C-F activation reactions of (pentafluorophenyl)cyclopentadiene and 3-(pentafluorophenyl)indene with tetrakis(dimethylamido)titanium(IV)

Article abstract of DOI:10.1021/om034385g

Reactions of either 3-(pentafluorophenyl)indene (1a) or (pentafluorophenyl)cyclopentadiene (1b) with Ti(NMe₂)₄ afford products in which one or both of the ortho fluorines of the C₆F₅ group are replaced with a dimethylamino group. The ortho selectivity suggests a titanium-mediated, intramolecular nucleophilic aromatic substitution mechanism. The organic products 3-[2-(dimethylamino)tetrafluorophenyl]indene (4a), 3-[2,6-bis(dimethylamino)trifluorophenyl]indene (7a), and 1-[2,6-bis(dimethylamino)trifluorophenyl]cyclopentadiene (7b) were isolated upon hydrolytic workup. The substituted cyclopentadiene (7b) was converted to the corresponding ferrocene derivative (12) by treatment first with sodium hydride and then with iron dibromide in THF. Three compounds (4a, 7b, and 12) were characterized crystallographically.

Full text of DOI:10.1021/om034385g

Organometallics 2004, 23, 1089-1097
1089
C-F Activa tion Rea ction s of
(P en ta flu or op h en yl)cyclop en ta d ien e a n d
3-(P en ta flu or op h en yl)in d en e w ith
Tetr a k is(d im eth yla m id o)tita n iu m (IV)
Paul A. Deck,* Mariam M. Konate´,† Bryte V. Kelly,† and Carla Slebodnick
Department of Chemistry, Virginia Tech, Blacksburg, Virginia 24061
Received December 18, 2003
Reactions of either 3-(pentafluorophenyl)indene (1a ) or (pentafluorophenyl)cyclopentadiene
(1b) with Ti(NMe2)4 afford products in which one or both of the ortho fluorines of the C6F5
group are replaced with a dimethylamino group. The ortho selectivity suggests a titanium-
mediated, intramolecular nucleophilic aromatic substitution mechanism. The organic
products 3-[2-(dimethylamino)tetrafluorophenyl]indene (4a ), 3-[2,6-bis(dimethylamino)-
trifluorophenyl]indene (7a ), and 1-[2,6-bis(dimethylamino)trifluorophenyl]cyclopentadiene
(7b) were isolated upon hydrolytic workup. The substituted cyclopentadiene (7b) was
converted to the corresponding ferrocene derivative (12) by treatment first with sodium
hydride and then with iron dibromide in THF. Three compounds (4a , 7b, and 12) were
characterized crystallographically.
In tr od u ction
than corresponding alkyls, allowing stepwise mecha-
nistic exploration, especially of late transition metal
oxidative addition pathways.11,18,19 Second, substituents
can impart rational activating/directing effects without
encumbering the reactive bonds. Finally, the desired
outcome of “activation” is usually formal substitution
rather than elimination, and although elimination is
known for arenes,20 the selectivity for substitution is
generally better.
As part of our investigation of perfluoroaryl-substi-
tuted metallocenes and related complexes,21-25 we want
to synthesize group 4 complexes of the 1-(pentafluo-
rophenyl)indenyl ligand. Certain zirconium and hafni-
um complexes of the isomeric 2-(pentafluorophenyl)-
indenyl ligand are stable and useful as catalysts for
olefin polymerization.26-29 Using established methods30-32
Selective activation of polyfluorinated organic com-
pounds is an important synthetic challenge and a
rapidly advancing area of investigation.1 Transition
metal complexes offer diverse mechanisms for CF
activation.2-17 In both C-H and C-F activation studies,
aromatic substrates are used more commonly than
aliphatic substrates for three main reasons. First, metal
aryls are often more stable and more readily isolated
* Corresponding author. E-mail: pdeck@vt.edu. Fax: (540)-231-
3255.
† Undergraduate research participant.
(1) Kiplinger, J . L.; Richmond, T. G.; Osterberg, C. E. Chem. Rev.
1994, 94, 373-431.
(2) Braun, T.; Noveski, D.; Neumann, B.; Stammler, H. G. Angew.
Chem., Intl. Ed. 2002, 41, 2745-2747.
(3) Albietz, P. J .; Houlis, J . F.; Eisenberg, R. Inorg. Chem. 2002,
41, 2001-2003.
(4) Fujiwara, M.; Ichikawa, J .; Okauchi, T.; Minami, T. Tetrahedron
Lett. 1999, 40, 7261-7265.
(18) Sladek, M. L.; Braun, T.; Neumann, B.; Stammler, H. G. J .
Chem. Soc., Dalton Trans. 2002, 297-299.
(5) Hughes, R. P.; Zhang, D.; Zakharov, L. N.; Rheingold, A. L.
Organometallics 2002, 21, 4902-4904.
(19) Bo¨hm, V. P. W.; Gsto¨ttmayr, C. W. K.; Weskamp, T.; Herrmann,
W. A. Angew. Chem., Intl. Ed. 2001, 40, 3387-3389.
(20) Kraft, B. M.; Lachiotte, R. J .; J ones, W. D. Organometallics
2002, 21, 727-731.
(6) Hughes, R. P.; Smith, J . M.; Incarvito, C. D.; Lam, K. C.;
Rhatigan, B.; Rheingold, A. L. Organometallics 2002, 21, 2136-2144.
(7) Kraft, B. M.; J ones, W. D. J . Organomet. Chem. 2002, 658, 132-
140.
(21) Deck, P. A.; J ackson, W. F.; Fronczek, F. R. Organometallics
1996, 15, 5287-5291.
(8) Kraft, B. M.; J ones, W. D. J . Am. Chem. Soc. 2002, 124, 8681-
8689.
(22) Deck, P. A.; Fronczek, F. R. Organometallics 2000, 19, 327-
333.
(9) Edelbach, B. L.; Rahman, A. K. F.; Lachicotte, R. J .; J ones, W.
D. Organometallics 1999, 18, 3170-3177.
(23) Thornberry, M. P.; Slebodnick, C.; Deck, P. A.; Fronczek, F. R.
Organometallics 2000, 19, 5352-5369.
(10) Edelbach, B. L.; J ones, W. D. J . Am. Chem. Soc. 1997, 119,
7734-7742.
(24) Deck, P. A.; Maiorana, C. R. Macromolecules 2001, 34, 9-13.
(25) Thornberry, M. P.; Slebodnick, C.; Deck, P. A.; Fronczek, F. R.
Organometallics 2001, 20, 920-926.
(11) Braun, T.; Perutz, R. N. Chem. Commun. 2002, 2749-2757.
(12) Godoy, F.; Higitt, C. L.; Klahn, A. H.; Oelckers, B.; Parsons,
S.; Perutz, R. N. J . Chem. Soc., Dalton Trans. 1999, 2039-2047.
(13) Cronin, L.; Higgitt, C. L.; Karch, R.; Perutz, R. N. Organome-
tallics 1997, 4920-4928.
(26) Waymouth, R. M.; Kravchenko, R. L.; Maciejewski; Petoff, J .
(Stanford University) US, 6169151, J an 2, 2001.
(27) Hauptman, E.; Waymouth, R. M.; Ziller, J . W. J . Am. Chem.
