Structural and electronic effects of pentafluorophenyl substituents on cyclopentadienyl complexes of Fe, Co, Mn, and Re

5352

Organometallics 2000, 19, 5352-5369

Str u ctu r a l a n d Electr on ic Effects of P en ta flu or op h en yl

Su bstitu en ts on Cyclop en ta d ien yl Com p lexes of F e, Co,

Mn , a n d Re¹

Matthew P. Thornberry, Carla Slebodnick, and Paul A. Deck*

Department of Chemistry, Virginia Tech, Blacksburg, Virginia 24061-0212

Frank R. Fronczek

Department of Chemistry, Louisiana State University, Baton Rouge, Louisiana 70803-1804

Received September 18, 2000

The reaction of CpNa (Cp ) C₅H₅) with C₆F₆(diglyme, 110 °C, 14 h) afforded a mixture of

products, from which two new triarylated cyclopentadienes, 1,2,4-Ar₃C₅H₃(3) and 1,2,3-

Ar₃C₅H₃(4), were separated by silica gel chromatography (Ar ) C₆F₅). The dienes 3 and 4

were readily converted (NaH, THF) to the corresponding triarylated sodium cyclopentadi-

enides, (1,2,4-Ar₃C₅H₂)Na (7) and (1,2,3-Ar₃C₅H₂)Na (8). Reactions of the ligands 7 and 8

with FeBr₂afforded the ferrocenes (1,2,4-Ar₃C₅H₂)₂Fe (11) and (1,2,3-Ar₃C₅H₂)₂Fe (12),

respectively. The cobaltocenes (ArC₅H₄)₂Co (13), (1,3-Ar₂C₅H₃)₂Co (14), (1,2,4-Ar₃C₅H₂)₂Co

(15), and (1,2,3-Ar₃C₅H₂)₂Co (16) were prepared from CoBr₂and (ArC₅H₄)Na (5), (1,3-

Ar₂C₅H₃)Na (6), 7, and 8, respectively. Oxidation of 13 (air, dilute aqueous HCl) followed by

-

precipitation with KPF₆afforded [(ArC₅H₄)₂Co]⁺PF₆(17). The homologous cobaltocenium

ions [1,3-Ar₂C₅H₃]₂Co⁺(18), (1,2,4- Ar₃C₅H₂)₂Co⁺(19), and (1,2,3-Ar₃C₅H₂)₂Co⁺(20) were

observed (NMR) only after dissolving the corresponding Co^IIspecies 14-16 in D₂SO₄; efforts

to isolate 18-20 as hexafluorophosphates failed. Reactions of the ligands 5-8 with M(CO)₅Br

(M ) Mn, Re) afforded the piano stool complexes (ArCp)Mn(CO)₃(21), (1,3-Ar₂Cp)Mn(CO)₃

(22), (1,2,4-Ar₃C₅H₂)M(CO)₃(23, M ) Mn; 27, M ) Re), and (1,2,3-Ar₃C₅H₂)M(CO)₃(24, M

) Mn; 28, M ) Re). Both the voltammetric oxidation potentials (E_1/2) of the substituted

metallocenes 9-16 and the carbonyl stretching wavenumbers of the substituted piano stool

complexes 21-24 and 25-28 increased linearly with increasing number of C₆F₅substituents.

Crystal structures were obtained for 2 (two polymorphs), 4, 11, 16, 23, 26, and 27. Noteworthy

features of these structures include C₆F₅-C₆F₅arene stacking, C₆F₅stacking with benzene

solvate molecules, C₆F₅‚‚‚OtC-M interactions, and extensive C-H‚‚‚F-C interactions.

Barriers toward Cp-C₆F₅rotation and Cp-Fe-Cp rotation were quantified for several of

1

the complexes using dynamic ¹⁹F and H NMR spectroscopy.

In tr od u ction

other ligands (L, X) and bonds (M-L, M-X) in the inner

coordination sphere,⁶which is particularly significant

for catalysis.^7,8While the size and shape (steric effects)

of Cp substituents can control the orientation of sub-

strate molecules approaching the metal, thereby opti-

mizing regio- and stereoselectivity, electronic effects

influence the degree to which a ligated metal atom can

present orbitals having appropriate energy and spatial

distribution to interact productively with organic sub-

strates.⁹

Cyclopentadienyl (Cp) complexes (e.g., A, B, and C)

are known for nearly all metals, including several main

group metals²and f-elements.³Ancillary ligand sub-

stituents (R) influence their molecular structures, physi-

cal properties,⁴and electrochemical and spectroscopic

features.⁵Substituents also modulate the reactivity of

(1) Parts of this work have been communicated orally: (a) Deck, P.

A.; Thornberry, M. P. Southeast Regional Meeting of the American

Chemical Society, Knoxville TN, October 17-20, 1999, Abstract 314.

(b) Deck, P. A.; Thornberry, M. P. 215th National Meeting of the

American Chemical Society, Dallas, TX, March 29-April 2, 1998,

Abstract INOR 73. (c) Deck, P. A.; Thornberry, M. P. 218th National

Meeting of the American Chemical Society, New Orleans, LA, August

22-26, 1999, Abstract INOR 525.

(2) (a) Overby, J . S.; Hanusa, T. P.; Boyle, P. D. Angew. Chem., Int.

Ed. Engl. 1997, 36, 2378. (b) Burkey, D. J .; Hanusa, T. P. J .

Organomet. Chem. 1996, 512, 165. (c) Overby, J . S.; Hanusa, T. P.

Organometallics 1996, 15, 2205. (d) Burkey, D. J .; Hanusa, T. P.

Commun. Inorg. Chem. 1995, 17, 41. (e) J utzi, P.; Burford, N. Chem.

Rev. 1999, 99, 969. (f) Harder, S. Coord. Chem. Rev. 1998, 176, 17. (g)

Bridgeman, A. J . J . Chem. Soc., Dalton Trans. 1997, 17, 2887. (h)

Sitzmann, H.; Dezember, T.; Ruck, M. Angew. Chem., Int. Ed. 1998,

37, 3114.

Nevertheless, the study of electronic effects has lagged

behind the study of steric effects.¹⁰One reason is the

(3) (a) Evans, W. J .; Nyce, G. W.; Clark, R. D.; Doedens, R. J .; Ziller,

J . W. Angew. Chem., Int. Ed. 1999, 38, 1801. (b) Arredondo, V. M.;

McDonald, F. E.; Marks, T. J . Organometallics 1999, 18, 1949. (c)

Klooster, W. T.; Brammer, L.; Schaverien, C. J .; Budzelaar, P. H. M.

J . Am. Chem. Soc. 1999, 121, 1381. (d) Gunko, Y. K.; Edelmann, F. T.

Comm. Inorg. Chem. 1997, 19, 153. (e) Molander, G. A.; Corrette, C.

P. J . Org. Chem. 1999, 64, 9697.

(4) (a) Overby, J . S.; Hanusa, T. P.; Sellers, S. P.; Yee, G. T.

Organometallics 1999, 18, 3561. (b) Hays, M. L.; Burkey, D. J .; Overby,

J . S.; Hanusa, T. P.; Sellers, S. P.; Yee, G. T.; Young, V. G. Organo-

metallics 1998, 17, 5521.

Publication on Web 11/14/2000

Cyclopentadienyl Complexes of Fe, Co, Mn, and Re

Organometallics, Vol. 19, No. 25, 2000 5353

sodium salts, and they can be used in standard synthe-

ses of various complexes of both early and late transition

metals.¹⁶Although the fluoro (-F) substituent is also

stable and highly electron-withdrawing, the synthesis

of ring-fluorinated metallocene complexes has not yet

been generalized beyond a few systems.¹⁷After our

report, Hughes and co-workers showed that C₆F₅groups

are useful “crystal engineering” design elements,¹⁶while

Rausch and co-workers showed that C₆F₅substituents

have significant effects on the polymerization of olefins

by titanium Cp complexes.¹⁸Complementary studies at

the electron-donating extreme have also appeared.¹⁹

This contribution reports the electron-withdrawing

substituent effects that we have observed in several

complexes bearing one to three C₆F₅substituents at-

tached to Cp ligands. First, we describe the syntheses

of two regioisomeric tris(pentafluorophenyl)cyclopen-

tadienes, which extend the series of mono- and bis-

(pentafluorophenyl)cyclopentadienes we reported ear-

lier.¹³Then, we show how Cp ligands with 1-3 C₆F₅

groups can be used as ligands for Fe, Co, Mn, and Re.

Electrochemical and spectroscopic data are presented

to quantify the electronic substituent effects in these

complexes. “Proof of concept” experiments using cobal-

tocene illustrate the potential of C₆F₅substituents to

stabilize 19-electron organometallic complexes. We then

present the results of crystallographic studies that

demonstrate the influence of C₆F₅groups on molecular

structure and crystal packing. Finally, we show how

C₆F₅groups affect static and dynamic aspects of met-

allocene conformations.

synthetic investment required to obtain a homologous

series of complexes in which electronic effects vary

widely.¹¹Highly reactive substituents can require exist-

ing synthetic routes to be redesigned. For this reason,

expected behavior of Cp complexes at the electron-

donating or electron-withdrawing extremes is often

inferred from trends observed over a narrow range of

weakly electron-donating substituents such as CH₃or

SiMe₃, which are easy to attach and are usually inert

toward the reagents and reaction conditions necessary

for synthesis and catalytic screening.¹²

Our goal is to probe the electron-withdrawing extreme

of substituent effects in Cp complexes using perfluoro-

aryl substituents. We showed earlier that one or two

pentafluorophenyl (C₆F₅) groups are readily attached to

cyclopentadiene and indene by nucleophilic aromatic

substitution.^13,14Unlike their perfluoroalkyl-substituted

congeners,¹⁵C₆F₅-substituted Cp anions are stable as

(5) (a) Blais, M. S.; Rausch, M. D. Organometallics 1994, 13, 3557.

(b) Graham, P. B.; Rausch, M. D.; Taschler, K.; Von Philipsborn, W.

Organometallics 1991, 10, 3049. (c) Meier, E. J . M.; Kozminski. W.;

Linden, A.; Lustenberger, P.; von Philipsborn, W. Organometallics

1996, 15, 2469. (d) von Philipsborn, W. Chem. Soc. Rev. 1999, 28, 95.

(e) Langmaier, J .; Samec, Z.; Varga, V.; Horacek, M.; Choukroun, R.;

Mach, K. J . Organomet. Chem. 1999, 584, 323. (f) Barthel-Rosa, L. P.;

Sowa, J . R.; Gassman, P. G.; Fischer, J .; McCarty, B. M.; Goldsmith,

S. L.; Gibson, M. T.; Nelson, J . H. Organometallics 1997, 16, 1595. (g)

Gassman, P. G.; Sowa, J . R.; Hill, M. G.; Mann, K. R. Organometallics

1995, 14, 4879. (h) Gassman, P. G.; Deck, P. A.; Winter, C. H.; Dobbs,

D. A.; Cao, D. H. Organometallics 1992, 11, 959. (i) Bu¨hl, M.

Organometallics 1997, 16, 261.

(6) (a) Cheong, M.; Basolo, F. Organometallics 1988, 7, 2041. (b)

Wang, D. M.; Angelici, R. J . Inorg. Chem. 1996, 35, 1321. (c) Nataro,

C.; Thomas, L. M.; Angelici, R. J . Inorg. Chem. 1997, 36, 6000.

(7) (a) Witte, P.; Lal, T. K.; Waymouth, R. M. Organometallics 1999,

18, 4147. (b) Tagge, C. D.; Kravchenko, R. L.; Lal, T. K.; Waymouth,

R. M. Organometallics 1999, 18, 380.

(8) (a) Batsanov, A. S.; Bridgewater, B. M.; Wilson, C. J . Organomet.

Chem. 1999, 590, 169. (b) Halterman, R. L. Chem. Rev. 1992, 92, 965.

(c) Mo¨hring, P. C.; Coville, N. J . J . Organomet. Chem. 1994, 479, 1.

(d) Hoveyda, A. H.; Morken, J . P. Angew. Chem., Int. Ed. Engl. 1996,

35, 1262.

(9) (a) Torrent, M.; Sola, M.; Frenking, G. Chem. Rev. 2000, 100,

439-493. (b) Gladysz, J . A.; Boone, B. J . Angew. Chem., Int. Ed. Engl.

1997, 36, 551-583. (c) Doucet, H.; Fernandez, E.; Layzell, T. P.; Brown,

J . M. Chem. Eur. J . 1999, 5, 1320-1330.

(10) (a) Richardson, D. E.; Alameddin, N. G.; Ryan, M. F.; Hayes,

T.; Eyler, J . R.; Siedle, A. R. J . Am. Chem. Soc. 1996, 118, 11244. (b)

Alameddin, N. G.; Ryan, M. F.; Eyler, J . R.; Siedle, A. R.; Richardson,

D. E. Organometallics 1995, 14, 5005. (c) Richardson, D. E. Elect.

Trans. React. 1997, 253, 79. (d) Ryan, M. F.; Siedle, A. R.; Burk, M.

J .; Richardson, D. E. Organometallics 1992, 11, 4231. (e) Kucht, A.;

Kucht, H.; Song, W.; Rausch, M. D.; Chien, J . C. W. Appl. Organomet.

Chem. 1994, 8, 437. (f) Schnabel, R. C.; Scott, B. L.; Smith, W. H.;

Burns, C. J . J . Organomet. Chem. 1999, 591, 14. (g) Bursten, B. E.;

Green, M. R. Prog. Inorg. Chem. 1988, 36, 393. (h) Slocum, D. W.;

Ernst, C. R. Adv. Organomet. Chem. 1972, 10, 79. (i) Senoff, C. V.

Coord. Chem. Rev. 1980, 32, 111. (j) King, R. B. Coord. Chem. Rev.

1976, 20, 155.

Exp er im en ta l Section

Gen er a l P r oced u r es. Standard inert-atmosphere tech-

niques were used for all reactions. C₆F₆(99% purity) was

received as a gift from Albemarle Corporation and used

without further purification. FeBr₂, Mn(CO)₅Br, and Re-

(CO)₅Br were used as received from Aldrich. NaH was pur-

chased as a 60% mineral oil dispersion from Aldrich, washed

with hexanes, dried under vacuum, and stored in a glovebox.

Anhydrous CoBr₂was prepared by heating the red commercial

hydrate (Fisher) under vacuum to obtain a bright green

powder. NaCp was prepared from freshly distilled cyclopen-

tadiene and excess NaH in THF. Two ligands, 5 and 6, were

prepared as previously reported.¹³NMR experiments used

J EOL Eclipse (500), Varian Unity (400), and Bruker EM (360)

instruments. THF-d₈was vacuum-transferred from Na/K alloy.

¹⁹F NMR spectra were referenced to external C₆F₆in CDCl₃

at -163.00 ppm. {¹⁹F}¹³C NMR spectra were particularly

useful for resolving signals in the aromatic C-F regions,

particularly where extensive ¹⁹F couplings in the ¹H-decoupled

spectra prevented us from distinguishing closely spaced sig-

nals. C-F coupling constants were approximated from appar-

ent splittings. Some of the complexes were insufficiently

soluble for ¹³C NMR analysis. Infrared spectra were recorded

on a Midac M-series instrument operating at 1 cm^-1resolution.

