomethyl)ferrocene (4.86 g, 20.0 mmol), n-BuLi (13
mL of a 1.6 M solution in hexanes, 21 mmol), and TMEDA
(2.44 g, 21.0 mmol) in hexanes (100 mL) was stirred for 15 h.
An orange precipitate formed. The precipiate was allowed to
settle, and the orange supernatant was decanted into satu-
rated aqueous NaCl, using an additional 50-mL portion of
hexanes to wash the precipitate. The biphasic, hydrolyzed
supernatant mixture was extracted with hexanes, and the
organic layer was washed with brine, dried over MgSO4,
filtered, and evaporated to afford 1.2 g (4.9 mmol) of pure N,N-
(dimethylaminomethyl)ferrocene (5) as determined by 1H NMR
analysis. This suggested that only 15.1 mmol of the desired
2-lithiated intermediate (6) remained in the reaction flask. To
the orange solid in the reaction flask was added hexanes (100
mL) followed by hexafluorobenzene (1.4 g, 7.5 mmol) in one
portion, and the mixture was heated under reflux for 2 h,
cooled, and evaporated. The resulting orange residue was
taken up in 100 mL of dichloromethane, filtered through
Celite, and evaporated to afford 4.9 g of an orange solid. The
crude product was triturated with pentane (2 × 50 mL at 0
°C) to afford 3.58 g (5.66 mol, 76%) of an orange solid, which
was found by 1H NMR and 19F NMR analyses to comprise a
3:2 mixture of m eso-8 and d l-8. The combined triturant liquors
were evaporated to afford 1.1 g of a waxy, orange solid, which
was analyzed by 19F NMR; about 45% of the integrated signal
intensity corresponded to d l-8 and about 40% to 7, and the
remaining signals were not assigned. The crude mixture of
m eso-8 and d l-8 was extracted with hexanes until the filtrate
ran clear (10 × 50 mL with stirring). [Note: Soxhlet extraction
with pentane would probably work very well for this separa-
tion.] The remaining orange filter cake was nearly pure m eso-8
(2.0 g, 45% yield based on C6F6, containing 5% of d l-8 as
determined by 1H NMR analysis). The filtrate was evaporated
to afford nearly pure d l-8 (1.4 g, 30% based on C6F6, containing
3% of m eso-8 as determined by 1H NMR analysis). Data for
m eso-8: 1H NMR (CDCl3) δ 4.48 (m, 1 H), 4.46 (m, 1 H), 4.40
3
-158.86 (t, J ) 21 Hz, 1 F), -163.73 (m, 2 F), -163.94 (m, 2
F); 13C NMR (CDCl3) δ 144.3 (d, 1J CF ) 240 Hz, two CF), 139.6
1
1
(d, J CF ) 250 Hz, CF), 138.8 (d, J CF ) 245 Hz, CF), 137.8 (d,
1J CF ) 250 Hz, CF), 137.6 (d, J CF ) 240 Hz, CF), 113.3 (dt,
1
2J CF ) 15 Hz, J CF ) 4 Hz, C6F5 ipso C), 112.4 (dt, J CF ) 16
4
2
4
Hz, J CF ) 4 Hz, C6F5 ipso C), 86.2 (s, CH), 73.0 (s, CH), 72.5
3
3
(t, J CF ) 2 Hz, C-C6F5), 72.4 (t, J CF ) 2 Hz, C-C6F5), 72.0
(t, J CF ) 2 Hz, CH), 71.9 (t, J CF ) 1 Hz, CH), 71.5 (t, J CF ) 4
Hz, CH), 71.4 (m, C), 70.9 (t, J CF ) 5 Hz, CH), 57.0 (t, J CF
4 Hz, CH2), 44.8 (s, CH3). Anal. Calcd for C25H15F10FeN: C,
)
52.20; H, 2.63; N, 2.43. Found: C, 52.25; H, 2.78; N, 2.43.
Isola tion a n d Ch a r a cter ia tion of 1,1′-(2-Dim eth yla m i-
n om eth yl)fer r ocen ylen e-1,4-tetr a -flu or op h en ylen e Co-
oligom er s (11). C6F6 (1.1 g, 6.0 mmol) was added to a solution
of freshly prepared, hexanes-washed 7 (2.2 g, 5.9 mmol) in 40
mL of THF at 25 °C. After the initial exotherm subsided, the
mixture was stirred under reflux for 2 h. After cooling the
reaction and removing the solvent under vacuum, aqueous
workup of the benzene-soluble red residue afforded 2.6 g of a
red-orange oil, which solidified upon standing. A small portion
(200 mg) was purified by column chromatography on alumina
(20 mm × 10 cm). A forerun (150 mL of ether) was eluted,
which was subsequently evaporated to afford about 30 mg of
1
a red solid identified as a mixture of 7 and 10 by H and 19F
NMR analysis. A dark red band was then eluted with
methanol. Evaporation of the solvent afforded 0.17 g of a red
solid designated as the “purified” oligomeric product 11. A
brown band remained at the top of the column. Details of the
NMR spectroscopic data and end-group analysis of 11 are
presented in the Supporting Information. None of the EI, CI,
FAB, or MALDI spectra that we obtained for 11 showed
masses higher than 2000 amu and lacked clear progressions
that we could assign to the repeat unit of 11. Evidently, the
dimethylaminomethyl substituents make 11 susceptible to
extensive fragmentation. Gel-permeation chromatography (GPC)
experiments gave spurious results (very high elution volumes
or long retention times), suggesting molecular weights of less
than 500 when calibrating to polystyrene standards. In-line
viscometry gave insufficient signal for an absolute molecular
weight determination.