Soc. 1995, 117, 11586-11587.
(14) Rausch, M.; Bruhn, C.; Steinborn, D. J . Organomet. Chem.
2001, 622, 172-179.
(28) Bruce, M. D.; Coates, G. W.; Hauptman, E.; Waymouth, R. M.;
Ziller, J . W. J . Am. Chem. Soc. 1997, 119, 11174-11182.
(29) Lin, S.; Hauptman, E.; Lal, T. K.; Waymouth, R. M.; Quan, R.
W.; Ernst, A. B. J . Mol. Catal. A Chem. 1998, 136, 23-33.
(30) Morris, R. J .; Shaw, S. L.; J efferis, J . M.; Storhoff, J . J .; Goedde,
D. M. Inorg. Synth. 1998, 32, 215-221.
(15) Aizenberg, M.; Milstein, D. J . Am. Chem. Soc. 1995, 117, 8674-
8675.
(16) Schrock, R. R.; Adamchuk, J .; Ruhland, K.; Lopez, L. P. H.
Organometallics 2003, 22, 5079-5091.
(17) Cochran, F. V.; Bonitatebus, P. J .; Schrock, R. R. Organo-
metallics 2000, 19, 2414-2416.
(31) Diamond, G. M.; Rodewald, S.; J ordan, R. F. Organometallics
1995, 14, 5-7.
10.1021/om034385g CCC: $27.50 © 2004 American Chemical Society
Publication on Web 01/27/2004

1090 Organometallics, Vol. 23, No. 5, 2004
Deck et al.
F igu r e 1. Partial 1H NMR spectra (Cp and vinylic region)
for the reaction of 1a with 2 equiv of Ti(NMe2)4 in C6D6
recorded at the following intervals: (a) 15 min at 25 °C,
(b) 18 h at 25 °C, then 3 h at 50 °C, (c) 17 h at 50 °C, (d) 40
h at 65 °C, (e) 3 days at 80 °C.
F igu r e 2. Partial 1H NMR spectra (methyl region) for the
reaction of 1a with 2 equiv of Ti(NMe2)4 in C6D6 recorded
at the following intervals: (a) 15 min at 25 °C, (b) 18 h at
25 °C, then 3 h at 50 °C, (c) 17 h at 50 °C, (d) 40 h at 65
°C, (e) 3 days at 80 °C, (f) TsF + Ti(NMe2)4 in C6D6 after
18 h at 65 °C.
for the synthesis of indenyl complexes, our initial
attempts to isolate stable C6F5-substituted indenyl
complexes largely failed. However, in an amine elimina-
tion reaction of 3-(pentafluorophenyl)indene with Ti-
(NMe2)4, we observed the formation of products consis-
tent with intramolecular SNAr substitution of C-F at
the ortho positions of the C6F5 substituent.33-36 We
describe here our initial exploration of this unexpected
C-F activation pathway.
Resu lts a n d Discu ssion
Rea ction s of 3-(P en ta flu or op h en yl)in d en e (1a )
w ith Ti(NMe2)4. A solution of 1a and Ti(NMe2)4 (1:2
ratio) in benzene-d6 in a flame-sealed NMR tube was
examined using 1H and 19F NMR spectroscopy at several
1
intervals (Figures 1-3). The initial H NMR spectrum
(Figures 1a, 2a) was consistent with a mixture of
unreacted arylindene (1a ) and the corresponding η5-
indenyltitanium complex (2a ) along with unreacted Ti-
(NMe2)4 (singlet at 3.11 ppm) and free Me2NH (doublet
at 2.20 ppm, broad singlet at 0.1 ppm, 6:1 intensity
ratio). The initial 19F NMR spectrum (Figure 3a)
likewise showed three signals each for 1a (major) and
2a (minor). After 18 h at 25 °C, the ratio of 1a and 2a
was essentially unchanged, suggesting the initial equi-
librium shown in Scheme 1. After warming the same
sample at 50 °C for 3 h, a new indene product (4a ) was
identified by a slightly shifted vinyl signal (Figure 1b)
F igu r e 3. 19F NMR spectra (aromatic region) for the
reaction of 1a with 2 equiv of Ti(NMe2)4 in C6D6 recorded
at the following intervals: (a) 15 min at 25 °C, (b) 18 h at
25 °C, then 3 h at 50 °C, (c) 17 h at 50 °C, (d) 40 h at 65
°C, (e) 3 days at 80 °C.
NMR spectrum (Figure 3d) and the upfield-shifted
vinylic hydrogen signal (Figure 1d) clearly indicated
1
formation of indene 7a . The H NMR spectrum showed
1
and a new aryl-NMe2 signal (Figure 2b) in the H NMR
numerous signals in the upfield region (Figure 2d), some
of which appear exchange-broadened. The methylene
signals of 4a and 7a were not well resolved, but their
presence was confirmed using a 1D-TOCSY experiment
by irradiating the corresponding vinylic hydrogens at
6.1 ppm. Complete conversion to the diaminated aryl-
indene (7a ) was essentially complete after 3 days at 80
°C (Figures 1e, 2e, 3e).
spectrum, as well as four signals of mutually equal
intensity in the 19F NMR spectrum (Figure 3b). The
formation of 4a is nearly complete after an additional
17 h at 50 °C (Figures 1c, 3c). After an additional 40 h
at 65 °C, the triplet and doublet (1:2 ratio) in the 19F
(32) Diamond, G. M.; J ordan, R. F.; Petersen, J . L. J . Am. Chem.
Soc. 1996, 118, 8024-8033.
Whereas the 19F NMR spectra suggest a clean, highly
selective reaction with respect to the indenyl moiety,
the coordination environment about titanium in general
and the fate of the “activated” fluoride ions in particular
are unclear. Formally, removal of NMe2 groups from
titanium leaves vacant coordination sites for fluoride,
affording (Me2N)3TiF and (Me2N)2TiF2. Examination of
(33) O’Connor, P. E.; Berg, D. J .; Barclay, T. Organometallics 2002,
21, 3947-3954.
(34) Barrio, P.; Castarlenas, R.; Esterualas, M. A.; Lledo´s, A.;
Maseras, F.; OZate, E.; TomBs, J . Organometallics 2001, 20, 442-
452.
(35) Hughes, R. P.; Lindner, D. C.; Rheingold, A. L.; Yap, G. P. A.
Organometallics 1996, 15, 5678-5686.
(36) Bellabarba, R. M.; Nieuwenhuyzen, M.; Saunders, G. C. J .
Chem. Soc., Dalton Trans. 2001, 512-514.