Mass spectra were obtained using a Fisons VG7070 instru-

ment. Elemental analyses were performed by Desert Analytics

(Tucson, AZ).

(11) (a) Blais, M. S.; Rausch, M. D. J . Organomet. Chem. 1995, 502,

1. (b) Rausch, M. D.; Lewison, J . F.; Hart, W. P. J . Organomet. Chem.

1988, 358, 161. (c) Hart, W. P.; Macomber, D. W.; Rausch, M. D. J .

Am. Chem. Soc. 1980, 102, 1196.

(16) Blanchard, M. D.; Hughes, R. P.; Concolino, T. E.; Rheingold,

A. L. Chem. Mater. 2000, 12, 1604.

(12) Mo¨hring, P. C.; Coville, N. J . J . Mol. Catal. A 1995, 96, 181.

(13) Deck, P. A.; J ackson, W. F.; Fronczek, F. R. Organometallics

1996, 15, 5287.

(14) Deck, P. A.; Fronczek, F. R. Organometallics 2000, 19, 327.

(15) (a) Baschky, M. C.; Sowa, J . R.; Gassman, P. G.; Kass, S. R. J .

Chem. Soc., Perkin Trans. 2 1996, 213. (b) Burk, M. J .; Arduengo, A.

J .; Calabrese, J . C.; Harlow, R. L. J . Am. Chem. Soc. 1989, 111, 8938.

(17) Richardson, D. E.; Lang, L.; Eyler, J . R.; Kircus, S. R.; Zheng,

X. M.; Morse, C. A.; Hughes, R. P. Organometallics 1997, 16, 149.

(18) Maldanis, R. J .; Chien, J . C. W.; Rausch, M. D. J . Organomet.

Chem. 2000, 599, 107.

(19) (a) Lehmus, P.; Kokko, E.; Ha¨rkki, O.; Leino. R.; Luttikhedde,

H. J . G.; Nasman, J . H.; Seppa¨la¨, J . V. Macromolecules 1999, 32, 3547.

(b) Plenio, H.; Burth, D. Angew. Chem., Int. Ed. Engl. 1995, 34, 800.

5354 Organometallics, Vol. 19, No. 25, 2000

Thornberry et al.

Electr och em ica l Stu d ies. Single-sweep cyclic and Ostery-

oung square-wave voltammetric data were obtained for the

sparingly soluble substituted ferrocenes 11 and 12 at nominal

concentrations of about 2 µM in CH₂Cl₂using 0.10 M [n-Bu₄N]-

[PF₆] as the electrolyte and activated alumina as an internal

desiccant. The substituted cobaltocenes 13-16 were analyzed

similarly, using instead acetonitrile as the solvent. The ap-

paratus was a BioAnalytical Systems BAS-100B automated

digital potentiostat with a Pt-disk working electrode, a Pt wire

auxiliary electrode, and a Ag/AgCl reference electrode. Cyclic

voltammetric (CV) sweeps for substituted ferrocenes were

typically initialized at +100 mV, scanned to +1100 mV, and

reversed to +100 mV, at a scan rate 100 mV s^-1. CV sweeps

for substituted cobaltocenium hexafluorophosphates were typi-

cally initialized at -100 mV, scanned to -1300 mV, and

The mixture was cooled and filtered through Celite, using 30

mL of THF to wash the filter. The filtrate was evaporated,

and the residue was taken up in 50 mL of THF and added

using a canulla to a stirred 10% aqueous NH₄Cl solution (250

mL). The mixture was stirred for 20 min and then rotary-

evaporated until the organic solvents were removed. The

resulting aqueous mixture was extracted with 150 mL of

benzene. The organic layer was washed with water, dried over

MgSO₄, filtered, and evaporated to afford a dark brown residue

that was shown by ¹⁹F NMR analysis to comprise a mixture

of C₆F₅-substituted cyclopentadienes. The mixture was redis-

solved in about 1 L of hexanes, filtered through alumina to

remove highly colored impurities, and evaporated. The result-

ing residue was taken up in about 25 mL of hexanes and

filtered; the insoluble portion was pure 4 (see following

procedure). The filtrate was loaded onto a 10 × 70 cm column

of silica gel, which was eluted with hexanes. After a long

forerun, an initial band afforded 1,4-bis(pentafluorophenyl)-

cyclopentadiene (2). A subsequent band afforded 1.41 g (2.50

mmol, 23%) of 3 as an off-white solid. Mp: 106-108 °C. ¹H

NMR (CDCl₃): δ 7.32 (s, 1 H), 4.14 (s, 2 H). ¹⁹F NMR (CDCl₃):

reversed to -100 mV, at a scan rate of 100 mV s^-1. |E_ox- E_red

|

were less than 80 mV, and I_c/I_aratios were within 15% of unity,

both indicators of substantially reversible couples.

Shifts in oxidation potential of substituted ferrocenes were

obtained by Osteryoung square-wave voltammetry (OSWV)

using ferrocene as an internal standard. The cell voltage was

typically swept from -200 to +1100 mV with a step resolution

of 4 mV, a square wave-amplitude of 25 mV, and frequency of

15 Hz. Shifts in oxidation potentials of substituted cobaltocenes

were similarly determined by OSWV using a sweep from -100

to -1300 mV and either octamethylferrocene or 9 as an

internal standard.

3

δ -141.03 (d, J ) 21 Hz, 2 F), -141.40 (d, J ) 22 Hz, 2 F),

-141.56 (d, ³J ) 22 Hz, 2 F), -154.21 (t, ³J ) 21 Hz, 1 F),

-154.41 (t, ³J ) 21 Hz, 1 F), -156.49 (t, ³J ) 21 Hz, 1 F),

-162.30 (m, 4 F), -163.46 (m, 2 F). {¹⁹F}¹³C NMR (CDCl₃): δ

144.6 (s, CF), 143.9 (s, CF), 143.8 (s, CF), 141.2 (s, CF), 141.0

(s, CF), 140.0 (s, CF), 138.0 (s, CF), 137.8 (s, CF), 137.8 (s,

CF), 137.3 (d, J ) 174 Hz, CH), 134.7 (t, J ) 5 Hz, C-C₆F₅),

133.6 (m, 2 unresolved C-C₆F₅signals), 110.3 (s, C_ipso), 110.1

(d, J ) 2 Hz, C_ipso), 109.9 (s, C_ipso), 47.7 (td, J ) 132 Hz, J )

8 Hz, CH₂). {¹H}¹³C NMR (CDCl₃): δ 144.7 (d, J ) 247 Hz,

CF), 144.0 (d, J ) 246 Hz, CF), 141.1 (d, J ) 251 Hz, CF),

140.0 (d, J ) 256 Hz, CF), 138.0 (d, J ) 244, CF), 137.9 (d, J

) 247 Hz, CF), 137.8 (d, J ) 262, CF), 137.3 (t, J ) 8 Hz,

CH), 134.7 (m, C-C₆F₅), 133.5 (s, 2 unresolved C-C₆F₅

signals), 110.2 (m, C_ipso), 110.1 (d, J ) 4 Hz, C_ispo), 109.9 (m,

Cr ysta llogr a p h ic Stu d ies. Two crystalline polymorphs of

2a were found under slightly different crystallization condi-

tions. The first, 2a -i, was obtained by cooling a moderately

concentrated hexanes solution for 24 h in a refrigerator at 5

°C. The second, 2a -ii, was obtained serendipitously during

studies of the reaction of 2a with dimethyl acetylenedicar-

boxylate (DMAD). These investigations will be reported in

more detail elsewhere. One crude product mixture contained

unreacted DMAD; unreacted 2a ; a product tentatively identi-

fied as 1,2-di(carbomethoxy)-3,5-bis(pentafluorophenyl)nor-

bornadiene (a Diels-Alder adduct of DMAD), 2b; and toluene

(residual reaction solvent). Crystallization of this mixture from

hexanes at -5 °C afforded 2a -ii. No attempts were made to

identify specific conditions that will lead predictably to the two

different polymorphs of 2a . Crystals of 4 were obtained by slow

evaporation of a concentrated hexanes solution at 25 °C.

Crystals of 11‚4(C₆H₆) were grown from benzene by slow

evaporation. Crystals of 16 were obtained by slow evaporation

of an n-octane solution at 25 °C. Crystals of 23‚C₆D₆were

obtained by cooling a solution of 23 in C₆D₆to 25 °C. Crystals

of 26 were obtained by dissolving the complex in warm hexanes

and cooling the solution to 0 °C for several days. Crystals of

27 were obtained by slow evaporation of a hexanes solution

at 25 °C. No unusual procedures were required in the prepara-

tion of any of these crystals, except that the benzene solvates

must be handled quickly after removal from their respective

mother liquors. The single-crystal X-ray diffraction analyses

(data collection, solution, and refinement) were unremarkable.

Relevant crystallographic data are presented in Table 1.

Sem iem p ir ica l Mod elin g. Calculations were carried out

at the PM3 level of semiempirical theory using PC Spartan

Pro software from Wavefunction, Inc. To map the conforma-

tional energies of 1,2-diarylcyclopentadienyl anions (aryl ) Ph,

C₆F₅), one aryl group was fixed at a given angle (0-90°, 5°

intervals), and the remaining geometric parameters were

energy-minimized. An analogous technique was used with

1,2,3-triarylcyclopentadienyl anions by fixing the torsion angle

of the central aryl group. Plots of the minimized energies as a

function of the fixed dihedral angle are provided in the

Supporting Information.

C

_ipso), 47.7 (s, CH₂). Anal. Calcd for C₂₃H₃F₁₅: C, 48.96; H, 0.54.

Found: C, 49.02; H, 0.35.

1,2,3-Tr is(pen taflu or oph en yl)cyclopen tadien e (4). From

the chromatographic separation described in the synthesis of

3, a third band was eluted to afford 1.04 g (1.84 mmol, 17%)

of 4 as an off-white crystalline solid. Mp: 216-217 °C. This

yield includes the solid obtained in the filtration immediately

prior to the chromatographic separation (see above). ¹H NMR

(CDCl₃): δ 6.97 (t, J ) 1.5 Hz, 1 H), 3.80 (d, J ) 1.2 Hz, 2 H).

¹⁹F NMR (CDCl₃): δ -141.61 (m, 4 F), -142.11 (d, ³J ) 23

3

Hz, 2 F), -153.24 (t, J ) 21 Hz, 1 F), -154.27 (t, J ) 20 Hz,

3

1 F), -154.66 (t, J ) 21 Hz, 1 F), -161.76 (m, 2 F), -162.22

(m, 2 F), -162.56 (m, 2 F). {¹⁹F}¹³C NMR (CDCl₃): δ 144.1 (s,

CF), 143.9 (s, CF), 143.7 (s, CF), 141.2 (s, CF), 141.0 (s, CF),

140.9 (s, CF), 139.7 (d, J ) 175 Hz, CH), 137.8 (s, CF), 137.6

(s, CF), 137.4 (s, CF), 134.4 (m, C-C₆F₅), 133.9 (m, C-C₆F₅),

132.3 (m, C-C₆F₅), 110.0 (s, C_ipso), 109.5 (d, J ) 2 Hz, C_ispo),

108.9 (s, C_ipso), 45.6 (td, J ) 134 Hz, J ) 8 Hz, CH₂). {¹H}¹³C

NMR (CDCl₃): δ 143.8 (br d, 3 unresolved CF), 141.1 (br d, J

) 284 Hz, 3 unresolved CF), 139.7 (s, CH), 137.6 (br d, 3

unresolved CF), 134.4 (s, C-C₆F₅), 133.9 (m, C-C₆F₅), 132.3

(s, C-C₆F₅), 45.5 (s, CH₂); the three C_ipsosignals were not

observed. Anal. Calcd for C₂₃H₃F₁₅: C, 48.96; H, 0.54. Found:

C, 49.21; H, 0.48.

Sod iu m

1,2,4-Tr is(p en t a flu or op h en yl)cyclop en t a -

d ien id e (7). A mixture of 1,2,4-tris(pentafluorophenyl)cyclo-

pentadiene (3, 3.04 g, 5.39 mmol), NaH (0.75 g, 31 mmol), and

THF (50 mL) was stirred at 25 °C for 1 day. The mixture was

filtered to remove unreacted NaH, and the filtrate was

evaporated. The residue was triturated with hexane, collected

on a fritted filter, and dried under vacuum at 80 °C to yield

1

1,2,4-Tr is(p en ta flu or op h en yl)cyclop en ta d ien e (3). A

mixture of lithium cyclopentadienide (0.790 g, 11.0 mmol),

hexafluorobenzene (9.74 g, 52.3 mmol), NaH (1.36 g, 56.7

mmol), and diglyme (50 mL) was stirred at 110 °C for 14 h.