2
2
(m, 1 H), 4.21 (s, 5 H), 3.48 (d, J ) 13.6 Hz, 1 H), 3.28 (d, J
) 13.6 Hz, 1 H), 2.01 (s, 6 H); 19F NMR (CDCl3) δ (referenced
to internal C6F6 at -163.00) -138.80 (s, 4 F); {1H}13C NMR
(CDCl3) δ 85.1 (C), 72.7 (C), 71.1 (CH), 70.1 (C5H5) 69.4 (CH),
68.1 (CH), 57.4 (CH2), 44.8 (NMe2); {19F}13C NMR (CDCl3) δ
1
144.5 (CF), 116.3 (C1 and C4 of C6F4), 70.1 (d, J CH ) 175 Hz,
C5H5). Anal. Calcd for C32H32F4Fe2N2: C, 60.79; H, 5.10; N,
4.43. Found: C, 61.05; H, 5.05; N, 4.08. Data for d l-8: 1H NMR
(CDCl3) δ 4.47 (m, 2 H), 4.40 (m, 1 H), 4.21 (s, 5 H), 3.46 (d, 2J
) 13.5 Hz, 1 H), 3.25 (d, J ) 13.5 Hz, 1 H), 1.99 (s, 6 H); 19F
2
NMR (CDCl3) δ (referenced to internal C6F6 at -163.00)
-138.65 (s, 4 F).
m eso-8-Bis(m eth iod id e) (m eso-12). A 20-mL test tube
was charged with 101 mg (0.160 mmol) of m eso-8. Methanol
(5 mL) and iodomethane (2 mL) were added, and the resulting
orange slurry was stirred at room temperature. No reaction
was apparent. Upon boiling for 2 min, the mixture first became
a clear orange solution, and then a yellow-orange precipitate
separated. A second 2-mL aliquot of iodomethane was added
(no visible change), and the mixture was boiled for an ad-
ditional 2 min and then cooled. The yellow solid was collected
on a Hirsch funnel, washed with ether, and dried in a vacuum
oven at 60 °C for 4 h to obtain 132 mg (0.144 mmol, 90.2%) of
a fine yellow powder: 1H NMR (DMSO-d6) δ 4.90 (m, 1 H),
Isola tion of 1,1′-Dilith io-2-(d im eth yla m in om eth yl)fer -
r ocen e (9). Treatment of N,N-dimethylaminomethylferrocene
with 2.2 equiv of n-BuLi and 1.1 equiv of TMEDA formed a
dark precipitate, which was collected on a filter, washed with
pentane, and dried under vacuum. The 1H NMR spectrum
(THF-d8) of the resulting red solid showed several broad
signals consistent with 9 containing about 1 equiv of residual
TMEDA. About 20% of the integrated signal intensity was not
assigned. This organolithium compound was found to be
unstable over a period of a few days even when stored in our
glovebox freezer (-35 °C).
2
4.78 (m, 1 H), 4.73 (m, 1 H), 4.67 (d, J ) 14 Hz, 1 H), 4.48 (s,
Syn th esis of 1,1′-Bis(p en ta flu or op h en yl)-2-(N,N-d im -
eth yla m in om eth yl)fer r ocen e (10). Using a canulla, a solu-
5 H), 4.17 (d, 2J ) 14 Hz, 1 H), 2.76 (s, 9 H); 19F NMR (DMSO-
d6) -135.2 (br s). Satisfactory elemental analysis was not

1024 Organometallics, Vol. 19, No. 6, 2000
Deck et al.
obtained in two attempts. Anal. Calcd for C34H38F4Fe2I2N2: C,
44.57; H, 4.18; N, 3.06. Found: C, 43.25 and 43.26; H, 3.78
and 3.81; N, 3.15 and 2.95. The assigned structure was
confirmed by single-crystal X-ray diffraction (see above).
the EI, CI, and FAB mass spectra, and the Washington
University Mass Spectrometry Resource for the MALDI-
TOF mass spectra. We thank Prof. H. W. Gibson and
Prof. M. R. Anderson for helpful discussions.
Ack n ow led gm en t. This work was supported by the
Petroleum Research Fund (30738-G1). We thank Prof.
K. J . Brewer for the use of her digital potentiostat and
Dr. M. R. J ordan for assistance with electrochemical
experiments. We thank Prof. J . E. McGrath and Dr. Q.
J i for assistance with the GPC analyses, K. Harich for
Su p p or tin g In for m a tion Ava ila ble: Crystallographic
data for 3, 7, 10, and 12; NMR spectral assignments and end-
group analysis of 11. This material is available free of charge
via the Internet at http://pubs.acs.org.
OM990930V