Ti-Mediated C-F Activation Reactions
Organometallics, Vol. 23, No. 5, 2004 1091
Sch em e 1
F igu r e 4. Partial 1H NMR spectra (Cp region) for the
reaction of 1b with 2 equiv of Ti(NMe2)4 in C6D6 recorded
at the following intervals: (a) 15 min at 25 °C, (b) 17 h at
25 °C, (c) 3 h at 50 °C, (d) 24 h at 50 °C. (e) Ti(NMe2)4 +
7b (2:1) in C6D6.
dependent NMR spectra, suggesting extensive aggrega-
tion and ligand scrambling, presumably through bridg-
ing fluorides.38 Aggregation would probably be even
more likely for our less encumbered dimethylamido
analogues, whereas for the more encumbered TiF
moiety of [tBuN(3,5-Me2C6H3)]3TiF, a clear signal was
reported at -71.6 ppm.39
When arylindene (1a ) reacted with 2.0 equiv of
Ti(NMe2)4 on a preparative scale, the diaminated aryl-
indene (7a ) was obtained selectively in 71% crude yield
after hydrolytic workup. Likewise, a preparative-scale
reaction of arylindene (1a ) with 0.5 equiv of Ti(NMe2)4
gave the monoaminated arylindene (4a ) in 93% crude
yield as the major product, although subsequent chro-
matographic purification was rather inefficient. NMR
spectroscopic analyses of the pure substances (4a and
7a ) confirmed their assignment in the spectra of the
reaction mixtures (Figures 1-3). Two related negative
findings are noteworthy: In reactions conducted in
flame-sealed NMR tubes, arylindene (1a ) did not react
with either Ti(OiPr)4 (80 °C, benzene-d6, 18 h) or CpTi-
(NMe2)3 (80 °C, benzene-d6, 18 h).
the 19F NMR spectrum from +250 to -250 ppm showed
only the intense aryl CF groups and two weak, broad-
ened, unassigned complex doublets (J ≈ 200 Hz) around
0 ppm. In a small minority of duplicate experiments,
weak/broadened signals were sometimes observed, but
these signals would then disappear again after ad-
ditional reaction time. The additional 19F signals ac-
counted for only about 20% of the expected signal
intensity for the abstracted fluoride. Longer relaxation
delays (20 s) did not change integration data signifi-
cantly. The samples do slowly deposit a small amount
of a fine precipitate that is insoluble in common
solvents.
Rea ction s of (P en ta flu or op h en yl)cyclop en ta d i-
en e (1b) w ith Ti(NMe2)4. When the arylcyclopentadi-
ene (1b) was treated with 2.0 equiv of Ti(NMe2)4 in C6D6
1
at 25 °C, the initial H NMR spectrum (Figures 4a, 5a;
t < 15 min) showed that all of the diene (1b) had been
consumed to form 2b and 1 equiv of Me2NH cleanly.
Generally, monosubstituted Cp complexes show 1H
NMR splitting patterns that resemble two triplets
because the vicinal (3J HH) and distal (4J HH) coupling
constants are nearly the same (about 2 Hz).40 The signal
at 6.42 ppm in 2b showed a pentet instead, and we
assigned this additional splitting to long-range H-F
coupling to the two ortho fluorines of the C6F5 group
We tried to identify (Me2N)3TiF spectroscopically by
preparing it in situ from p-toluenesulfonyl fluoride (TsF)
with excess Ti(NMe2)4 in C6D6. Slow formation of N,N-
dimethyl-p-toluenesulfonamide37 was confirmed by com-
parison to authentic samples. We hypothesized that the
putative byproduct (Me2N)3TiF would meet the same
fate as in the reaction of 1a with Ti(NMe2)4, and indeed
1
the H NMR spectra of the two reactions (Figure 2e,f)
3
4
(5J HF ≈ J HH ≈ J HH ≈ 2 Hz). A similar observation was
reported for [η5-C5H4(C6F5)]2Fe.21 The accidental equiva-
lence of the three coupling constants provided a means
are strikingly similar in the region 2.5-3.7 ppm,
whereas neither 19F NMR spectrum shows signals in
the range +250 to -250 ppm that we could assign to
titanium fluorides. Furthermore the diethylamido ana-
logues (Et2N)3TiF and (Et2N)2TiF2 are known to exhibit
complex, broadened, concentration- and temperature-
1
to assign C-F activated species in the H NMR spec-
(38) Bu¨rger, H.; Wiegel, K. Z. Anorg. Allg. Chem. 1973, 398, 257-
272.
(39) Agapie, T.; Diaconescu, P. L.; Mindiola, D. J .; Cummins, C. C.
Organometallics 2002, 21, 1329-1340.
(40) Gamboa, J . A.; Sundararaman, A.; Kakalis, L.; Lough, A. J .;
J a¨kle, F. Organometallics 2002, 21, 4169-4181.
(37) Steinkopf, W.; Buchheim, K.; Beythien, K.; Dudek, H.; Eisold,
J .; Gall, J .; J aeger, P.; Beumuth, H.; Semenoff, A.; Wemme, A. J . Prakt.
Chem. 1927, 117, 1-82.

1092 Organometallics, Vol. 23, No. 5, 2004
Deck et al.
F igu r e 7. Partial 1H NMR spectra (Cp region) for the
reaction of 1b with 1 equiv of Ti(NMe2)4 in C6D6 recorded
at the following intervals: (a) 15 min at 25 °C, (b) 17 h at
25 °C, (c) 3 h at 50 °C, (d) 24 h at 50 °C. (e) Ti(NMe2)4 +
7b (2:1) in C6D6.
F igu r e 5. Partial 1H NMR spectra (methyl region) for the
reaction of 1b with 2 equiv of Ti(NMe2)4 in C6D6 recorded
at the following intervals: (a) 15 min at 25 °C, (b) 17 h at
25 °C, (c) 3 h at 50 °C, (d) 24 h at 50 °C. (e) Ti(NMe2)4 +
7b (2:1) in C6D6.
byproduct FTi(NMe2)3 probably does not have the same
chemical shift based on data obtained for the corre-
sponding Et2N derivatives.38 The presence of this signal,
its intensity (equal to that of 5b ), and its broadened
appearance may arise from aggregation and dispropor-
tionation of FTi(NMe2)3 as discussed earlier, but we
have no additional evidence to support that speculation.
As in the reactions of 1a , Ti-F signals were not
identified in any of the 19F spectra in the expected
range,21,41 and again we must concede that the fate of
the displaced fluoride remains unknown.
After an additional 3 h at 50 °C, all of the initial
complex (2b) has been consumed, and after 24 h at 50
°C, the intermediate 5b is nearly completely consumed
as well. The corresponding 19F NMR spectra (Figure
6c,d) show that the four-line pattern of 5b is gradually
replaced by two new doublets at about -146 ppm and
two new triplets at about -160 ppm. These observa-
tions, together with the appearance of a new triplet in
the downfield 1H NMR spectrum (Figure 4c), suggest
new complexes in which both ortho fluorines are re-
placed by Me2N groups. One of these was assigned by
treating an authentic sample of 7b with Ti(NMe2)4 to
afford 8b (Figures 4e, 5e, 6e). The other is presumably
intermediate 6b, although the extensive broadening of
the NMR signals (Figures 4d, 5d) suggests ligand-
scrambling at titanium on the NMR time scale.
When an essentially identical reaction is conducteds
except that a 1:1 ratio of 1b and Ti(NMe2)4 is usedsa
qualitatively analogous process occurs, except that the
chemical shifts of the major intermediates are different.