2.99 g (5.10 mmol, 95%) of a tan solid. H NMR (THF-d₈): δ

6.71 (2 H). ¹⁹F NMR (THF-d₈): δ -146.64 (d, ³J ) 22 Hz, 4 F),

-148.01 (d, ³J ) 18 Hz, 2 F), -169.36 (t, ³J ) 20 Hz, 4 F),

-169.82 (tt, ³J ) 21 Hz, ⁴J ) 3 Hz, 2 F), -170.09 (m, 2 F),

Cyclopentadienyl Complexes of Fe, Co, Mn, and Re

Organometallics, Vol. 19, No. 25, 2000 5355

Ta ble 1. Cr ysta llogr a p h ic Da ta

2-i 2-ii

C₁₇H₄F₁₀C₁₇H₄F₁₀

4

11‚4C₆H₆

empirical formula

fw

diffractometer

C₂₃H₃F₁₅

C₄₆H₄F₃₀Fe‚4C₆H₆

1494.77

398.20

564.26

Enraf-Nonius

CAD4

0.30 × 0.30 × 0.10

tetragonal

7.4767(9)

35.794(2)

90

Enraf-Nonius

CAD4

Siemens P4

Enraf-Nonius

Kappa CCD

0.12 × 0.15 × 0.25

triclinic

11.7507(4)

13.1551(4)

20.7085(6)

84.443(2)

78.008(2)

70.1770(10)

2944.41(16)

P1h

cryst dimens (mm)

cryst system

a (Å)

b (Å)

c (Å)

0.22 × 0.28 × 0.48

monoclinic

14.5190(10)

5.0173(3)

20.5780(10)

90

0.44 × 0.34 × 0.15

monoclinic

13.1079(10)

13.7150(11)

8.3515(7)

90

R (deg)

â (deg)

108.009(7)

90

105.812(7)

90

γ (deg)

V (Å)³

1425.6(2)

P2₁/c

1444.6(2)

P2₁/c

2000.9(6)

P4₁2₁2

4

space group

Z

4

2

D

_calc(Mg m^-3

)

1.855

1.831

1.869

0.199

1104

1.686

abs coeff (mm^-1

F₀₀₀

)

1.8

0.197

0.4

784

1488

radiation

(K) (Å)

temp (K)

Cu K

1.54184

297

Mo K

0.71073

293

Mo K

0.71073

103

Mo K

0.71073

100

range for collection

no. of reflns colld

no. of indep reflns

abs corr method

max, min transm

data/restrts/params

R [I > 2σ(I)]

0.0-74.9

1.61-24.99

0.0-31.0

0.0-30.0

3418

3082

7175

16 606

2934

2513

2070

none

16 603

empirical

1.00, 0.89

2606/0/261

0.037

0.052

3.029

integration

0.95, 0.93

2531/0/261

0.0374

0.0788

0.732

empirical

0.95, 0.92

16603/0/910

0.040

0.078

0.762

1820/0/187

0.047

0.060

0.949

0.25, -0.20

R_w[I > 2σ(I)]

GoF on F²

largest diff peak, hole (e Å^-3

)

0.20, -0.16

0.13, -0.12

0.58, -0.42

16

23‚1/2C₆D₆

26

27

empirical formula

fw

diffractometer

cryst dimens (mm)

cryst system

a (Å)

C

₄₆H₄CoF₃₀

C₂₆H₂F₁₅MnO₃‚1/2C₆D₆

744.29

Siemens P4

0.48 × 0.49 × 0.32

triclinic

7.0264(13)

13.269(3)

15.229(3)

105.837

99.893

93.846

1335.9(4)

P1h (No. 2)

2

C₂₀H₃F₁₀O₃Re

667.43

Enraf-Nonius CAD4

0.18 × 0.25 × 0.30

triclinic

7.4561(8)

10.0723(8)

13.1941(13)

83.379(7)

74.710(9)

89.520(8)

949.18(16)

P1h (No. 2)

2

C₂₆H₂F₁₅O₃Re

833.48

1185.42

Siemens P4

0.51 × 0.16 × 0.08

monoclinic

7.5933(8)

14.2469(15)

38.263(4)

90

91.728(8)

90

4137.5(7)

P2₁/c (No. 14)

4

Siemens P4

0.17 × 0.19 × 0.48

monoclinic

7.6570(7)

10.8461(10)

29.606(2)

90

92.696(7)

90

2452.8(4)

P2₁/c (No. 14)

4

b (Å)

c (Å)

R (deg)

â (deg)

γ (deg)

V (Å)³

space group

Z

D_calc(Mg m^-3

)

1.903

0.588

2308

1.850

0.633

726

2.335

6.603

624

2.257

5.100

1568

abs coeff (mm^-1

F₀₀₀

)

radiation

λ (Å)

temp (K)

Mo KR

0.71073

293(2)

Mo KR

0.71073

293(2)

0.71073

293(2)

θ range for collection

no. of reflns colld

no. of indep reflns

abs corr method

max, min transm

no. of data/restrts/params

R [I > 2σ(I)]

1.06-21.50

1.81-25.00

0.0-30.4

2.00-26.00

6508

5919

11496

6677

4752

4670

ψ-scans

0.6518, 0.5727

4670/0/441

0.0454

0.1159

1.000

0.449, -0.418

5748

ψ-scans

0.9966, 0.7654

4463/0/308

0.0490

0.0590

1.680

2.11, -0.43

4828

integration

0.9667, 0.9052

4750/0/694

0.0467

0.1054

0.915

integration

0.6235, 0.5067

4826/0/406

0.0348

0.0986

0.896

R_w[I > 2σ(I)]

GoF on F²

diff peak, hole (e Å^-3

)

0.266, -0.319

1.052, -1.358

-175.29 (tt, ³J ) 21 Hz, ⁴J ) 5 Hz, 1 F). {¹⁹F}¹³C NMR (THF-

d₈): δ 143.5 (s, CF), 143.0 (s, CF), 137.9 (s, CF), 137.6 (s, CF),

136.3 (s, CF), 118.9 (s, CF), 117.6 (s, C_ipso), 115.2 (d, J ) 165

Hz, CH), 107.0 (m, C-C₆F₅), 105.3 (m, C-C₆F₅). {¹H}¹³C NMR

(THF-d₈): δ 143.0 (d, J ) 240 Hz, CF), 143.0 (d, J ) 243 Hz,

CF), 138.0 (d, J ) 242 Hz, CF), 137.6 (d, J ) 220 Hz, CF),

136.4 (d, J ) 240 Hz, CF), 133.9 (d, J ) 234 Hz, CF), 118.9 (t,

J ) 18 Hz, C_ipso), 117.6 (t, J ) 15 Hz, C_ipso), 115.2 (s, CH),

107.0 (s, C-C₆F₅), 105.3 (s, C-C₆F₅).

using instead 1,2,3-tris(pentafluorophenyl)cyclopentadiene (4,

1.73 g, 3.07 mmol) and NaH (0.41 g, 17 mmol) as the starting

materials to obtain 1.53 g (2.62 mmol, 85%) of tan solid. ¹H

NMR (THF-d₈): δ 6.16 (s, 2 H). ¹⁹F NMR (THF-d₈): δ -144.15

(d, ³J ) 20 Hz, 2 F), -144.45 (d, ³J ) 22 Hz, 4 F), -166.14 (tt,

4

3

³J ) 21 Hz, J ) 2 Hz, 1 F), -167.40 (m, 6 F), -168.12 (tt, J

4

) 21 Hz, J ) 3 Hz, 2 F). {¹⁹F}¹³C NMR (THF-d₈): δ 145.7 (s,

CF), 145.4 (s, CF), 139.4 (s, CF), 139.2 (s, CF), 138.8 (CF), 138.1

(CF), 120.7 (s, C_ipso), 113.9 (d, J ) 166 Hz, 2 CH), 107.2 (t, J

) 7.6 Hz, C-C₆F₅), 105.2 (s, C-C₆F₅); no signal that we could

assign to the second C_ipsowas observed. {¹H}¹³C NMR (THF-

Sod iu m

1,2,3-Tr is(p en t a flu or op h en yl)cyclop en t a -

d ien id e (8). The procedure for the synthesis of 7 was followed,

5356 Organometallics, Vol. 19, No. 25, 2000

Thornberry et al.

d₈): δ 145.7 (d, J ) 240 Hz, CF), 145.4 (d, J ) 240 Hz, CF),

139.4 (m, 2 unresolved CF), 138.8 (d, J ) 246 Hz, CF), 138.1

(d, J ) 243 Hz, CF), 120.7 (t, J ) 20 Hz, C_ipso), 113.9 (s, CH),

107.2 (s, C-C₆F₅), 105.2 (s, C-C₆F₅); no signal that we could

assign to the second C_ipsowas observed.

a filter, and dried to afford 0.086 g (0.072 mmol, 34%) of a

purple solid. ¹H NMR (THF-d₈): δ -20.5 (br s). ¹⁹F NMR (THF-

d₈): δ -43.3 (s, 2 F), -84.4 (br s, 8 F), -125.8 (s, 4 F), -142.5

3

(s, 2 F), -144.0 (s, 2 F), -146.2 (t, J ) 21 Hz, 2 F), -160.3

(br s, 8 F), -163.0 (s, 2F). Anal. Calcd for C₄₆H₄F₃₀Co: C, 46.73;

H, 0.34. Found: C, 46.58; H, 0.05.

Bis[η⁵-1,2,4-tr is(p en ta flu or op h en yl)cyclop en ta d ien yl]-

ir on (II) (11). A mixture of sodium 1,2,4-tris(pentafluoro-

phenyl)cyclopentadienide (7, 0.493 g, 0.841 mmol), FeBr₂

(0.095 g, 0.44 mmol), and THF (50 mL) was refluxed for 1 day,

and then the solvent was evaporated. The residue was

extracted with hot toluene (50 mL) and filtered, and the filtrate

was evaporated. The resulting orange solid was washed with

pentane (50 mL) on a fritted filter to afford 0.284 g (0.24 mmol,

57%) of an orange solid. ¹H NMR (THF-d₈): δ 5.61 (s, 4 H).

¹⁹F NMR (THF-d₈): δ -137.36 (br s, 8 F), -138.84 (d, ³J ) 16

Bis[η⁵-(p en t a flu or op h en yl)cyclop en t a d ien yl]cob a lt -

(III) Hexa flu or op h osp h a te (17). Air was bubbled through

a biphasic mixture of 13 (148 mg, 0.284 mmol), toluene (50

mL), water (50 mL), and 12 M aqueous HCl (20 mL) until the

dark color of the starting material in the organic phase

disappeared. The layers were separated, and the crude product

was precipitated with aqueous KPF₆, collected on a filter, and

dried in a vacuum desiccator over anhydrous calcium sulfate

1

to afford 161.9 mg (0.243 mmol, 86%) of a yellow powder. H

3

Hz, 4 F), -155.08 (t, J ) 20 Hz, 4 F), -156.78 (t, J ) 20 Hz,

2 F), -163.11 (m, 8 F), -163.92 (m, 4 F). Anal. Calcd for

NMR (acetone-d₆): δ 6.52 (t, ³J _HH) 2.0 Hz, 4 H), 6.27 (t, ³J _HH

) 2.1 Hz, 4 H). ¹⁹F NMR (acetone-d₆): δ -68.50 (d, ¹J _PF) 707

3

C

₄₆H₄F₃₀Fe: C, 46.73; H, 0.34. Found: C, 46.62; H, 0.16.

Bis[η⁵-1,2,3-tr is(p en ta flu or op h en yl)cyclop en ta d ien yl]-

Hz, 6 F), -133.49 (d, J _FF) 19 Hz, 2 F), -149.21 (tt, J _FF)

5

21 Hz, J _FH) 4 Hz, 4 F), -158.99 (m, 4 F). {¹H}¹³C NMR

1

(acetone-d₆): δ 146.2 (d, J _CF) 243 Hz), 142.9 (d, J _CF) 259

ir on (II) (12). A mixture of sodium 1,2,3-tris(pentafluoro-

phenyl)cyclopentadienide (8, 0.305 g, 0.520 mmol), FeBr₂

(0.052 g, 0.243 mmol), and THF (25 mL) was refluxed for 1

day. The THF was then evaporated. The residue was extracted

with hot toluene (25 mL), filtered, and cooled to 25 °C to afford

1

Hz), 139.2 (d, J _CF) 251 Hz), 106.6 (s), 91.7 (s), 88.1 (s), 86.6

2

(t, J _CF) 3 Hz). Anal. Calcd for C₂₂H₈F₁₆PCo: C, 39.66; H,

1.21. Found: C, 39.35; H, 1.11.

Bis{bis[η⁵-1,3-bis(pen taflu or oph en yl)cyclopen tadien yl]-

coba lt(III)} Su lfa te (18). In a 15 mL test tube fitted with a

stir bar, concentrated sulfuric acid (1 mL) was added to 40.0

mg of 14, and the mixture was stirred for 2 h at 25 °C. Using

a pipet, the resulting orange solution was added dropwise to

10 g of ice and then stirred until the ice melted. The yellow

precipitate was collected on a filter and washed with 10 mL

of water (upon which an indigo color suggesting partial

reduction to 14 had occurred) and ether (3 × 10 mL). The dark

precipitate was dissolved in 5 mL of acetone, filtered, repre-

cipitated in 100 mL of rapidly stirred anhydrous ether,

collected on a filter, and then dried in a vacuum oven at 60 °C

for 4 h to obtain 28 mg of a black-yellow solid. ¹H NMR

(acetone-d₆): δ 6.9 (br s, 4 H), 6.8 (br s, 2 H). ¹⁹F NMR (acetone-

d₆): δ -137.5 (m, 8 F), -152.9 (m, 4 F), -163.6 (m, 8 F).

Satisfactory microanalytical data were not obtained.

1

0.100 g (0.084 mmol, 35%) of orange crystals. H NMR (THF-

d₈): δ 5.63 (s, 4 H). ¹⁹F NMR (THF-d₈): δ -134.82 (s, 2 F),

3

-137.33 (br s, 8 F), -140.69 (d, J ) 24 Hz, 2 F), -153.96 (t,

3

³J ) 21 Hz, 2 F), -155.27 (t, J ) 20 Hz, 4 F), -162.17 (m, 2

F), -163.14 (m, 10 F). Anal. Calcd for C₄₆H₄F₃₀Fe: C, 46.73;

H, 0.34. Found: C, 46.50; H, 0.24.

1,1′-Bis(p en ta flu or op h en yl)coba ltocen e (13). A solution

of anhydrous cobalt(II) bromide (219 mg, 1.00 mmol) and

sodium (pentafluorophenyl)cyclopentadienide (5, 508 mg, 2.00

mmol) in THF was refluxed for 2 h and then cooled. The

solvent was evaporated, and the dark residue was extracted

with 50 mL of toluene and filtered. The dark filtrate was

evaporated to afford a violet-black solid, which was subse-

quently recrystallized from pentane at -20 °C to afford 405

mg (0.777 mmol, 78%) of violet crystals. An analytical sample

was obtained by sublimation (100-120 °C, 0.050 Torr). ¹H

NMR (THF-d₈): δ 29.1 (br), -44.9 (br). Anal. Calcd for

Tr ica r bon yl[η⁵-p en ta flu or op h en ylcyclop en ta d ien yl]-

m a n ga n ese(I) (21). A mixture of Mn(CO)₅Br (0.42 g, 1.5

mmol), (C₆F₅C₅H₄)Na (5, 0.42 g, 1.6 mmol), and THF (50 mL)

was refluxed for 1 day. The THF was evaporated, and the

resulting crude solid was dissolved in hexane and filtered. The

filtrate was cooled to -20 °C for 1 day, and 0.38 g (1.0 mmol,

C

₂₂H₈F₁₀Co: C, 50.70; H, 1.55. Found: C, 50.52; H, 1.61.

Bis[η⁵-1,3-b is(p en t a flu or op h en yl)cyclop en t a d ien yl]-

coba lt(II) (14). A mixture of 6 (0.500 g, 1.20 mmol), CoBr₂

(0.130 g, 0.600 mmol), and THF (100 mL) was stirred for 0.5

h at 25 °C, and then the solvent was then evaporated. Under

air, the residue was dissolved in 1 L of benzene, washed with

50 mL of 1.0 M aqueous HCl, then water (3 × 50 mL), dried

over MgSO₄, filtered, and evaporated. The resulting solid was

triturated with pentane (3 × 50 mL) and dried under vacuum

to afford 410 mg (0.48 mmol, 80%) of a deep indigo micro-

crystalline solid. In CD₂Cl₂solution, the ¹H NMR spectrum

showed no signals, whereas ¹⁹F NMR showed a set of two broad

(fwhm ≈ 10 000 Hz), unresolved signals at about -160 and

-175 ppm. Anal. Calcd for C₃₄H₆F₂₀Co: C, 47.83; H, 0.71.

Found: C, 48.09; H, 0.77.