Complex 2b is still obtained cleanly (Figures 7a, 8, and
9a), and free Ti(NMe2)4 is consumed. However after 17
h at 25 °C, two quartets are observed in the downfield
1H NMR spectrum (Figure 7b) along with two four-line
systems in the 19F NMR spectrum (Figure 9b), perhaps
indicating a mixture of 3b and 5b. After 24 h at 50 °C,
a small amount of doubly activated species is indicated
by the triplet at 6.43 ppm in the downfield 1H NMR
spectrum (Figure 7d) and the two-line pattern in the
19F NMR spectrum (Figure 9d). This species is not
complex 8b (compare to Figures 7e and 9b) but corre-
F igu r e 6. 19F NMR spectra (aromatic region) for the
reaction of 1b with 2 equiv of Ti(NMe2)4 in C6D6 recorded
at the following intervals: (a) 15 min at 25 °C, (b) 17 h at
25 °C, (c) 3 h at 50 °C, (d) 24 h at 50 °C. (e) Ti(NMe2)4 +
7b (2:1) in C6D6.
trum as the reaction ensued. The 19F NMR spectrum of
2b (Figure 6a) shows the expected three signals.
After 17 h at 25 °C, analysis of the mixture using 19
F
NMR (Figure 6b) showed about 75% conversion to the
ortho-CF-activated intermediate 3b (four equally inte-
grating signals). In the downfield 1H NMR spectrum
(Figure 4b), the quartet at 6.42 ppm (Figure 4b) is
consistent with coupling of Cp hydrogens to only one
1
ortho fluorine. In the upfield H NMR spectrum (Figure
5b), a new doublet (J ) 2 Hz) appears at 2.45 ppm for
the aryl dimethylamino group coupled to the fluorine
at position 3. (Corresponding dd-septet coupling is
observed in the signal at -150 ppm in the 19F NMR
spectrum.) However compound 3b requires a 1:2 inte-
gration ratio for the aryl- and Ti-bound NMe2 groups,
respectively, whereas a 1:3 ratio is observed instead. We
propose that the reaction has proceeded by exchange
with unreacted Ti(NMe2)4 to give intermediate 5b
instead, possibly by bimolecular ligand exchange or via
the free diene 4b. In this scenario, the signal at 3.11
corresponding to Ti(NMe2)4 should be depleted, because
2 equiv of the starting titanium complex is formally
consumed. Instead the signal is broadened. The putative
(41) Kaminsky, W.; Lenk, S.; Scholz, V.; Roesky, H. W.; Herzog, A.
Macromolecules 1997, 30, 7647-7650.

Ti-Mediated C-F Activation Reactions
Organometallics, Vol. 23, No. 5, 2004 1093
ation of fluoroaryl-substituted tripodal amine ligands
to molybdenum17 and to hafnium.16 In the latter ex-
ample the hafnium source was Hf(NMe2)4, and the
product was a well-characterized quasi-octahedral hafni-
um monofluoride complex in which an erstwhile 2,6-
difluorophenyl group was converted to a 2-fluoro-6-
(dimethylamino)phenyl substituent. This result suggests
that our C-F activation reaction might also proceed
with less highly fluorinated aryl substituents.
F igu r e 8. Partial 1H NMR spectra (methyl region) for the
reaction of 1b with 1 equiv of Ti(NMe2)4 in C6D6 recorded
after 15 min at 25 °C.
The equilibrium between free and complexed indenes
(1a and 2a ) warrants more discussion. As the aminated
indenes (4a and 7a ) are formed, they are observed only
as free indenes (Figure 1, Scheme 1); indenyltitanium
complexes 3a , 5a , and 8a were not observed. The Me2N
groups might disrupt indenyl-titanium bonding by
changing the conformation of the aryl groups (see
“Structural Studies” below), although steric effects
should be relatively minor especially for the monoami-
nated indene (3). Indenes 4 and 7 are probably also less
acidic than 1. More acidic indenes should be more
reactive toward the basic Ti-NMe2 moiety. However,
variations in Ti-indenyl bond strengths would also need
to be considered in a complete thermodynamic treat-
ment, and one might expect indenyl anions bearing
electron-withdrawing substituents to form weaker Ti-
indenyl bonds. When the diarylated indene (10, 12.2 mg)
was treated with 1 equiv of Ti(NMe2)4 (15.4 mg) in
benzene-d6, we observed a qualitatively similar equi-
librium, even though 10 is about 7 pK units more acidic
than 1a in THF.44 Similarly, when unsubstituted in-
dene, which has a pK about 6 units higher than that of
1a ,44 was treated with 1 equiv of Ti(NMe2)4 at 80 °C for
F igu r e 9. 19F NMR spectra (aromatic region) for the
reaction of 1b with 1 equiv of Ti(NMe2)4 in C6D6 recorded
at the following intervals: (a) 15 min at 25 °C, (b) 17 h at
25 °C, (c) 3 h at 50 °C, (d) 24 h at 50 °C. (e) Ti(NMe2)4 +
7b (2:1) in C6D6.
1
sponds to the signals we tentatively assigned to inter-
mediate 6b above. An increase in the relative concen-
trations of fluorine-rich intermediates is consistent with
the change in reaction stoichiometry.
18 h in benzene-d6 in a sealed tube, H NMR spectro-
scopic analysis (at 25 °C) showed a mixture of uncom-
plexed indene and (Ind)Ti(NMe2)3. Martins and co-
workersreporteda high-yieldingsynthesisof(Ind)Ti(NMe2)3
from the same two reactants in toluene at 80 °C, but
their procedure called for a continuous stream of inert
gas to carry away the Me2NH byproduct.45 In contrast,
cyclopentadienyl ligands generally form stronger bonds
to transition metals than indenyl ligands. Thus, even
though the pK of 1b lies between that of the two
arylindenes 1a and 10,44 the free dienes (1b, 4b, 7b)
were not observed in our reactions with excess Ti-
(NMe2)4. Also, 9-(pentafluorophenyl)fluorene (11),46 which
is about 2 pK units less acidic than 1a ,47 does not react
at all with 2 equiv of Ti(NMe2)4 (C6D6, 80 °C, 18 h).
The ortho selectivity of the CF activation processes
suggests an intramolecular substitution pathway, per-
haps via an intermediate like 9. Reactions of 1a with
either excess LiNMe2 in THF or excess Me2NH in
aqueous THF resulted in para-selective substitutions
characterized by the appearance of two new equally
integrating signals in the 19F NMR spectrum. Bimo-
lecular reactions of C6F5-substituted ferrocenes with
“external” nucleophiles such as alkoxides and alkyl-
lithiums are also para-selective.15,42 Pentafluorotoluene
(CH3C6F5), which cannot bind titanium in the same
manner as 1a or 1b, did not react with Ti(NMe2)4
(toluene, 100 °C, 24 h).
Berg and co-workers observed a similar formal in-
tramolecular ortho CF substitution in zirconium(III)
aryldiamides, which they theorized might proceed by
oxidative addition to Zr(II) following disproportion-
ation.33 Redox chemistry is unlikely in our reactions.
More relevant here are the analogous intermolecular
SNAr pathway proposed43 for dehydrofluorination of
C6H5F by (Me5C5)2ZrH2 and the dehydrofluorinative
couplings of Me5C5 and fluoroaryl ligands observed in
some late transition metal complexes.35,36 The most
compelling predecent for our findings is the observation
of analogous fluoride/amide exchanges upon complex-
(42) Deck, P. A.; Lane, M. J .; Montgomery, J . L.; Slebodnick, C.;
Fronczek, F. R. Organometallics 2000, 19, 1013-1024.
(43) Kraft, B. M.; Lachiotte, R. J .; J ones, W. D. J . Am. Chem. Soc.
2001, 123, 10973-10979.

1094 Organometallics, Vol. 23, No. 5, 2004
Deck et al.
Ta ble 1. Cr ysta llogr a p h ic Da ta
4a 7b
17H13F4N
12
empirical formula
fw
C
C15H17F3N2
282.31
C
30H32F6FeN4
307.28
618.45
diffractometer
cryst dimens (mm)
cryst system
a (Å)
b (Å)
c (Å)
Oxford Diffraction X2
0.14 × 0.10 × 0.08
monoclinic
7.5668(8)
16.3012(15)
11.2873(13)
90
92.976(9)
90
1390.4(3)
P21/c
Oxford Diffraction X2
0.25 × 0.19 × 0.17
monoclinic
10.7977(18)
8.0970(13)
15.816(2)
90
94.422(12)
90
1378.6(4)
P21/n
Oxford Diffraction X2
0.18 × 0.18 × 0.12
monoclinic
6.0695(5)
34.001(2)
13.2317(12)
90
92.535(7)
90
2727.9(4)
P21/c
R (deg)
â (deg)
γ (deg)
V (Å3)
space group
Z
4
4
4
D
calc (Mg m-3
)
1.468
0.123
632
1.360
0.109
592
1.506
0.622
1280
abs coeff (mm-1
F000
)
radiation
λ (KR) (Å)
temp (K)
Mo KR
0.71073
100(2)
Mo KR
0.71073
100(2)
Mo KR
0.71073
100(2)
θ range for collection (deg)
no. of reflns colld
no. of indep reflns
abs corr method
2.97 to 27.60
8270
3213
3.35 to 30.07
15733
4032
3.08 to 27.50
18700
2513
none
none
none
no. of data/restrts/params
R [I > 2σ(I)]
3213/0/201
0.0439
0.1099
1.139
0.572, -0.662
4032/0/185
0.0494
0.1477
1.157
0.417, -0.372
6250/0/378
0.0463
0.1004
1.171
0.440, -0.368
Rw [I > 2σ(I)]
GoF on F2
largest diff peak, hole (e A-3
)
Sch em e 2
F igu r e 10. Thermal ellipsoid plot (50% probability) of 4a .
Methyl hydrogens omitted for clarity. Selected interatomic
distances (Å) and angles (deg): C1-C2, 1.507(3); C2-C3,
1.348(2); C3-C4, 1.481(2); C4-C9, 1.395(2); C9-C1,
1.508(3); C3-C31, 1.487(2); N-C36, 1.389(2); N-C37,
1.456(2); N-C38, 1.456(2); C36-N-C37, 120.96(14); C37-
N-C38, 115.82(15); C38-N-C36, 121.62(14).
Isola tion of th e (Dia m in oa r yl)cyclop en ta d ien e
(7b) a n d its Use a s a Liga n d . As shown in Scheme 2,
when the arylcyclopentadiene (1b) was treated with
excess Ti(NMe2)4 on a preparative scale, hydrolytic
workup gave 7b in 71% yield. Treatment of 7b with
sodium hydride followed by 0.5 equiv of FeBr2 in THF
gave the substituted ferrocene derivative (12) in 74%
yield as a deep red solid. The NMR spectrum of 12 in
CDCl3 at 25 °C shows free rotation of the aryl substi-
tutents (1 Me2N signal in 1H NMR spectrum and 2
sharp aryl CF signals in the 19F NMR spectrum).
Str u ctu r a l Stu d ies. Compounds 4a , 7b, and 12 were
subjected to single-crystal X-ray diffraction studies
(Table 1), and their solid-state molecular structures
(Figures 10-12) exhibited some noteworthy features. In
4a , the indene-aryl torsion angle is about 55°, whereas
in 7b and 12, the aryl substituents are approximately
perpendicular to their respective cyclopentadienyl moi-
eties, presumably because of the steric effects of the
Me2N groups. In contrast, 1,4-bis(pentafluorophenyl)-
cyclopentadiene, which lacks bulky ortho substituents,
is a nearly planar molecule.23 This strong conforma-
tional preference of the aryl groups in 12 apparently
precludes intramolecular arene stacking of the type
reported elsewhere for 1,1′-diarylferrocenes.23,25,48 The
packing diagram of 4a , which has only one Me2N group,
(44) Thornberry, M. P.; Warren, A. D.; Deck, P. A.; McKeown, A.
E.; Hasayan, F.; Streitwieser, A. Manuscript in preparation.
(45) Martins, A. M.; Ascenso, J . R.; de Azevedo, C. G.; Calhorda, M.
J .; Dias, A. R.; Rodrigues, S. S.; Toupet, L.; de Leonardis, P.; Veiros,
L. F. J . Chem. Soc., Dalton Trans. 2000, 4332-4338.
(46) Andrews, A. F.; Mackie, R. K.; Walton, J . C. J . Chem. Soc.,
Perkin Trans. 2 1980, 96-102.
(47) Bordwell, F. G.; Branca, J . C.; Bares, J . E.; Filler, R. J . Org.
Chem. 1988, 1988, 780-782.
(48) Blanchard, M. D.; Hughes, R. P.; Concolino, T. E.; Rheingold,
A. L. Chem. Mater. 2000, 12, 1604-1610.

Ti-Mediated C-F Activation Reactions
Organometallics, Vol. 23, No. 5, 2004 1095
lecular, nucleophilic substitution of the ortho C-F
groups by Ti-NMe2 groups to afford aminated arylin-
denes and arylcyclopentadienes. NMR-scale experi-
ments demonstrate the indermediacy of indenyltitanium
and cyclopentadienyltitanium compounds. Taken in
context of complementary work published by others, it
is clear that aryl CF bonds positioned in close proximity
to the inner coordination sphere of early transition
metal amides are subject to nucleophilic CF activation.
The scope and mechanism of these processes are under
continued investigation in our laboratories.
F igu r e 11. Thermal ellipsoid plot (50% probability) of 7b.
Methyl hydrogens omitted for clarity. Selected interatomic
distances (Å) and angles (deg): C1-C2, 1.357(2); C2-C3,
1.450(2); C3-C4, 1.376(2); C4-C5, 1.470(2); C5-C1,
1.458(2); C1-C11, 1.480(2); C2-C1-C5, 108.4(1); C2-C1-
C11, 128.4(1); C5-C1-C11, 128.2(1); C11-C12-N12,
119.2(1); C11-C16-N16, 118.6(1); C2-C1-C11-C12,
82.1(2).
Exp er im en ta l Section
Gen er a l P r oced u r es. All reactions were carried out using
standard inert-atmosphere techniques. NMR spectra were
recorded on Varian Unity-400 or Inova-400 instruments. 19F
NMR spectra were referenced externally to C6F6 in CDCl3 (δF
) -163.0 ppm). Elemental microanalyses were performed by
Desert Analytics (Tuscon AZ). Ti(NMe2)4 was used as received
from Aldrich. The arylindene (1a ) was prepared as previously
reported22 and purified by repeated fractional crystallization
from methanol followed by sublimation at 100 °C (0.1 Torr).