Bis[η⁵-1,2,4-tr is(p en ta flu or op h en yl)cyclop en ta d ien yl]-

coba lt(II) (15). A mixture of 7 (0.494 g, 0.843 mmol), CoBr₂

(0.093, 0.425 mmol), and THF (25 mL) was refluxed for 1 day,

cooled, and filtered. Evaporation of the filtrate afforded a violet

solid, which was triturated with toluene, collected on a filter,

and dried under vacuum to afford 0.337 g (0.284 mmol, 67%)

of a purple solid. ¹H NMR (THF-d₈): δ 1.7 (br s). ¹⁹F NMR

(THF-d₈): δ -84.60 (br s, 4 F), -100.76 (br s, 10 F), -129.73

(s, 4 F), -163.09 (s, 8 F), -179.62 (s, 4 F). Anal. Calcd for

67%) of a yellow solid was collected. IR (hexane): ν_CO(cm^-1

)

) 2032 (A₁), 1954 (E). ¹H NMR (C₆D₆): δ 4.67 (pentet, J ) 2.0

Hz, 2 H), 3.87 (t, J ) 2.2 Hz, 2 H). ¹⁹F NMR (C₆D₆): δ -139.56

3

5

3

(dd, J ) 22.0 Hz, J ) 5.5 Hz, 2 F), -155.50 (t, J ) 22.2 Hz,

1 F), -162.33 (m, 2 F). Anal. Calcd for C₁₄H₄F₅MnO₃: C, 45.43;

H, 1.09. Found: C, 45.64; H, 0.95.

Tr ica r b on yl[η⁵-1,3-b is(p en t a flu or op h en yl)cyclop en -

ta d ien yl]m a n ga n ese(I) (22). A mixture of Mn(CO)₅Br (0.23

g, 0.84 mmol), [1,3-(C₆F₅)₂C₅H₃]Na (6, 0.36 g, 0.86 mmol), and

THF (50 mL) was stirred under reflux for 2 days. The THF

was then evaporated, and the resulting crude solid was

dissolved in hexane and filtered. Cooling the filtrate to -78

°C for about 5 h precipitated the product, which was collected

on a filter and dried under vacuum to afford 0.31 g (0.58 mmol,

69%) of a yellow solid. IR (hexane): ν_CO(cm^-1) 2035 (A₁), 1966,

1

1960. H NMR (C₆D₆): δ 5.65 (pentet, J ) 1.3 Hz, 1 H), 4.76

(s, 2 H). ¹⁹F NMR (C₆D₆): δ -139.34 (d, ³J ) 18.0 Hz, 4 F),

3

-154.24 (tt, J ) 21.8 Hz, 2 F), -161.66 (m, 4 F). Anal. Calcd

for C₂₀H₃F₁₀MnO₃: C, 44.80; H, 0.56. Found: C, 45.03; H, 0.37.

Tr ica r bon yl[η⁵-1,2,4-tr is(p en ta flu or op h en yl)cyclop en -

ta d ien yl]m a n ga n ese(I) (23). A mixture of Mn(CO)₅Br (0.102

g, 0.371 mmol), [1,2,4-(C₆F₅)₃C₅H₂]Na (7, 0.224 g, 0.382 mmol),

and THF (20 mL) was refluxed for 1 day. The THF was

evaporated, and the resulting crude solid was recrystallized

from hexanes to afford 0.065 g (0.093 mmol, 25%) of an orange

crystalline solid. IR (hexane): ν_CO(cm^-1) 2040 (A₁), 1973, 1965.

C

₄₆H₄F₃₀Co: C, 46.73; H, 0.34. Found: C, 46.58; H, 0.05.

Bis[η⁵-1,2,3-tr is(p en ta flu or op h en yl)cyclop en ta d ien yl]-

coba lt(II) (16). A mixture of 8 (0.265 g, 0.422 mmol), CoBr₂

(0.047 g, 0.215 mmol), and THF (50 mL) was refluxed for 1

day, cooled, and filtered. Evaporation of the filtrate afforded

a violet solid, which was triturated with hexanes, collected on

Cyclopentadienyl Complexes of Fe, Co, Mn, and Re

Organometallics, Vol. 19, No. 25, 2000 5357

¹H NMR (C₆D₆): δ 5.51 (s, 2 H). ¹⁹F NMR (C₆D₆): δ -137.37

lated cyclopentadienes. The arylation reaction presum-

ably occurs via formal nucleophilic aromatic substitution.

Product yields as a function of solvent, reaction tem-

perature, and reaction time are summarized in Table

2. A large excess of C₆F₆was used in all reactions. The

choice of either lithium or sodium as the counterion for

the starting material made no significant difference in

the observed product ratios or isolated yields. Extended

reaction times using diglyme as the solvent (entries 6

and 7) gave the best selectivity for the triarylated

products. Although significant decomposition to dark,

intractible byproducts decreased the isolated yields, the

benefit of this procedure is the ability to isolate ap-

preciable amounts of both regioisomers from a simple,

inexpensive, one-pot procedure.

3

(d, J ) 18 Hz, 4 F), -140.24 (d, J ) 18 Hz, 2 F), -152.17 (d,

3

³J ) 22 Hz, 2 F), -154.20 (t, J ) 22 Hz, 1 F), -161.66 (td, J

) 23 Hz, ⁵J ) 7 Hz, 4 F), -162.10 (m, 2 F). Anal. Calcd for

₂₆H₂F₁₅MnO₃: C, 44.47; H, 0.29. Found: C, 44.32; H, 0.12.

Tr ica r bon yl[η⁵-1,2,3-tr is(p en ta flu or op h en yl)cyclop en -

C

ta d ien yl]m a n ga n ese(I) (24). A mixture of Mn(CO)₅Br (0.104

g, 0.378 mmol), [1,2,3-(C₆F₅)₃C₅H₂]Na (0.227 g, 0.387 mmol),

and THF (20 mL) was stirred under reflux for 1 day. The THF

was evaporated, and the resulting crude solid was recrystal-

lized from toluene at -78 °C to afford 0.124 g (0.176 mmol,

47%) of an orange solid. IR (hexane): ν_CO(cm^-1) 2041 (A₁),

1975, 1963. H NMR (C₆D₆): δ 4.47 (s, 1 H). ¹⁹F NMR (C₆D₆):

1

3

δ -131.13 (d, J ) 22 Hz, 1 F), -137.13 (d, J ) 18 Hz, 4 F),

-141.14 (d, ³J ) 23 Hz, 1 F), -150.34 (t, ³J ) 22 Hz, 1 F),

3

5

-151.83 (t, J ) 22 Hz, 2 F), -160.61 (td, J ) 22 Hz, J ) 7

Hz, 1 F), -161.51 (td, ³J ) 23 Hz, ⁵J ) 8 Hz, 4 F), -161.76

Compounds 1-4 are separated by silica gel chroma-

tography and isolated as colorless, crystalline solids.

Comparisons of crude yields determined by ¹H NMR

(entries 2 and 4) with isolated, purified compounds

(entries 3 and 5) show that the separation procedure is

relatively inefficient. Chromatography is impeded by the

sparing solubility of 3 and 4. As we noted earlier,

(pentafluorophenyl)cyclopentadiene (1) is obtained as a

mixture of double-bond regioisomers and dimerizes

slowly at room temperature.¹³The Supporting Informa-

tion contains an improved procedure for isolating 1

using a cold aqueous workup to minimize dimerization.

The diarylated cyclopentadiene 2 forms two double-bond

isomers. One isomer (2a ) has time-averaged C_2vsym-

metry, while the other isomer (2b) has C_ssymmetry,

(td, ³J ) 22 Hz, ⁵J ) 8 Hz, 1 F). Anal. Calcd for C₂₆H₂F₁₅

-

MnO₃: C, 44.47; H, 0.29. Found: C, 44.07; H, 0.17.

Tr ica r bon yl[η⁵-1,2,4-tr is(p en ta flu or op h en yl)cyclop en -

ta d ien yl]r h en iu m (I) (27). A mixture of Re(CO)₅Br (0.149 g,

0.367 mmol), 7 (0.225 g, 0.384 mmol), and THF (20 mL) was

stirred under reflux for 1 day, and then the solvent was

evaporated. The residue was dissolved in hexane and filtered

through a pad of neutral alumina. The filtrate was evaporated,

and the residue was recrystallized from methanol to yield

0.153 g (0.184 mmol, 50%) of a tan solid. IR (octane): ν_CO(cm^-1

)

1

2042 (A₁), 1963, 1957. H NMR (CDCl₃): δ 6.36 (s, 2 H). ¹⁹F

3

NMR (CDCl₃): δ -136.42 (s, 4 F), -139.22 (d, J ) 15 Hz, 2

3

F), -152.22 (t, J ) 21 Hz, 2 F), -153.95 (t, J ) 21 Hz, 1 F),

-161.00 (m, 4 F), -161.46 (m, 2 F). Anal. Calcd for C₂₆H₂F₁₅O₃-

Re: C, 37.47; H, 0.24. Found: C, 37.65; H, 0.08.

Tr ica r bon yl[η⁵-1,2,3-tr is(p en ta flu or op h en yl)cyclop en -

ta d ien yl]r h en iu m (I) (28). A mixture of Re(CO)₅Br (57.5 mg,

0.142 mmol), 8 (81.0 mg, 0.138 mmol), sodium hydride (35 mg,

1.4 mmol), and toluene (25 mL) was stirred at 105 °C for 24

h, and then the hot mixture was filtered. The filtrate was

evaporated, and the residue was triturated with hexanes and

collected on a filter to afford 50 mg (0.060 mmol, 43%) of a

tan solid. IR (octane): ν_CO(cm^-1) 2043 (A₁), 1965, 1954. ¹H

NMR (CDCl₃): δ 5.87 (s, 2 H). ¹⁹F NMR (CDCl₃): δ -130.53

1

enabling structural differentiation by H and ¹⁹F NMR

spectroscopy. The isomers 2a and 2b coelute in our

chromatographic procedure. The structure of 2a is

confirmed by crystallographic analysis (see below). Upon

standing for several weeks, crystalline 2a will slowly

convert to a mixture of 2a and 2b. No effort has yet been

made to study this interconversion in detail. A mixture

of 2a and 2b is converted to the common conjugate base

6 upon treatment with NaH.

3

(m, 1 F), -135.96 (d, J ) 22 Hz, 4 F), -140.79 (d, J _FF) 24

3

Hz, 1 F), -150.40 (t, J _FF) 21 Hz, 1 F), -151.75 (t, J ) 24

Hz, 2 F), -159.80 (m, 1 F), -160.72 (m, 5 F). Anal. Calcd for

C

₂₆H₂F₁₅O₃Re: C, 37.47; H, 0.24. Found: C, 37.27; H, 0.20.

Obser va t ion of Cob a lt ocen iu m Ion s. To about 10 mg

each of cobaltocenes 14, 15, and 16 in three separate 5 mm

NMR tubes was added about 1.0 mL each of D₂SO₄. The tubes

were capped and shaken occasionally over 20 min. The

cobaltocene complexes appeared to dissolve completely to

afford faintly turbid yellow-orange solutions of 18, 19, and 20,

respectively. The samples were subjected to ¹H and ¹⁹F NMR

spectroscopic analysis in D₂SO₄solvent referenced to the

residual solvent isotopomer (HDSO₄) at 10.90 ppm. Due to the

viscosity of the sulfuric acid solvent, all of the NMR spectra

exhibited sufficient broadening to preclude observation of

spin-spin couplings, so no multiplicity information is pre-

sented. Data for 18: ¹H NMR (D₂SO₄): 6.86 (2 H), 6.64 (2 H).

¹⁹F NMR (D₂SO₄): -140.7 (8 F), -151.0 (4 F), -163.0 (8 F).

Data for 19: ¹H NMR (D₂SO₄): δ 6.77 (4 H). ¹⁹F NMR (D₂-

SO₄): δ -138.2 (4 F), -138.5 (8 F), -147.8 (2 F), -148.9 (4

F), -161.4 (8 F), -161.9 (4 F). Data for 20: ¹H NMR (D₂SO₄):

δ 6.80 (4 H). ¹⁹F NMR (D₂SO₄): δ -137.8 (2 F), -138.5 (8 F),

-141.9 (2 F), -148.2 (4 F), -148.6 (2 F), -160.4 (2 F), -160.9

(10 F).

Both of the two triarylated regioisomers 3 and 4

exhibit time-averaged C_ssymmetry in solution. In the

1,2,4-isomer 3 both the CH₂and CH protons are

adjacent to two electron-withdrawing C₆F₅groups,

whereas in the 1,2,3-isomer 4, the CH₂and CH groups

are each adjacent to only one C₆F₅group. Accordingly,

1

3 shows H NMR chemical shifts for the CH₂(4.11 ppm)

and CH (7.29 ppm), downfield of the corresponding

signals (CH₂, 3.80 ppm; CH, 6.97 ppm) in the spectrum

of the 1,2,3 isomer 4. Furthermore, only 4 exhibits CH-

CH₂coupling (J ≈ 1.5 Hz). This regiochemical assign-

ment is confirmed by an X-ray diffraction analysis of

crystalline 4 (see below). Regioselectivity in the forma-

tion of the triarylated cyclopentadienes 3 and 4 is

minimal (Table 2).

Both triarylated cyclopentadienes 3 and 4 are con-

verted efficiently to their corresponding sodium salts 7

and 8, respectively, under similar deprotonation condi-

tions (Scheme 1). The triarylated sodium cyclopentadi-

enides 7 and 8 both demonstrated C_2vsymmetry in THF

Resu lts a n d Discu ssion

Liga n d Syn th eses. As shown in Scheme 1, sodium

cyclopentadienide (NaCp) reacts with hexafluoroben-

zene (C₆F₆) to afford mixtures of mono-, di-, and triary-

5358 Organometallics, Vol. 19, No. 25, 2000

Thornberry et al.

Sch em e 1

Ta ble 2. Con d ition s a n d Yield s of C₆F ₅-Su bstitu ted

Cyclop en ta d ien e Syn th eses

E, along with their use as ligands for transition metals,

are underway.

reaction conditions

isolated yields (% based on NaCp)

T

time

entry solvent (°C) (h)

1

2

3

4

1.

2.

3.

4.

5.

6.

7.

DME

85

12

24

96

14

25

9

19

26^a

17

6^a

8

10

22^a

5

10

19^a

11

20^a

8

2^a

0

0^a

0

32^a

18

23

22

Syn th esis of P en taflu or oph en yl-Su bstitu ted Fer -

r ocen es. As Scheme 1 shows, reactions of iron(II)

bromide with 2 equiv of an arylated cyclopentadienide

ligand (5-8) afford the corresponding ferrocene com-

plexes 9-12, which are isolated as air-stable, red-orange

solids. We reported the diarylated and tetraarylated

ferrocenes 9 and 10 earlier,¹³and the structure of 9 was

also reported elsewhere.¹⁶Although ferrocene is highly

soluble in hydrocarbons, 9 is noticeably less soluble, and

10 is only sparingly soluble. The hexaarylated com-

plexes 11 and 12 are essentially insoluble in hydrocar-

bons at room temperature, nor are they appreciably

soluble in “fluorous” solvents including hexafluoroben-

zene,²⁰but they are sufficiently soluble in THF-d₈for

¹H and ¹⁹F NMR measurements.

diglyme 110

3

1

17

14

0

a

¹H NMR yields determined from the crude product mixture

prior to separation.

solution on the basis of the number and intensity of

signals in their H and ¹⁹F NMR spectra. Like the mono-

1

and diarylated ligand salts 5 and 6 that we reported

earlier, the triarylated salts 7 and 8 remain partly

solvated by small but significant amounts of THF (up

to 10 mol %) even after heating to 80 °C under high

vacuum for several hours.