Indene was distilled at 20 mmHg. The monoarylated cyclo-
pentadiene (1b) was prepared by treating hexafluorobenzene
with sodium cyclopentadienide and sodium hydride at 0 °C in
THF following our “improved” procedure given on the last page
of the Supporting Information for ref 23. After hydrolytic
workup, the diene (1b) is obtained as a mixture of double-bond
regioisomers. The diene dimerizes slowly even when stored
in a freezer (-10 °C), so all samples of 1b used for these
experiments were freshly sublimed at 50 °C (0.1 Torr). When
highly pure, both 1a and 1b are white solids. Common solvents
were purified by published methods. Benzene-d6 was vacuum-
transferred from NaK alloy and stored in the glovebox.
NMR-Sca le Rea ction of 3-(P en ta flu or op h en yl)in d en e
(1a ) a n d Ti(NMe2)4. In a nitrogen glovebox, a solution of 1a
(9.4 mg, 33 µmol) and Ti(NMe2)4 (15.0 mg, 67 µmol) in benzene-
d6 (about 0.7 mL) was prepared in an NMR tube, which was
then flame-sealed. After initial NMR spectroscopic analysis,
the tube was allowed to stand for the following intervals and
temperatures and reanalyzed after each interval: 18 h at 25
°C, 3 h at 50 °C, 17 h at 50 °C, 10 h at 65 °C, 30 h at 65 °C, 12
h at 80 °C, and 60 h at 80 °C. Representative NMR data are
presented in Figures 1-3. A concurrent experiment in which
1a (18.8 mg, 66 µmol) was treated instead with 1 equiv of Ti-
(NMe2)4 (15.0 mg, 67 µmol) gave qualitatively similar results.
Components of the reaction mixture were assigned as fol-
lows: Unreacted Ti(NMe2)4: 1H NMR (C6D6) δ 3.11(s). Me2-
NH: 1H NMR (C6D6) δ 2.20 (d, 6 H), 0.1 (br s). Complex 2a :
1H NMR (C6D6) δ 7.26 and 7.00 (m, 2 H each, C6H4), 6.66 (m,
1 H2), 6.21 (dd, 1 H3), 2.80 (s, 18 H, Ti[NMe2]3); 19F NMR (C6D6)
δ -139.2 (d, 2 Fo), -158.2 (t, 1 Fp), -163.0 (m, 2 Fm). Data for
4a and 7a are provided below.
F igu r e 12. Thermal ellipsoid plot (50% probability) of 12.
Hydrogens omitted for clarity. Selected interatomic dis-
tances (Å) and angles (deg): Fe1-C1, 2.069(2); Fe1-C2,
2.051(2); Fe1-C3, 2.051(2); Fe1-C4, 2.040(2); Fe1-C5,
2.035(2); Fe1-C6, 2.074(2); Fe1-C7, 2.032(2); Fe1-C8,
2.035(2); Fe1-C9, 2.051(2); Fe1-C10, 2.051(2); C2-C1-
C11-C16, 34.3(3); C10-C6-C21-C26, -35.0(3).
reveals an arene stacking motif in which the Me2N
groups are in an anti arrangement. The distance
between arene centroids is 3.53 Å, and the orthogonal
interplanar distance is 3.27 Å, indicating a moderately
strong stacking interaction. The ferrocene (12) also
exhibits some strain, reflected in slightly longer Fe-C
distances to the ipso carbons of the (η5-C5H4R) groups,
a distortion not observed in (C5H4C6F5)2Fe.48 The Fe-N
distances (3.72 Å) are far too long to consider the Me2N
group to be a pendant donor group in this system. We
have considered the possibility that species such as 4b
might serve as “hybrid” ligands in a manner similar to
Saunders’s Cp-linked phosphines49,50 or Rothwell’s in-
denyl-linked phenoxides,51,52 but the fluoroaromatic
substituent probably renders the Me2N moiety a rather
unwilling electron pair donor.
NMR-Sca le Rea ction of (P en ta flu or op h en yl)cyclop en -
ta d ien e (1b) a n d Ti(NMe2)4. In a nitrogen glovebox, a
solution of 1b (15 mg, 67 µmol) and Ti(NMe2)4 (30 mg, 132
µmol) in benzene-d6 (about 0.7 mL) was prepared in an NMR
tube, which was then flame-sealed. After initial NMR analysis,
the tube was allowed to stand for the following intervals and
temperatures and reanalyzed after each interval: 17 h at 25
°C, 3 h at 50 °C, and 24 h at 50 °C. Representative NMR data
are presented in Figures 4-6. Spectra from a concurrent
experiment in which 1b (15 mg, 67 µmol) was treated instead
with 1 equiv of Ti(NMe2)4 (15 mg, 67 µmol) are shown in
Figures 7-9. Spectral assignments are presented in the
Results and Discussion.
Con clu sion s
Ti(NMe2)4 reacts with 3-(pentafluorophenyl)indene
and (pentafluorophenyl)cyclopentadiene via intramo-
(49) Ballabarba, R. M.; Saunders, G. C. J . Fluorine Chem. 2001,
112, 139-144.
(50) Bellabarba, R. M.; Niewenhuyzen, M.; Saunders, G. C. Orga-
nometallics 2002, 21, 5726-5737.
(51) Turner, L. E.; Thorn, M. G.; Fanwick, P. E.; Rothwell, I. P.
Chem. Commun. 2003, 1034-1035.
(52) Thorn, M. G.; Parker, J . R.; Fanwick, P. E.; Rothwell, I. P.
Organometallics 2003, 22, 4658-4664.

1096 Organometallics, Vol. 23, No. 5, 2004
Deck et al.
NMR-Sca le Rea ction of In d en e a n d Ti(NMe2)4. In a
nitrogen glovebox, a solution of indene (20 mg, 170 µmol) and
Ti(NMe2)4 (38.5 mg, 172 µmol) in benzene-d6 (about 0.7 mL)
was prepared in an NMR tube, which was then flame-sealed.
The tube was placed in an 80 °C oil bath for 18 h. 1H NMR
spectroscopic analysis showed unreacted indene and (Ind)Ti-
(NMe2)3 (verified by comparison with published data45) in a
ratio of about 4:1.
6 H), 1.91 (s, 3 H). These chemical shifts were confirmed by
comparison to an authentic sample of TsNMe2 prepared by
adding excess 40% aqueous dimethylamine to a solution of
tosyl chloride in benzene followed by aqueous workup.54 The
1H NMR spectrum also contained several unassigned signals
in the NMe2 region (see Figure 8), whereas the 19F NMR
spectrum showed only unreacted TsF at 66.7 ppm.
Syn th esis of 3-[2-(Dim eth yla m in o)tetr a flu or op h en yl]-
in d en e (4a ). A solution of 1a (1.13 g, 4.00 mmol), Ti(NMe2)4
(450 mg, 2.00 mmol), and toluene (50 mL) was stirred at 100
°C under a nitrogen atmosphere for 5 days and then cooled.
The volatile components were evaporated. Hexane (50 mL) and
water (20 mL) were added. The biphasic mixture was filtered
through Celite, and the filter was rinsed with 50 mL of hexane.