In our preliminary explorations of the synthesis of

C₆F₅-substituted cyclopentadienes,¹³we dismissed many

small, poorly reproducible NMR signals as arising from

“minor impurities” (such as the disubstituted diene

isomer 2b without pursuing their assignments. Re-

cently, we also learned that two additional compounds

observed in some reactions are the vicinally disubsti-

tuted regioisomer D and the tetrasubstituted cyclopen-

tadiene E. The isolation and characterization of D and

Syn th esis of P en ta flu or op h en yl-Su bstitu ted Co-

ba ltocen es. The syntheses of the substituted cobal-

tocenes are analogous to those of the ferrocenes (Scheme

(20) (a) Barthel-Rosa, L. P.; Gladysz, J . A. Coord. Chem. Rev. 1999,

192, 587. (b) Horvath, I. T. Acc. Chem. Res. 1998, 31, 641.

Cyclopentadienyl Complexes of Fe, Co, Mn, and Re

Organometallics, Vol. 19, No. 25, 2000 5359

E_1/2, mV

Ta ble 3. M^II|M^IIIOxid a tion P oten tia ls of Ar yla ted F er r ocen es a n d Coba ltocen es^a

E_1/2, mV

ferrocenes^b

vs Cp₂Fe

vs SCE^c

cobaltocenes^d

Cp₂Co^e

[(C₆F₅)C₅H₄]₂Co (13)

[1,3-(C₆F₅)₂C₅H₃)₂Co (14)

vs Cp₂Co

vs SCE^e

vs NHE^f

Cp₂Fe^c

410

760

1090

1350

1360

-910

-510

-150

-670

-270

+90

[(C₆F₅)C₅H₄]₂Fe (9)^g

348(5)

680(5)

940(10)

951(10)

400(5)

760(20)

[1,3-(C₆F₅)₂C₅H₃)₂Fe (10)^g

[1,2,4-(C₆F₅)₃C₅H₂]₂Fe (11)

[1,2,3-(C₆F₅)₃C₅H₂]₂Fe (12)

a

Experimental errors in parentheses are average deviations from the mean over at least two experiments. Complexes [1,2,4-

b

(C₆F₅)₃C₅H₂]₂Co (15) and [1,2,3-(C₆F₅)₃C₅H₂]₂Co (16) were not sufficiently soluble for voltammetric analysis. Dichloromethane solvent,

0.10 M TBAH electrolyte. ^cThe potential of Cp₂Fe|Cp₂Fe⁺vs SCE was reported in ref 48a. The values in this column are approximations

intended for comparison to the cobaltocene data and were obtained by adding 410 mV to the potentials obtained vs Cp₂Fe|Cp₂Fe⁺in

dichloromethane. Acetonitrile solvent, 0.10 M TBAH electrolyte. ^eThe potential of Cp₂Co|Cp₂Co⁺vs SCE in acetonitrile was reported in

d

ref 48b. The values in this column were obtained by subtracting 910 mV from the potentials obtained vs Cp₂Co|Cp₂Co⁺in acetonitrile.

^fThe values in this column were obtained by adding 240 mV to the potentials obtained vs SCE. See ref 21. Data from ref 13.

g

1). The diarylated complex 13 and the tetraarylated

complex 14 are dark blue, while the hexaarylated

complexes 15 and 16 are purple. The NMR spectra of

13-16 vary widely in their utility. Because these are

paramagnetic compounds, extensive broadening and

shifting of the resonances is expected. For example, the

diarylated cobaltocene 13 showed two broad signals in

the ¹H NMR spectrum (+29 and -45 ppm), but we could

not identify any signals in the ¹⁹F NMR spectrum. The

tetraarylated cobaltocene 14 was the most recalcitrant,

such as NH₃and is reduced by water to the correspond-

ing Co^IIion.

These findings, which are corroborated by electro-

chemical measurements (see below), represent a “proof

of concept” that C₆F₅substituents can stabilize a 19-

electron complex toward oxidation. We believe this

stabilization arises from the electron-withdrawing ten-

dency of the C₆F₅groups, although crsytallographic

characterization of a hexaarylated cobaltocene 16 sug-

gests that fluorine atoms in ortho positions of the C₆F₅

groups may surround the Co atom and protect it from

an approaching oxidant molecule. We are currently

exploring the use of C₆F₅groups to stabilize other “low

oxidation state” organometallics metal species including

zirconocene(II) species and 19-electron carbonylate an-

ions.

Syn th esis of P en ta flu or op h en yl-Su bstitu ted Mn

a n d Re P ia n o Stool Com p lexes. Substitution reac-

tions of either Re(CO)₅Br or Mn(CO)₅Br with arylated

cyclopentadienylsodium ligands (Scheme 1) afforded the

corresponding “piano stool” tricarbonyl complexes as

yellow to orange (M ) Mn, 21-24) or colorless to tan

(M ) Re, 25-28) solids. The infrared spectra of these

complexes exhibit the expected two bands in alkane

solution corresponding to A₁and E symmetry-adapted

internal coordinates for carbonyl-stretching deforma-

tions. The E bands are split in some cases as a result of

perturbed C_3vsymmetry.

The synthetic yields of substituted ferrocenes, cobal-

tocenes, and piano stool complexes reported here are low

to moderate. Often, hydrolyzed ligands 1-4 were re-

covered during the workup and isolation of the desired

complexes, suggesting either incomplete substitution

reactions with the metal-containing precursors or hy-

drolysis of the ligands by adventitious moisture. The

latter is a common problem with reactions carried out

on a small scale, although we were careful to exclude

moisture from our reactions as much as possible.

P h ysicoch em ica l Eva lu a tion of Electr on ic Su b-

stitu en t Effects. The hexaarylated ferrocenes 11 and

12 as well as the diarylated and tetraarylated cobal-

tocenes 13 and 14 were analyzed by solution voltam-

metry. The reversibility of the Fe^III|Fe^IIand Co^III|Co^II

couples was established using cyclic voltammograms

(CVs), while the potential relative to internal Fc⁺|Fc (for

9-12) and to internal Cp₂Co⁺|Cp₂Co (for 13 and 14) was

determined by square wave voltammetry. The potential

data for these complexes (Table 3) reveal the expected

monotonic and roughly linear increase in E_1/2(Figure

1) with increasing ancillary arylation. The hexaarylated

1

providing neither H or ¹⁹F NMR spectra that we could

interpret, even when analyzing a concentrated sample.

However, the hexaarylated cobaltocenes 15 and 16

provided satisfactory ¹H NMR spectra (one broad signal

for each compound at 25 °C) and surprisingly sharp,

well-resolved ¹⁹F NMR spectra. The ¹H NMR spectra

were most useful in confirming the absence of significant

diamagnetic impurities such as hydrolyzed ligands or

residual solvent in the purified complexes. Elemental

microanalysis ultimately confirmed the composition and

purity of these species.

Oxid a tion of Su bstitu ted Coba ltocen es. The rela-

tive stabilities of the cobaltocenes toward oxidation was

particularly striking. Whereas Cp₂Co (formally a 19-

electron complex) is relatively sensitive to air, the

diarylated cobaltocene 13 is air-stable, and the tetraary-

lated 14 and hexaarylated cobaltocenes 15, 16 are stable

to air even in the presence of dilute aqueous hydrochlo-

ric acid. As the diarylated cobaltocene 13 is readily

oxidized using air and dilute aqueous HCl (Scheme 1),

we were able to isolate the corresponding diarylated

cobaltocenium hexafluorophosphate 17 as an air-stable,

yellow solid. In contrast, we observed the tetraarylated

and hexaarylated cobaltocenium ions 18-20 only after

dissolving the corresponding cobaltocenes 14-16 in

concentrated sulfuric acid-d₂, whereupon the ¹H and ¹⁹F

NMR spectra showed spectra characteristic of the

assigned diamagnetic [(Ar_nCp)₂Co^{III +}

] ions (18, n ) 2;

19 and 20, n ) 3), similar to those obtained for the

isoelectronic ferrocenes 10-12. The counterion associ-

2-

ated with 18-20 was presumed to be either SO₄or

-

DSO₄

.

Attempts to isolate the tetraarylated cobaltocenium

ion 18 as the hexafluorophosphate failed; diluting the

sulfuric acid solution of 18 led to the appearance of a

blue-black solid having a ¹⁹F NMR spectrum consistent

with the cobaltocene 14. This shows that water is a

suitable reductant for [(Ar₂Cp)₂Co^III]⁺. This behavior is

reminiscent of classical coordination chemistry: [(H₂O)₆-

Co ]

^{III 3+}is unstable in the absence of stabilizing ligands

5360 Organometallics, Vol. 19, No. 25, 2000

Thornberry et al.

F igu r e 2. Plot of carbonyl stretching wavenumbers for

CpMn(CO)₃(b) and CpRe(CO)₃(O) as a function of the

number of C₆F₅substituents. Wavenumber values have

errors of (1 cm^-1. Data were taken from Table 4.

F igu r e 1. Plot of electrochemical potentials for ferrocenes

(O) and cobaltocenes (b) as a function of the number of C₆F₅

substituents. Data were taken from Table 3.

Ta ble 5. Selected Metr ic Da ta for 2a in Tw o

Cr ysta llin e P olym or p h s (i a n d ii)

Ta ble 4. Ca r bon yl Str etch in g Wa ven u m ber s of

P ia n o Stool Com p lexes (Ar _nCp )M(CO)₃

(n ) 0-3, M ) Mn , Re, Ar ) C₆F ₅)

bond distances (Å)

2-i 2-ii

bond angles (deg)

2-i

2-ii

a

wavenumbers (cm^-1

)

C1-C2 1.481(2) 1.490(5) C1-C2-C3 107.57(14) 107.7(3)

C2-C3 1.362(2) 1.338(5) C2-C3-C4 109.79(14) 110.4(3)

C3-C4 1.443(3) 1.441(5) C3-C4-C5 109.14(15) 109.6(3)

C4-C5 1.375(2) 1.356(4) C4-C5-C1 107.67(14) 107.4(3)

C5-C1 1.471(2) 1.489(5) C5-C1-C2 105.82(13) 105.0(3)

C2-C6 1.462(2) 1.471(4) C1-C2-C6 124.56(13) 123.7(3)

C5-C12 1.467(2) 1.461(4) C3-C2-C6 127.84(15) 128.6(3)

C1-C5-C12 124.73(13) 124.6(3)

entry

complex

ν_CO(A)

ν_CO(E)

1944

1954

b

1.

2.

3.

4.

5.

6.

7.

8.

9.

CpMn(CO)₃

2028

2032

2035

2040

2041

2031

2034

2038

2042

2043

(ArCp)Mn(CO)₃(21)

(1,3-Ar₂Cp)Mn(CO)₃(22)

(1,2,4-Ar₃Cp)Mn(CO)₃(23)

(1,2,3-Ar₃Cp)Mn(CO)₃(24)

1966, 1960

1973, 1965

1975, 1963

1939

b

CpRe(CO)₃

C4-C5-C12 127.56(14) 128.0(3)

(ArCp)Re(CO)₃(25)^c

1947

1953

(1,3-Ar₂Cp)Re(CO)₃(26)^c

(1,2,4-Ar₃Cp)Re(CO)₃(27)

(1,2,3-Ar₃Cp)Re(CO)₃(28)

identical in the two polymorphs. The two double bonds

of the five-membered core are significantly shorter than

the three single bonds, and the C3-C4 single bond is

also somewhat shortened, as expected for a conjugated

diene. The nearly planar structures show Cp-C₆F₅

interplanar torsional angles (Table 6, entries 1-4) that

are much smaller than those observed in fluorinated

biphenyls such as 2,3,4,5,6-pentafluorobiphenyl (52.9%).²²

The corresponding angles observed for biphenyl (gas

phase) are comparable,²³whereas crystalline biphenyl

is nearly planar.²⁴Fluorinated biphenyls also typically

exhibit arene stacking in the solid state.²²Interestingly,

neither 2-i nor 2-ii exhibits intermolecular arene stack-

ing; however several intermolecular C-H‚‚‚F-C inter-

actions were identified (see below).

Although both polymorphs crystallized in monoclinic

(P2₁/c) crystal systems, an interesting difference be-

tween the three-dimensional packing arrangements of

2-i and 2-ii is apparent in the relative orientations of

face-to-face neighbors. In 2-i (Figure 4(i)), neighboring

molecular faces are related by simple translation along

the crystallographic b axis, whereas in 2-ii (Figure 4(ii)),

molecular faces are instead related by the c-glide

operation, leading to alternating orientations of the

methylene carbon. Despite these differences, the dis-

1963, 1957

1965, 1954

10.

a

Recorded in octane or hexanes solvent. Reproducibility is

within 1 cm^-1

.

Values of 2030 and 1946 cm^-1for CpMn(CO)₃and

b

2032 and 1940 for CpRe(CO)₃were reported elsewhere.⁴⁹We

prepared these complexes again for direct comparison with our

compounds. ^cPrior results from our laboratories.¹³

cobaltocenes 15 and 16 were insufficiently soluble for

reliable voltammetric measurements. Using standard

reference potentials,²¹we found that the oxidation

potentials of 13 and 14 vs NHE (Table 3) lie on either

side of 0 V, rationalizing the observed difference in

reactivity toward dilute aqueous HCl.

A clearly linear trend (Table 4, Figure 2) is also

observed in the infrared spectra of substituted CpM-

(CO)₃complexes (M ) Mn, Re). The linear trends in

these infrared spectroscopic and electrochemical data

demonstrate that effects of up to six C₆F₅substituents

do not approach an electronic “limit”, but rather the

electronic structure of the metallocene and piano stool

frameworks adjust to the influence of the substituents

over a wide range.

Cr yst a l St r u ct u r es of Ar yla t ed Cyclop en t a -

d ien es 2 a n d 4. As described in the Experimental

Section, we obtained two distinct crystalline polymorphs

of 2a (2-i and 2-ii) serendipitously. Relevant crystal data

are collected in Table 1, selected metric data are

presented in Table 5, and thermal ellipsoid plots are

shown in Figure 3. The molecular structures are nearly

(22) (a) Naae, D. G. Acta Crystallogr. 1979, B35, 2765. (b) Brock,

C. P.; Naae, D. G.; Goodhand, N.; Hamor, T. A. Acta Crystallogr. 1978,

B34, 3691.

(23) Almenningen, A.; Hartmann, A. O.; Seip, H. M. Acta Chem.

Scand. 1968, 22, 1013.

(24) (a) Trotter, J . Acta Crystallogr. 1961, 14, 1135. (b) Charbonneau,

G. P.; Delugeard, Y. Acta Crystallogr. 1976, B32, 1420. (c) Charbon-

neau, G. P.; Delugeard, Y. Acta Crystallogr. 1977, B32, 1586. (d)

Gleason, W. B.; Britton, D. Cryst. Struct. Commun. 1976, 5, 483.

(21) (a) Bard, A. J .; Faulkner, L. R. Electrochemical Methods;

Wiley: New York, 1980; p 3. (b) Lu, S.; Strelets, V. V.; Ryan, M. F.;

Pietro, W. J .; Lever, A. B. P. Inorg. Chem. 1996, 35, 1013.

Cyclopentadienyl Complexes of Fe, Co, Mn, and Re

Organometallics, Vol. 19, No. 25, 2000 5361

Ta ble 6. Cr ysta llogr a p h ic C₆F ₅-Cp In ter p la n a r Con for m a tion a l An gles

entry

compound

C₆F₅group

Cp-C₆F₅(deg)

entry

compound

C₆F₅group

Cp-C₆F₅(deg)

1.

2.

3.

4.

5.

6.

7.

8.

9.

2-i

C6-C11

C12-C17

C6-C11

C12-C17

C4-C7

C8-C13

C11-C16

C21-C26

C41-C46

C61-C66

C71-C76

C91-C96

13.0

15.4

10.4

4.5

57.0

54.8

39.6

77.0

30.7

36.7

75.2

27.3

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

16

C11-C16

C21-C26

C31-C36

C61-C66

C71-C76

C81-C86

C9-C14

C15-C20

C21-C26

C9-C14

C15-C20

C9-C14

C15-C20

C21-C26

46.9

83.8

49.4

52.6

56.6

45.4

43.5

47.7

36.4

23.0

39.3

66.3

51.0

41.5

2-ii

4^a

11

23‚1/2C₆D₆

10.

11.

12.

26

27

a

Only two C₆F₅groups in 4 are crystallographically unique.

that the -C(3)H₂-C(3a)H- moiety fits snugly into a

fluorinated “pocket” with crystallographically enforced

C₂symmetry. The Cp-C₆F₅torsional angles (Table 6,

entries 5 and 6) are much greater than the correspond-

ing angles in 2a (entries 1-4). Clearly, steric congestion

among the three C₆F₅substituents prevents the C₆F₅

groups from attaining more coplanar conformations

relative to the C1-C5 ring. A similar difference is

observed between biphenyl (torsion angle near 0°)²⁴and

1,2-diphenylbenzene (torsion angles of 43° and 60°).²⁵

Cr ysta l Str u ctu r es of Meta llocen es. The molecu-

lar structure of the hexaarylated ferrocene 11 is shown

in Figure 6. Examination of the metric data (Table 8)

shows that the metallocene “core” of the complex

exhibits nearly ideal pentahapticity within anticipated

experimental errors. Both Cp rings are planar, with no

carbon atom deviating more than 0.01 Å from its

respective least-squares plane, and the interplanar

angle between the Cp rings is 5.8(2)°.

In crystalline 11‚4(C₆H₆), the six C₆F₅substituents

are held at varying torsional angles (27-77°) relative

to the Cp ligands to which they are respectively attached

(Table 6, entries 7-12). In previously reported C₆F₅-

substituted ferrocenes lacking substituents vicinal to the

C₆F₅group, the C₆F₅torsional angles are in the range

15.7-32.4°,¹⁶but the presence of a vicinal substituent

(CH₂NMe₂or CH₂NMe₃⁺) increases the angle (40.7-

54.8°).³⁵Arene stacking is another striking feature of

F igu r e 3. Thermal ellipsoid plots (50% probability) of the

molecular structure of 2a in two polymorphs. Numbering

scheme is identical for both. (a) and (b) are 2-i. (c) and (d)

are 2-ii.

(25) (a) Aikawa, S.; Maruyama, Y.; Ohashi, Y.; Sasada, Y. Acta

Crystallogr. 1978, B34, 2901. (b) Brown, G. M.; Levy, H. A. Acta

Crystallogr. 1979, 35, 785.

(26) (a) Weck, M.; Dunn, A. R.; Matsumoto, K.; Coates, G. W.;

Lobkovsky, E. B.; Grubbs, R. H. Angew. Chem., Int. Ed. 1999, 38, 2741.

(b) Coates, G. W.; Dunn, A. R.; Henling, L. M.; Dougherty, D. A.;

Grubbs, R. H. Angew. Chem., Int. Ed. Engl. 1997, 36, 248. (c) Williams,

J . H. Acc. Chem. Res. 1993, 26, 593. (d) Lorenzo, S.; Lewis, G. R.; Dance,

I. New J . Chem. 2000, 24, 295.

(27) (a) Basolo, F. Pure Appl. Chem. 1988, 60, 1193. (b) O’Connor,

J . M.; Casey, C. P. Chem. Rev. 1987, 87, 307. (c) Marder, T. B.;

Calabrese, J . C.; Roe, C. D.; Tulip, T. H. Organometallics 1987, 6, 2012.

(d) Wescott, S. A.; Kakkar, A. K.; Stringer, G.; Taylor, N. J .; Marder,

T. B. J . Organomet. Chem. 1990, 394, 777. (e) Faller, J . W.; Crabtree,

R. H.; Habib, A. Organometallics 1985, 4, 929. (f) Carl, R. T.; Hughes,

R. P.; Rheingold, A. L.; Marder, T. B.; Taylor, N. J . Organometallics

1988, 7, 1613.

(28) Nangia, A.; Desiraju G. R. Top. Curr. Chem. 1998, 198, 57.

(29) (a) Braga, D.; Grepioni, F. Coord. Chem. Rev. 1999, 183, 19.

(b) Rivas, J . C. M.; Brammer, L. Coord. Chem. Rev. 1999, 183, 43.

(30) (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, 2023. (d) Williams, J . H.;

Cockcroft, J . K.; Fitch, A. N. Angew. Chem., Int. Ed. Engl. 1992, 31,

1655.

tance between molecular least-squares planes for the

two polymorphs was nearly identical (3.60 Å for 2-i and

3.56 Å for 2-ii), and the unit cell volumes are within

2% of one another.

We obtained crystals of 4 both to confirm our NMR-

spectroscopic regiochemical assignment and to examine

its molecular structure, particularly with regard to

effects of the vicinal C₆F₅substituents. Crystal data are

presented in Table 1, metric data are shown in Table

7, and a diagram of the molecular structure is shown

in Figure 5. The molecule exhibits crystallographic C₂

symmetry with the C1-C4 and C7-F3 bonds resting

on the C₂axis. Noteworthy is a 0.5:0.5 positional

distortion that interchanges the methylene and methine

carbon atoms (C3 and C3A). Packing diagrams reveal

5362 Organometallics, Vol. 19, No. 25, 2000

Thornberry et al.

F igu r e 4. Partial crystallographic packing stereoviews of 2. (i) Polymorph 2-i. (ii) Polymorph 2-ii. Hydrogen and fluorine

atoms were omitted for clarity.

F igu r e 5. Thermal ellipsoid plot of 4 shown at 50%

probability.

Ta ble 7. Selected Metr ic Da ta for Cr ysta llin e 4

distances (Å)

bond angles (deg)

F igu r e 6. Thermal ellipsoid plot of 11‚4(C₆H₆). Benzene

solvate molecules were omitted for clarity. H and F atoms

are numbered the same as the C atoms to which they are

attached.

C1-C2

1.4113(17)

1.420(3)

1.486(3)

1.4761(15)

1.477(2)

C1-C2-C3

108.94(14)

106.58(17)

108.95(11)

125.52(8)

C2-C3

C3-C3a

C1-C4

C2-C8

C2-C3-C3a

C2a-C1-C2

C2-C1-C4

C1-C2-C8

C3-C2-C8

124.01(13)

127.05(15)

C₆F₅-C₆F₅stacking are instead associated with benzene

solvate molecules (entries 3 and 6). Two additional

C₆F₅-C₆H₆interactions (entries 4 and 5) arise from a

benzene molecule “sandwiched” between two C₆F₅groups

of adjacent molecules. A packing diagram (Figure 7)

shows this extended stacking motif. We were unable to

grow crystals from C₆F₆in order to examine the pos-

sibility of C₆F₅-C₆F₆stacking in these complexes.

crystalline 11‚4(C₆H₆). Each ligand has two C₆F₅groups

that are engaged in intramolecular C₆F₅-C₆F₅stacking

(Table 9, entries 1 and 2). Hughes and co-workers have

rationalized this “slip-stacked” structural motif in crys-

talline 9 in terms of dipolar and quadrupolar charge

distributions of the C₆F₅groups.¹⁶

The striking tendency for nonfluorinated aryl groups

and perfluoroaryl groups to engage in mutual stacking

has also been the subject of recent experimental and

theoretical investigation.²⁶In 11‚4(C₆H₆), both of the

C₆F₅groups that are not engaged in intramolecular

The fact that all of the C₆F₅groups of 11 are engaged

in some form of arene stacking suggests that this motif

is a significant contributor to crystal packing. Metric

data (Table 9) suggest less than idealized stacking

interactions, as the least-squares planes are signifi-

cantly nonparallel, and the centroid-centroid distances

are relatively long, compared with those observed in

crystalline 9 (entry 10). Arene stacking also influences

the molecular structure (conformations) of 11. Compar-

ing the C₆F₅torsion angles of the two “vicinal” C₆F₅

groups on each Cp ligand (Table 6), we observe that the

C₆F₅groups engaged in intramolecular C₆F₅-C₆F₅

stacking (entries 7 and 10) exhibit much smaller C₆F₅-

Cp torsion angles than the C₆F₅groups not so engaged

(entries 8 and 11).

(31) Weck, M.; Dunn, A. R.; Matsumoto, K.; Coates, G. W.; Lob-

kovsky, E. B.; Grubbs, R. H. Angew. Chem., Int. Ed. 1999, 38, 2741.

(32) Thalladi, V. R.; Weiss, H.-C.; Bla¨ser, D.; Boese, R.; Nangia, A.;

Desiraju, G. R. J . Am. Chem. Soc. 1998, 120, 8702.

(33) This interaction was observed in bis(N-pentafluorophenyl)-

tetraphenylcyclodisilazane benzene solvate. Two centroid-centroid

distances are needed to describe this interaction, which does not appear

at the crystallographic center. Claire, P.; Sowerby, D. B.; Haiduc, I. J .

Organomet. Chem. 1986, 310, 161.

(34) (a) Gallivan, J . P.; Dougherty, D. A. Org. Lett. 1999, 1, 103. (b)

Herna´ndez-Trujillo, J .; Vela, A. J . Phys. Chem. 1996, 100, 6524.

(35) Deck, P. A.; Lane, M. J .; Montgomery, J . L.; Slebodnick, C.;

Fronczek, F. R. Organometallics 2000, 19, 1013.

Cyclopentadienyl Complexes of Fe, Co, Mn, and Re

Organometallics, Vol. 19, No. 25, 2000 5363

Ta ble 8. Selected Bon d Dista n ces (Å) a n d An gles (d eg) for th e Meta llocen e Cor es of 11‚4(C₆H₆)

(M ) F e) a n d 16 (M ) Co)

11‚4(C₆H₆)

16

11‚4(C₆H₆)

16

M-C1

2.071(2)

2.061(2)

2.067(2)

2.039(2)

1.669(2)

2.179(7)

2.178(7)

2.150(6)

2.069(6)

2.090(7)

1.760(7)

1.743(7)

M-C6

2.070(2)

2.063(2)

2.070(2)

2.043(2)

175.3(3)

2.168(7)

2.143(7)

2.129(7)

2.070(6)

2.097(7)

179.4(8)

M-C2

M-C7

M-C3

M-C8

M-C4

M-C9

M-C5

M-C10

M-Cp(1)^a

M-Cp(2)^b

Cp(1)^a-M-Cp(2)b

a

b

Cp(1) is the centroid of C1-C5. Cp(2) is the centroid of C6-C10.

Ta ble 9. Ar en e Sta ck in g In ter a ction s in Cr ysta llin e 11, 16, 23‚1/2C₆D₆, a n d 9

entry

complex

11

aryl (1)

aryl (2)

OO (Å)^a

OP (Å)^b

R (deg)^c

1.

2.

3.

4.

5.

6.

7.

8.

9.

C11-C16

C41-C46

C21-C26

C41-C46

C61-C66

C71-C76

C31-C36

C61-C66

C9-C14

C₆F₅

C91-C96

3.78

3.72

3.92

4.02

4.17

3.79

4.20

3.89

3.57

3.58

3.23, 3.44

3.32, 3.35

3.43, 3.64

3.45, 3.60

3.40, 3.83

3.78, 3.54

3.51, 3.61

3.72

11

12

9

C61-C66

C113-C118^d

C107-C112^d

C101-C106^d

C81-C86

6

14

17

0

5

16

C61-C66^e

C27-C29A^d

C₆F₅

23‚1/2C₆D₆

3.41, 3.48

3.28, 3.32

10.

9^f

a

b

The distance between the two aryl C₆centroids. The distance between an aryl C₆centroid and an aryl C₆least-squares plane. If the

two least-squares planes are not parallel, then two unique distances are calculated. ^cThe angle between the two C₆least-squares planes.

C₆D₆solvate molecule. ^eAryl (2) located on an adjacent molecule. ^fParameters calculated from data provided in ref 16.

d

F igu r e 7. Partial packing stereoview of 11‚4(C₆H₆), showing arene stacking motifs. H and F atoms omitted for clarity.

In the hexaarylated cobaltocene (16, Figure 8), the

metallocene core is slightly distorted, as shown by some

differences among the Co-C distances (Table 8), but not

enough to suggest a “slipped” bonding mode.²⁷The

conformations adopted by five of the C₆F₅substituents

and their respective Cp ligands are described by torsion

angles in the relatively narrow range of 45-57° (Table

6, entries 13 and 15-18), similar to the conformations

observed in the corresponding cyclopentadiene (entries

5 and 6). The remaining C₆F₅substituent (C21-C26)

shows a nearly perpendicular Cp-C₆F₅torsion angle of

about 84° (entry 14). Six of the “ortho” F atoms (F16,

F26, F32, F62, F72, and F82) encircle the cobalt atom

with Co-F distances ranging from 3.40 to 3.64 Å (Figure

9). Such a “belt” of fluorine atoms may partly account

for the astonishing stability of 19-electron complexes

such as 16 toward air and dilute aqueous acids.

Crystalline 16 exhibits two interesting arene-stacking

motifs (Table 9, entries 7 and 8), one intramolecular and

one intermolecular. According to an empirical guideline

that we established earlier,³⁵the C₆F₅-C₆F₅stacking

interactions in crystalline 11 and 16 do not exhibit

F igu r e 8. Thermal ellipsoid plot of 16 shown at 50%

probability. H and F atoms are numbered the same as the

C atoms to which they are attached.