The biphasic filtrate was separated, and the aqueous layer was
washed with 50 mL of hexane. The organic layers were
combined, washed with saturated sodium bicarbonate solution
(2 × 20 mL), dried over anhydrous MgSO4, filtered, and
evaporated to afford 1.14 g (3.71 mmol, 93%) of a red oil that
crystallized upon standing. 1H NMR analysis showed the crude
product to be about 85% pure, with 7a as the primary
impurities. An analytical sample was obtained by flash chro-
matography on silica gel, eluting with hexane (Rf ) 0.24)
followed by sublimation (60-80 °C, 0.1 mmHg) to afford a pale
NMR-Sca le Rea ction of 1,3-Bis(p en ta flu or op h en yl)-
in d en e (10) a n d Ti(NMe2)4. In a nitrogen glovebox, a solution
of diarylindene22 (10, 13.5 mg, 30.1 mmol) and Ti(NMe2)4 (13.5
mg, 60.3 µmol) in benzene-d6 (about 0.7 mL) was prepared in
an NMR tube, which was then flame-sealed. Immediate 19F
NMR spectroscopic analysis (t < 15 min) showed unreacted
10 and [1,3-(C6F5)2Ind]Ti(NMe2)3 in a ratio of about 2:1. Data
for 10 compared well with published data obtained in CDCl3.22
Data for [1,3-(C6F5)2Ind]Ti(NMe2)3: 19F NMR (C6D6) δ -142.3
(d, 2 F), -157.3 (t, 2 F), -162.6 (m, 2 F). 1H NMR assignments
were unclear because of apparent exchange broadening. After
several hours at room temperature a C-F activation process
analogous to the reaction of 1a and Ti(NMe2)4 was evident from
19F NMR spectroscopic analysis but was not pursued further.
NMR-Sca le Rea ction of Ar ylcyclop en ta d ien e (7b) w ith
Ti(NMe2)4. In a nitrogen glovebox, a solution of arylindene
(7b, 9.3 mg, 33 µmol) and Ti(NMe2)4 (14.8 mg, 66 µmol) in
benzene-d6 (about 0.7 mL) was prepared in an NMR tube,
which was then flame-sealed. After 4 h at 25 °C, NMR analysis
showed dimethylamine, unreacted Ti(NMe2)4, and complex 8b.
Data for 8b: 1H NMR (C6D6) δ 6.45 (t, J ) 2 Hz, 2 H), 5.90 (t,
3
yellow solid: 1H NMR (C6D6) δ 7.28 (d, J ) 8 Hz, 1 H), 7.20
3
3
3
(t, J ) 8 Hz,1 H), 7.15 (t, J ) 8 Hz, 1 H), 6.97 (d, J ) 8 Hz,
3
3
1 H), 6.18 (t, J ) 2 Hz, 1 H), 3.12 (d, J ) 2 Hz, 2 H), 2.32 (d,
5J FH ) 2 Hz, 6 H); 19F NMR (C6D6) δ -140.7 (dd, J ) 24 Hz,
3
5J ) 9 Hz, 1 F, F6), -150.9 (dd of sept, J ) 20 Hz, J ) 9 Hz,
5J HF ) 2 Hz, F3), -158.1 (td, 3J ) 21 Hz, 4J ) 2 Hz, F4), -164.5
(dd, 3J ) 24 Hz, 3J ) 21 Hz, F5). Anal. Calcd (found) for
3
5
J ) 2 Hz, 2 H), 3.06 (s, 18 H), 2.50 (d, J FH ) 1.6 Hz, 6 H); 19
F
)
NMR (C6D6) δ -146.4 (d, 3J FF ) 20 Hz, 2 F), -160.7 (d, 3J FF
C
17H13F4N: C, 66.45 (66.27); H, 4.26 (4.17); N, 4.56 (4.69).
20 Hz, 1 F).
NMR-Sca le Rea ction of 9-(P en ta flu or op h en yl)flu o-
r en e (11) w ith Ti(NMe2)4. In a nitrogen glovebox, a solution
of 9-(pentafluorophenyl)fluorene47 (10.8 mg, 32 µmol) and Ti-
(NMe2)4 (14.6 mg, 65 µmol) in benzene-d6 (about 0.7 mL) was
prepared in an NMR tube, which was then flame-sealed. After
4 h at 25 °C, NMR spectroscopic analysis showed only
unreacted starting materials. Data for 11: 1H NMR (C6D6) δ
7.57 (d, 2 H), 7.20 (d, 2 H), 7.08 (d, 2 H), 7.00 (d, 2 H), 5.09 (s,
1 H); 19F NMR (C6D6) δ -140.0 (br s, 1 F), -145.4 (br s, 1 F),
Syn t h esis of 3-[2,6-Bis(d im et h yla m in o)t r iflu or o-
p h en yl]in d en e (7a ). A solution of 1a (564 mg, 2.00 mmol),
Ti(NMe2)4 (900 mg, 4.00 mmol), and toluene (50 mL) was
stirred at 100 °C under a nitrogen atmosphere for 5 days and
then cooled. The volatile components were evaporated. Hexane
(50 mL) and water (20 mL) were added. The dark biphasic
mixture was filtered through Celite, and the filter was rinsed
with 50 mL of hexane. The biphasic filtrate was separated,
and the aqueous layer was washed with 50 mL of hexane. The
organic layers were combined, washed with saturated sodium
bicarbonate solution (2 × 20 mL), dried over anhydrous
MgSO4, filtered, and evaporated to afford 471 mg (1.42 mmol,
71%) of a yellow-orange crystalline solid, which was found to
be about 90% pure by 1H NMR spectroscopic analysis. An
analytical sample was obtained by flash chromatography on
silica gel, eluting with hexane (Rf ) 0.11) followed by sublima-
3
-156.0 (t, J FF ) 21 Hz, 1 H), -161.7 (br s, 1 F), -162.6 (br s,
1 F).
NMR-Sca le Rea ction of 3-(P en ta flu or op h en yl)in d en e
w ith Cp Ti(NMe2)3. In a nitrogen glovebox, a solution of 1a
53
(20 mg, 76 µmol) and CpTi(NMe2)3 in benzene-d6 (about 1
mL) was prepared in an NMR tube, which was then flame-
sealed. After immersion in an oil bath for 18 h at 80 °C,
analysis by 1H and 19F NMR spectroscopy showed only
unreacted starting materials.
3
tion (90 °C, 0.1 mm): 1H NMR (C6D6) δ 7.30 (d, J ) 8 Hz, 1
3
3
H), 7.22 (t, J ) 8 Hz, 1 H), 7.15 (t, J ) 8 Hz, 1 H), 6.90 (d,
3J ) 8 Hz, 1 H), 6.07 (t, J ) 2 Hz, 1 H), 3.24 (d, J ) 2 Hz, 2
H), 2.44 (d, 5J FH ) 1.5 Hz, 12 H); 19F NMR (C6D6) δ -148.0 (d,
3J ) 21 Hz, 2 F), -161.2 (t, 3J ) 21 Hz, 1 F). Anal. Calcd
(found) for C19H19F3N2: C, 68.66 (68.38); H, 5.76 (5.40), N 8.43
(8.30).