5364 Organometallics, Vol. 19, No. 25, 2000

Thornberry et al.

F igu r e 11. Thermal ellipsoid plot of 26 shown at 50%

probability.

F igu r e 9. Partial molecular structure diagram for 16

showing a “belt” of aromatic fluorides circling the cobalt

atom. Co‚‚‚F distances are given in Å.

F igu r e 12. Thermal ellipsoid plot of 27 shown at 50%

probability. H and F atoms are numbered the same as the

C atoms to which they are attached.

Ta ble 10. Selected Bon d Dista n ces (Å) a n d Bon d

An gles (d eg) for th e Cp M(CO)₃Cor es of 23‚1/

2(C₆D₆) (M ) Mn ), 26 (M ) Re), a n d 27 (M ) Re)

23‚1/2(C₆D₆)

26

27

M-C_ring

2.174(3)

2.164(3)

2.130(3)

2.149(3)

2.139(4)

1.780(4)

1.795(4)

1.790(4)

1.155(5)

1.147(4)

1.152(5)

178.5(4)

177.3(3)

176.5(3)

2.314(8)

2.322(8)

2.285(8)

2.277(8)

2.318(7)

1.955(8)

1.922(9)

1.907(9)

1.964(9)

1.136(12)

1.158(11)

1.087(13)

177.8(8)

176.3(9)

175.1(8)

2.309(6)

2.331(6)

2.299(5)

2.302(6)

2.298(6)

1.968(6)

1.904(7)

1.915(9)

1.902(6)

1.141(9)

1.141(11)

1.138(9)

179.5(6)

177.3(6)

177.0(6)

F igu r e 10. Thermal ellipsoid plot of 23‚1/2(C₆D₆) (50%

probability) showing arene stacking of C₆D₆solvate mol-

ecule between the C21-C26 and C21A-C26A pentafluo-

rophenyl groups. H, D, and F atoms are numbered the

same as the C atoms to which they are attached.

M-Cp^a

M-C_CO

C-O

the degree of predictability needed for “crystal engi-

neering”.^16,28-32

Cr ysta l Str u ctu r es of P ia n o Stool Com p lexes.

When the manganese tricarbonyl complex 23 was

crystallized from C₆D₆, crystals of an adduct having the

empirical formula 23‚1/2C₆D₆were obtained (Figure

10). Solvent-free crystal structures of the rhenium

tricarbonyl complexes (26 and 27) are shown in Figures

11 and 12, respectively. Metric parameters of the CpM-

(CO)₃cores of these three complexes (Table 10) indicate

no significant structural distortions. The three M-Cp

distances fall well within the ranges (1.75 < Mn-Cp <

1.80 Å, 1.94 < Re-Cp < 1.96 Å) anticipated from a

search of the Cambridge Crystallographic Database for

analogous compounds. Likewise, the M-C-O param-

eters are not unusual. The torsion angles adopted by

the C₆F₅groups in these structures vary between 23°

and 66° (Table 6, entries 19-26). Not surprisingly, the

vicinal C₆F₅groups (entries 19, 20, 24, and 25) are less

M-C-O

a

Cp represents the centroid of the 5 Cp carbons.

coplanar compared with the “isolated” C₆F₅groups

(entries 21-23 and 26).

The C₆D₆molecule of 23‚1/2C₆D₆rests on a crystal-

lographic inversion center in a stacking arrangement

between two C₆F₅groups of 23. This stacking feature

(Table 9, entry 9) appears more ideal (shorter distances,

rings closer to parallel) than that observed in 11‚

4(C₆H₆). Surprisingly, the interaction between C₆F₅

groups and solvate C₆H₆(or C₆D₆) is not well docu-

mented in the Cambridge Crystallographic Database.

The clearest, most compact example in the database is

a “stack” analogous to that found in 23‚1/2C₆D₆, having

slightly longer centroid-centroid distances of 3.713 and

Cyclopentadienyl Complexes of Fe, Co, Mn, and Re

Organometallics, Vol. 19, No. 25, 2000 5365

mented.³⁴Evidently the oxygen lone pairs of the CO

ligands are much weaker donors than those previously

studied. In the case of 23‚1/2C₆D₆, an additional

CtO‚‚‚Cp(centroid) interaction is observed (Table 11,

entry 2). This interaction is relatively common, and the

metric details of the interaction observed for 23‚1/2C₆D₆

are normal.

F igu r e 13. Definition of parameters used to characterize

the intermolecular interactions of C₆F₅groups with CO

ligands.

Ta ble 11. In ter m olecu la r Don or -Accep tor

In ter a ction s of Ca r bon yl Liga n d s a n d Ar om a tic

Rin gs in Cr ysta llin e P ia n o Stool Com p lexes

23‚1/2C₆D₆, 26, a n d 27

OX OP

R

D

entry

complex

donor acceptor (Å)^a(Å)^a(deg)^a(Å)^a

1.

2.

3.

4.

5.

6.

23‚1/2C₆D₆

O6

O7

O1

O6

O7

O8

C9-C14

C1-C5

3.12 3.10

3.15 3.11

126

171

170

133

150

0.33

0.51

0.86

1.88

1.76

0.85

C-H‚‚‚F -C In ter a ction s. Earlier reports from our

laboratory³⁵and elswhere³⁶have attempted to establish

specific parametric correlations as criteria for defining

C-H‚‚‚F-C as “hydrogen bonds.” The sum of the van

der Waals radii of hydrogen and fluorine is in the range

2.5-2.7 Å.³⁷H‚‚‚N and H‚‚‚O hydrogen bonds generally

exhibit distances that are significantly shorter than the

sum of their respective van der Waals radii. Dunitz and

Taylor suggested that a similar criterion should apply

to H‚‚‚F, which implies that an H‚‚‚F distance less than

2.3 Å is necessary. In fact, 39 compounds having

C-H‚‚‚F-C contacts characterized by H‚‚‚F distances

less than 2.35 Å were found among 360 compounds

containing sp²-hybridized C-F bonds in the Cambridge

Crystallographic Database.³⁸However, interactions

among organic C-F and C-H groups must often coexist

with various other intermolecular packing motifs and

compete with stronger donors and acceptors, which

leads to a broad range of observed H‚‚‚F distances and

C-H‚‚‚F angles.³²Desiraju and Boese therefore ad-

vanced an alternative criterion of an inverse correlation

between H‚‚‚F distances and C-H‚‚‚F angles based on

23‚1/2C₆D₆

26

27

C15-C20 3.06 2.93

C9-C14

3.65 3.13

C21-C26 3.36 2.86

C21-C26 3.18 3.06

a

The parameters OX, OP, R, and D are defined in Figure 13.

3.719 Å, correspondingly longer centroid-plane dis-

tances of 3.614, 3.596, 3.704, and 3.709 Å, and inter-

planar angles of 3.4° and 5.3°.³³

All three piano stool complexes 23‚1/2C₆D₆, 26, and

27 exhibit directed, intermolecular interactions between

the partially negatively charged oxygen terminae of the

carbonyl ligands and the partially positively charged

centroids of crystallographically neighboring C₆F₅sub-

stituents. Metric parameters for these CO‚‚‚C₆F₅inter-

actions are defined in Figure 13 and summarized in

Table 11. These interactions are most striking in 27, in

which all three CO groups “contact” adjacent C₆F₅

groups (Figure 14). Of the five compounds in the

Cambridge Crystallographic Database that exhibit OX

distances less than 4.0 Å, only one shows an OX distance

less than those reported here. The correlation (r ) 0.87)

of OX and D shown in Figure 15(a) suggests that donors

approaching the centroid (decreasing OX) avoid van der

Waals interactions with the fluorine atoms around the

perimeter. However, poor correlation of OP with D

shown in Figure 15(b) reveals that donors approaching

the C₆F₅plane (decreasing OP) do not tend to “zero in”

on the central axis of the ring (decreasing D). Further-

more, neither F nor G, which we reported earlier, show

any CO‚‚‚C₆F₅interactions.¹⁴Although this concept

warrants further experimental (and theoretical) inves-

tigation, we must conclude for now that CO‚‚‚C₆F₅

interactions are not useful design elements for crystal

engineering. Well-organized interactions of perfluoro-

aromatic groups and lone pairs of electrons is docu-

a meticulous and exhaustive study of the fluorinated

32

benzenes C₆H_nF_6-n

.

We recently extended their findings to fluoroarylated

ferrocenes, in which C-H‚‚‚F-C interactions are a

persistent structural feature.³⁵This extension was

justified, because the C-H acidity of metallocenes is

comparable with aromatic hydrocarbons and because

the addition of the metal atom does not provide a

competing hydrogen bond acceptor. We further classified

as “purposeful” those C-H‚‚‚F-C interactions that have

H‚‚‚F distances less than 3.0 Å and that lie within the

inverse correlation demonstrated by Desiraju and Boese.

Those having H‚‚‚F distances less than 3.0 Å that

diverge from the correlation were termed “diffuse”.

Importantly, our notion that “diffuse” and “purposeful”

F igu r e 14. Partial packing stereoview of 27 showing CO‚‚‚C₆F₅interactions.

5366 Organometallics, Vol. 19, No. 25, 2000

Thornberry et al.

F igu r e 15. Correlations of metric parameters describing

CO‚‚‚C₆F₅interactions. Parameters OX, OP, and D are

defined in Figure 13. (+) Data from this work (Table 11);

(O) data from the Cambridge Crystallographic Database.

C-H‚‚‚F-C interactions can coexist in the same struc-

ture is not consistent with the Desiraju-Boese criterion

of an inverse parametric correlation over all observed

interactions. We assert that considerable allowance for

“diffuse” interactions is not only justified but necessary,

because the range of structural motifs and packing

arrangements found in fluorinated benzenes (which are

nearly isostructural) is not representative of all com-

pounds in which one might want to look for C-H‚‚‚F-C

hydrogen bonds. In particular, other compounds often

have extensive arrays of intramolecular C-H‚‚‚F-C

interactions, which tend to “saturate” hydrogen atoms,

leading to much weaker intermolecular C-H‚‚‚F-C

interactions. The search for a more general understand-

ing of C-H‚‚‚F-C interactions is made much more

difficult by these complications.

F igu r e 16. Scatterplot of C-H‚‚‚F angles and H‚‚‚F

distances describing intermolecular C-H‚‚‚F-C interac-

tions in C₆F₅-substituted (cyclopentadienyl)metal com-

plexes. (b) Data for 2-i; (*) data for 2-ii; (4) data for 4; (×)

data for 11; (+) data for 16; (0) data for 23; (9) data for

27; (O) data for fluorinated benzenes from ref 32.

This section presents an analysis of the present

crystal data in terms of these principles. We conducted

an exhaustive search of all eight structures reported

here for C-H‚‚‚F-C interactions having H‚‚‚F distances

less than 3.0 Å and C-H‚‚‚F angles greater than 100°

(no angle restriction was enforced for intramolecular

interactions). A complete table of all interactions is

included in the Supporting Information (Tables S-65

and S-66). Figure 16 shows a scatterplot, in which the

data reported by Desiraju and Boese are represented

by small, unfilled circles. We added an elliptical region

to encompass their data (only one point was excluded)

and highlight the inverse correlation they proposed as

a criterion for hydrogen bonding. The plot is dominated

by 34 interactions (× symbols) observed among the C-F

bonds and solvated benzene molecules in 11‚4(C₆H₆).

Although many of the new points in Figure 16 lie within

the elliptical “correlation” region, the correlation be-

tween H‚‚‚F distances and C-H‚‚‚F angles for each

individual compound is relatively poor. In fact, each of

our compounds represented in the plot has a third to

half of its respective points outside the shaded region.

However, to pack into stable crystal structures, these

irregularly shaped molecules cannot optimize all

C-H‚‚‚F-C interactions. We conclude that the strict

requirement of a inverse distance-angle correlation is

not necessarily appropriate for irregularly shaped mol-

ecules.

F igu r e 17. Stacked plot of ¹⁹F NMR spectra of 27 obtained

at several temperatures in THF-d₈at 470 MHz.

Dyn a m ic NMR Stu d ies. Complexes 11, 12, and 27

were examined by variable-temperature ¹H and ¹⁹F

NMR spectroscopy in order to investigate their solution

conformational dynamics. In the ¹⁹F NMR spectrum of

27 (Figure 17), six signals were assigned to the ortho

(-130 to -145 ppm), meta (-160 to -165 ppm), and

para (-152 to -157) positions, respectively. At 25 °C,

the presence of two signals (4:2 ratio) for the ortho

fluorines signifies first that the two vicinal C₆F₅groups

are related by time-averaged molecular C_ssymmetry

and, second, that the endo and exo “edges” of the aryl

groups are in the fast-exchange regime of the experi-

ment. As temperature decreases, endo and exo ortho

fluorines of the vicinal C₆F₅groups decoalesce (T_coal

≈

-50 °C), whereas the “isolated” C₆F₅group remains in

fast exchange with respect to Cp-C₆F₅rotation at -100

°C. Using eq 1 (∆ν ) 4920 Hz at -100 °C), a first-order

rate constant of exchange k_coal(at -50 °C) of 10 800 Hz

is obtained for Cp-C₆F₅rotation of the vicinal C₆F₅

groups. Temperature-dependent rate data obtained from

line shape analysis³⁹of the slow (T e -60 °C) and fast

exchange regimes (eq 2 and eq 3, respectively) enable

(36) (a) Grepioni, F.; Cojazzi, G.; Draper, S. M.; Scully, N.; Braga,

D. Organometallics 1998, 17, 296. (b) Dunitz, J .; Taylor, R. Chem. Eur.

J . 1997, 3, 89. (c) Plenio, H. Chem. Rev. 1997, 97, 3363.

(37) (a) Bondi, A. J . Phys. Chem. 1964, 68, 441. (b) Rowland, R. S.;

Taylor, R. J . Phys. Chem. 1996, 100, 7384.

(38) Howard, J . A. K.; Hoy, V. J .; O’Hagan, D.; Smith, G. T.

Tetrahedron 1996, 52, 12613.

(39) Hore, P. J . Nuclear Magnetic Resonance; Oxford Science: New

York, 1995; pp 44-47.

Cyclopentadienyl Complexes of Fe, Co, Mn, and Re

Organometallics, Vol. 19, No. 25, 2000 5367

F igu r e 18. Arrhenius plot of C₆F₅-Cp rotational rate

constants for 27.

F igu r e 20. Conformational analyses of hexasubstituted

metallocenes. Notation is described in the text.