3
3
NMR-Sca le Rea ction of 3-(P en ta flu or op h en yl)in d en e
w ith Ti(OiP r )4. In a nitrogen glovebox, a solution of 1a (about
20 mg) and Ti(OiPr)4 (about 50 mg) in toluene-d8 (about 1 mL)
was prepared in a J -Young NMR tube. After immersion in an
oil bath for 18 h at 80 °C, analysis by 1H and 19F NMR
spectroscopy showed only unreacted starting materials and a
Syn th esis of [2,6-Bis(d im eth yla m in o)tr iflu or op h en yl]-
cyclop en ta d ien e (7b). In a Teflon-valved glass tube, a
solution of freshly sublimed (pentafluorophenyl)cyclopenta-
diene (1.15 g, 4.96 mmol) and Ti(NMe2)4 (1.11 g, 4.96 mmol)
in toluene (2 mL) was sealed and stirred at 110 °C for 15 h.
After cooling, the resulting solution was added to a rapidly
stirred mixture of ice-cold water (100 mL) and hexane (50 mL).
The resulting mixture was filtered through Celite. The bipha-
sic filtrate was separated, and the organic phase was washed
with water, dried over anhydrous MgSO4, filtered, and evapo-
rated to afford 1.02 g (3.61 mmol, 73%) of a dark solid. An
analytical sample was obtained by chromatography on silica
gel, eluting with hexanes, followed by sublimation (0.1 mmHg,
i
small amount of PrOH from the hydrolysis of Ti(OiPr)4 by
adventitious moisture.
NMR-Sca le Rea ction of p-Tolu en esu lfon yl F lu or id e
(TsF ) w ith Ti(NMe2)4. In a nitrogen glovebox, a solution of
TsF (8.7 mg, 50 µmol) and Ti(NMe2)4 (22.7 mg, 100 µmol) in
C6D6 (about 0.8 mL) was prepared in an NMR tube, which was
then flame-sealed. Initial 1H and 19F NMR spectra showed only
unreacted starting materials. Data for TsF: 1H NMR (C6D6)
δ 7.51 (d, 2 H), 6.50 (d, 2 H), 1.72 (s, 3 H); 19F NMR (C6D6) δ
+66.7 (s). After 18 h at 65 °C, the sample contained about 25%
of N,N-dimethyl p-toluenesulfonamide (TsNMe2). Data for
TsNMe2: 1H NMR (C6D6) δ 7.59 (d, 2 H), 6.78 (d, 2 H), 2.26 (s,
(53) Bu¨rger, H.; Da¨mmgen, U. J . Organomet. Chem. 1975, 101, 295-
306.
(54) Briscoe, P. A.; Challenger, F.; Duckworth, P. S. J . Chem. Soc.
1956, 1755-1768.

Ti-Mediated C-F Activation Reactions
Organometallics, Vol. 23, No. 5, 2004 1097
80 °C) to obtain a pale yellow solid. The product was found to
be a nearly equimolar mixture of double-bond regioisomers.
Data for 1-[2,6-C6(Me2N)2F3]C5H5: 1H NMR (CDCl3) δ 6.62 (m,
crystallized from hexanes at -10 °C. Crystals were cut prior
to mounting on nylon CryoLoops (Hampton Research) with
Krytox Oil (DuPont) and centering on the goniometer of an
Oxford Diffraction XCalibur2 diffractometer equipped with a
Sapphire 2 CCD detector. The data collection routines, unit
cell refinements, and data processing were carried out with
the program CrysAlis.55 Structure solution (direct methods),
refinement, and molecular graphics generation used the
SHELXTL NT program package.56 The Laue symmetry and
systematic absences were consistent with the monoclinic space
group P21/n for indene 4a , P21/n for diene 7b, and P21/c for
ferrocene 12. The final refinement models for both structures
involved anisotropic displacement parameters for non-hydro-
gen atoms and a riding model for all hydrogen atoms.
5
1 H), 6.54 (m, 1 H), 6.38 (m, 1 H), 3.36 (m, 2 H), 2.65 (d, J FH
) 2 Hz, 12 H); 19F NMR (CDCl3) δ -147.6 (d, 3J ) 21 Hz, 2 F)
-161.0 (t, 3J ) 21 Hz, 1 F). Data for 2-[2,6-C6(Me2N)2F3]C5H5:
1H NMR (CDCl3) δ 6.57 (m, 1 H), 6.46 (m, 1 H), 6.38 (m, 1 H),
5
3.18 (m, 2 H), 2.64 (d, J FH ) 2 Hz, 12 H); 19F NMR (C6D6) δ
3
3
-148.4 (d, J ) 21 Hz, 2 F), -161.1 (t, J ) 21 Hz, 1 F). Anal.
Calcd (found) for C15H17F3N2: C, 64.08 (63.80); H, 6.09 (5.99);
N, 9.96 (9.91).
1,1′-Bis(2,6-bis(d im eth yla m in o)tr iflu or op h en yl)fer r o-
cen e (12). A mixture of arylcyclopentadiene (7b, 564 mg, 2.00
mmol), sodium hydride (96 mg, 4.00 mmol), and THF was
stirred under nitrogen for 15 h. To the resulting purple solution
was added iron dibromide (216 mg, 1.00 mmol) in one portion.
The resulting mixture was stirred at 65 °C for 4 h and then
cooled. The solvent was evaporated, and the mixture was
hydrolyzed by adding toluene (50 mL) and water (40 mL) with
stirring. After filtration through Celite, the biphasic mixture
was separated, dried over anhydrous MgSO4, filtered, and
evaporated to afford 460 mg (0.74 mmol, 74%) of a red solid.
An analytical sample was obtained by recrystallization from
hexane: 1H NMR (CDCl3) δ 5.11 (t, 3J ) 4J ) 2 Hz, 2 H), 4.18
Ack n ow led gm en t. This work was supported by the
National Science Foundation (CHE-9875446). We also
thank NSF for funding the purchase of the Oxford
Diffraction Xcalibur2 single crystal diffractometer (CHE-
0131128) and the Varian Inova 400 MHz NMR spec-
trometer (CHE-0131124).
Su p p or tin g In for m a tion Ava ila ble: X-ray crystallo-
graphic data, including thermal ellipsoid plots of 4a , 7a , and
12, and a partial packing diagram of 4a showing the arene
stacking motif (PDF). This material is available free of charge
via the Internet at http://pubs.acs.org.
4
(t, 3J ) J ) 2 Hz, 2 H), 2.72 (s, 24 H); 19F NMR (CDCl3) δ
3
3
-146.2 (d, J ) 21 Hz, 4 F), -162.5 (t, J ) 21 Hz, 2 F). Anal.
Calcd (found): C, 58.26 (58.65); H, 5.22 (5.40); N, 9.06 (8.85).
Cr ysta llogr a p h ic Stu d ies. Crystal data are presented in
Table 1. Colorless prisms of 4a were obtained by sublimation
at ambient conditions over an 18-month period. Colorless
plates of substituted cyclopentadiene 7b were crystallized from
ethanol at 25 °C. Red needles of substituted ferrocene 12 were
OM034385G
(55) CrysAlis v1.170; Oxford Diffraction: Wrocław, Poland, 2002.
(56) Sheldrick, G. M. SHELXTL NT v 6.12; 2001.