F igu r e 19. Stacked plot of ¹⁹F NMR spectra of 11 obtained

at several temperatures in THF-d₈at 470 MHz.

the construction of an Arrhenius plot (Figure 18), from

which we obtain an activation energy (E_a) of 9(1) kcal

mol^-1for the rotation of adjacent C₆F₅groups in 27. The

corresponding activation parameters, ∆H^q) 8(1) kcal

mol^-1and ∆S^q) -2(5) cal mol^-1K^-1, obtained from an

Eyring treatment agree qualitatively with a semiem-

pirical (PM3) estimate (∆H^q) 4.2 kcal mol^-1) carried

out on 1,2-bis(pentafluorophenyl)cyclopentadienyl an-

ion.

k_coal) (π/x2)∆ν ) 2.22∆ν

k(T) ) π[fwhm(T) - fwhm(natural)]

(1)

(2)

k(T) ) π(∆ν)²/[fwhm(T) - fwhm(natural)] (3)

F igu r e 21. Stacked plot of ¹H NMR spectra of 11 obtained

at several temperatures in THF-d₈at 500 MHz.

Analysis of 11 by variable-temperature ¹⁹F NMR

spectroscopy reveals two dynamic processes. First, the

Cp-C₆F₅rotation of the “adjacent” C₆F₅groups (as in

27, discussed above) is evident from decoalescence of

the signal at -137 ppm in the spectra obtained at T e

-50 °C (Figure 19). The second dynamic process in-

volves relative rotation of the two Cp-Fe bonds⁴⁰and

is evident from the signal at -155 ppm (4 F), assigned

to the para fluorines of the vicinal C₆F₅substituents.

At 25 °C, these four fluorines are interrelated by time-

averaged C_2hsymmetry, resulting in a single resonance.

Decoalescence into two equally integrating resonances

just above -50 °C shows that the ground state is

represented by one of four canonical, diastereomeric

structures (Figure 20, b-e), in which adjacent C₆F₅

groups on the same Cp ligand are inequivalent (loss of

σ symmetry) while retaining C₂relationships with

counterparts on the opposite Cp ligand. Exchange of the

two inequivalent C₆F₅groups is a formal racemization

process (Figure 20, b a j, c a i, d a h , or e a g),

suggesting the use of a degenerate, two-site exchange

model.

Slowing the Cp-Cp rotation in 11 also leads to two

1

signals in the H NMR spectrum below about -50 °C

(Figure 21). An Arrhenius plot (Figure 22) resulting

from line shape analysis (eqs 2, 3) of the spectra shown

in Figure 21 provides an activation energy (E_a) of 11(2)

kcal mol^-1for this Cp-Cp rotation process with corre-

sponding activation parameters (∆H^q) 11(2) kcal mol^-1

and ∆S^q) -2(5) cal mol^-1K^-1) obtained from an Eyring

plot. Similar analysis of the ¹⁹F NMR signals at -155

(40) (a) Okuda, J . Top. Curr. Chem. 1991, 160, 97. (b) J aniak, C.;

Schumann, H. Adv. Organomet. Chem. 1991, 33, 291.

5368 Organometallics, Vol. 19, No. 25, 2000

Thornberry et al.

F igu r e 22. Arrhenius plot of Cp-Cp rotational rate

constants for 11.

F igu r e 24. Arrhenius plot of Cp-C₆F₅rotational rate

constants for 12.

on the “internal” phenyl groups at room temperature,

with coalescence at -65 °C and ∆G^q_c) 9(1) kcal mol^{-1 45}

.

Crude semiempirical modeling (PM3) predicted enthal-

pies of activation of 5.7 kcal mol^-1for 1,2,3-triphenyl-

cyclopentadienyl anion and 13.5 kcal mol^-1for 1,2,3-

tris(pentafluorophenyl)cyclopentadienyl anion. These

findings illustrate the different steric effects of hydrogen

and fluorine on aryl substituents.

The decoalescence of the signal assigned to the four

para fluorines of the “outer” C₆F₅groups of 12 (Figure

23, -100 °C spectrum, signal at -156 ppm) prohibits

strict C_2hsymmetry in the ground state and suggests a

degenerate exchange process similar to that observed

for 11. Unfortunately, no estimate can be made of the

activation barriers to Cp-Cp rotation in 12. Although

decoalescence occurs at a lower temperature in 12 than

in 11, the value of ∆ν needed for eq 1 requires measure-

ments in the slow-exchange spectral regime, which was

F igu r e 23. Stacked plot of ¹⁹F NMR spectra of 12 obtained

at several temperatures in THF-d₈at 470 MHz.

ppm (Figure 19) corroborates this value. This activation

energy⁴¹is comparable with those measured for [1,2,4-

(Me₃Si)₃C₅H₂]₂Fe (11.0 kcal mol^{-1 42}

) and [1,2,4-(Me₃C)₃-

1

inaccessible experimentally. The H NMR spectra ob-

C₅H₂]₂Fe (15.3 kcal mol^-1),⁴³but much larger than

tained at -100 °C show no line-broadening that we

could attribute to chemical exchange, which is not

surprising, as the spectral dispersion of ¹H NMR is

much less than that of ¹⁹F NMR.

ferrocene itself (1.8-2.3 kcal mol^-1).⁴⁴

Like 11, the isomeric hexaarylated ferrocene 12

exhibits two dynamic processes in THF solution. First,

rotation of the “outer” C₆F₅groups about their respective

C₆F₅-Cp bonds was evident from variable-temperature

¹⁹F spectroscopy (Figure 23), and line shape analysis of

the signals at -137 ppm led to an Arrhenius plot

(Figure 24), which yielded an activation energy (E_a) of

10(1) kcal mol^-1. Eyring analysis of the same data gave

Okuda proposed a simple but highly predictive strain-

enthalpic quantifier to understand Cp-Cp rotational

kinetics in substituted metallocenes (Figure 20).⁴²Each

eclipsing interaction (in eclipsed or “e” conformations)

is assigned a value of +1 arbitrary enthalpy units, and

each gauche interaction (in staggered or “s” conforma-

tions) is assigned a value of +0.5. Combining Okuda’s

conformational model (Figure 20) with the crystal-

lographic data for 11 and 16, and the NMR data for 11

and 12, provides some insight into these dynamic

processes while pointing to a potential limitation of

Okuda’s model for C₆F₅substituents. An Okuda dia-

gram for 11 (Figure 20) suggests that c/i (a pair of

enantiomeric rotamers) should be the ground state of

11, and indeed this is observed in [1,2,4-(Me₃Si)₃C₅H₂]₂-

Fe,⁴²whereas the crystallographic data presented above

for 11 (Figure 6) argue for e/g instead. Okuda’s model

also predicts either o/q or p for the ground conforma-

tional state of 12, but symmetry arguments (based on

NMR data) presented above rule out p . Although we

were unable to obtain crystals of 12, the corresponding

hexaarylated cobaltocene (16) adopts a structure about

halfway between o and p (Figure 8).

∆H^q) 9(1) kcal mol^-1and ∆S^q) -1(5) cal mol^-1K^-1

.

This value is within experimental error of the analogous

process observed in 27. The “middle” C₆F₅groups of 12

do not rotate about their C₆F₅-Cp bonds at a rate that

we could detect by ¹⁹F NMR spectroscopy. No change

in line shape of the diastereotopic ortho ¹⁹F resonances

was observed at 80 °C in 1,4-dioxane solution. The same

“rotational locking” of the “middle” C₆F₅group was also

observed in all the other complexes prepared with this

ligand (16, 20, 24, and 28). These results contrast

sharply with related studies in which (Ph₄Cp)₂Fe un-

derwent rapid exchange of endo and exo phenyl protons

(41) In some of the prior literature, E_aand ∆G^qappear to have been

used interchangeably. We have cited only primary sources rather than

review sources for the data presented here.

(42) Okuda, J .; Herdtweck, E. Chem. Ber. 1988, 121, 1899.

(43) Abel, E. W.; Long, N. J .; Orrell, K. G.; Osborne, A. G.; Sˇik, V.

J . Organomet. Chem. 1991, 403, 195.

(44) (a) Holm, C. H.; Ibers, J . A. J . Chem. Phys. 1959, 30, 885. (b)

Mulay, L. N.; Attalla, A. J . Am. Chem. Soc. 1963, 85, 702. (c) Makova,

M. K.; Leonova, E. V.; Karimov, Y. S.; Kochetokova, N. S. J . Organomet.

Chem. 1973, 55, 185.

(45) Castellani, M. P.; Wright, J . M.; Geib, S. J .; Rheingold, A. L.;

Trogler, W. C. Organometallics 1986, 3, 1116.

Cyclopentadienyl Complexes of Fe, Co, Mn, and Re

Organometallics, Vol. 19, No. 25, 2000 5369

According to Okuda’s enthalpic descriptors, 11 and

12 should undergo facile exchange processes (e a g and

o a q, respectively) with negligible enthalpic barriers.

It is important to note that Okuda’s model is derived

from analysis of tert-butyl and trimethylsilyl substitu-

ents, which are much more isotropic with respect to

Cp-R rotations than C₆F₅groups. One might therefore

expect that the anisotropic steric profile of C₆F₅groups

would facilitate Cp-Cp rotations, because transannular

steric strain could be relieved considerably by adopting

Cp-C₆F₅torsional angles closer to 0° (coplanar rings).

Indeed, detailed analysis of a three-dimensional model

of 11 suggests that the exchange e a g requires

synchronized (geared) rotations of the two pairs of

stacked C₆F₅groups. One member of each of these pairs

(Figure 6) is a “vicinal” C₆F₅group, for which the

enthalpic barrier to Cp-C₆F₅rotation is about 8 kcal

mol^-1based on our analysis of the rhenium complex 27.

Two such rotations would account for more than the

observed ∆H^qfor Cp-Cp rotation in 11 (11 kcal mol^-1).

In contrast, the exchange o a q proceeds with all C₆F₅

groups turning about their respective C₆F₅-Cp bonds

through 90° torsion angles. Semiempirical (PM3) cal-

culations on the 1,2-bis(pentafluorophenyl)cyclopen-

tadienyl and 1,2,3-tris(pentafluorophenyl)cyclopenta-

dienyl anions predict a much smaller barrier (∆H^q< 2

kcal mol^-1) for this process. We would therefore expect

12 to undergo Cp-Cp rotation between the two enan-

tiomeric ground-state rotamers to occur more rapidly.

These findings suggest the need to couple Okuda’s

model for transannular steric strain with the confor-

mational strain arising within a single ligand in order

to arrive at a complete description of the fluxionality of

perfluoroarylated metallocenes. The possibility of cor-

relation among torsional degrees of freedom has been

raised by others for related systems.⁴⁶

rocene (H), which indeed shows pairwise, intramolecular

stacking of all four C₆F₅groups.⁴⁷

Con clu sion s. Up to three pentafluorophenyl (C₆F₅)

substituents are readily attached to cyclopentadiene by

nucleophilic aromatic substitution reactions of cyclo-

pentadienyl anions with hexafluorobenzene. Cyclopen-

tadienyl anions bearing one, two, or three C₆F₅groups

serve as ligands to form stable, well-characterized

ferrocenes and cobaltocenes, as well as piano stool

complexes of manganese and rhenium. The electron-

withdrawing ability of the C₆F₅substituents is evident

from large shifts in the E° of oxidation of substituted

ferrocenes and cobaltocenes and in the carbonyl stretch-

ing wavenumbers of substituted Mn and Re piano stool

complexes. The ability of C₆F₅groups to stabilize 19-

electron complexes is illustrated for cobaltocene; with

four or more C₆F₅substitutents, cobaltocene is stable

to air in the presence of dilute hydrochloric acid. C₆F₅-

substituted (cyclopentadienyl)metal complexes exhibit

pervasive structural motifs: orderly arene stacking,

close C₆F₅‚‚‚OtC contacts, and diffuse C-H‚‚‚F-C

interactions. Conformational dynamics in ferrocenes

bearing multiple C₆F₅substituents are described by a

qualitative model in which transannular steric interac-

tions governing Cp-Fe-Cp rotational barriers are

synchronized (coupled) with Cp-C₆F₅torsions, resulting

in an overall enthalpic barrier to Cp-Fe-Cp that is

similar to corresponding Me₃Si-substituted ferrocenes.

The final inconsistency to address is the observation

of e/g in crystalline 11 instead of the c/i conformation

predicted by the Okuda diagram (Figure 20). Possibly,

the positive enthalpy of an eclipsed interaction is

reversed by attractive arene stacking forces. A distinct

preference for arene stacking in 1,1′-bis(perfluoroaryl)-

ferrocene complexes is already firmly established.^16,35

In that case, one might expect conformers a and k to

be the preferred ground-state structures of 11 and 12,

respectively. Examination of three-dimensional models

suggest that pairs of vicinal C₆F₅groups cannot engage

in ideal arene stacking, however, because of steric strain

between ortho F atoms on vicinal C₆F₅groups. This

rationalizes the crystallographic observation of con-

former o (a single pair of stacked C₆F₅groups) instead

of m /s (two stacking pairs) or k (three stacking pairs)

in the hexaarylated cobaltocene (16). Of course, a crystal

structure of 10 would address this hypothesis, but so

far we have been unable to obtain these data. The most

closely related complex that we have been able to

characterize crystallographically is the methylated fer-

Ack n ow led gm en t. J . L. Montgomery, an under-

graduate research participant, provided some of the

electrochemical analyses of arylated cobaltocenes. We

thank T. Glass for assistance with NMR analyses.

Fruitful discussions with Professors R. P. Hughes

(Dartmouth College), T. W. Hanks (Furman University),

and J . R. Sowa (Seton Hall University), are gratefully

acknowledged. This project was supported by the Pe-

troleum Research Fund (30738-G) and the Virginia Tech

Chemistry Department.

Su p p or tin g In for m a tion Ava ila ble: Tables (S1-S64) of

crystallographic data for 2-i, 2-ii, 4, 11‚4(C₆H₆), 16, 23‚1/

2(C₆D₆), 26, and 27; tables of C-H‚‚‚F-C interactions (S65,

S66); two figures (S1, S2) showing semiempirical modeling

results, and a detailed procedure for the preparation of 1. This

information is available free of charge on the Internet at

http://pubs.acs.org.

OM000798V

(47) Unpublished results from our laboratories.

(48) (a) Robbins, J . L.; Edelstein, N.; Spencer, B.; Smart, J . C. J .

Am. Chem. Soc. 1982, 104, 1882. (b) Koelle, U.; Khouzami, F. Angew.

Chem., Int. Ed. Engl. 1980, 19, 640.

(46) (a) Gupta, H. K.; Stradiotto, M.; Hughes, D. W.; McGlinchey,

M. J . J . Org. Chem. 2000, 65, 3652. (b) Iwamura, H.; Mislow, K. Acc.

Chem. Res. 1988, 21, 175.

(49) (a) Benson, I. B.; Knox, S. A. R.; Stansfield, R. F. D.; Woodward,

P. J . Chem. Soc., Dalton Trans. 1981, 51. (b) Casey, C. P.; Rutter, E.

W. Inorg. Chem. 1990, 29, 2333.

Article Doi

Structural and electronic effects of pentafluorophenyl substituents on cyclopentadienyl complexes of Fe, Co, Mn, and Re

DOI: 10.1021/om000798v

Source and publish data:

Authors:

Article abstract of DOI:10.1021/om000798v

Full text of DOI:10.1021/om000798v

Products guided by the article

R&D Labs maybe for 320425-50-1

Relevant to this article

Hot Product