Through-space nuclear spin-spin JPP coupling in tetraphosphine ferrocenyl derivatives: A 31P NMR and x-ray structure correlation study for coordination complexes

Article abstract of DOI:10.1021/ja048907a

Herein, we report on ³¹P³¹P solution-phase through-space nuclear spin-spin coupling constants (JPP) from a novel family of organometallic tetraphosphine nickel and palladium complexes. These JPP constants were accurately determined through NMR iterative simulation based on the second-order spectra obtained for the compounds. The corresponding solid-state X-ray structures of the complexes were determined, and the through-space P...P distances are reported. Due to the blocked conformation of the species in solution, a qualitative and semiquantitative experimental correlation is obtained, which links the geometric parameters and the intensity of the corresponding P-P coupling constant. The lone-pair overlap theory developed for ¹⁹F¹⁹F and ¹⁵N ¹⁹F through-space couplings in organic compounds [J. Am. Chem. Soc. 1973, 95, 7747-7752; 2000, 122, 4108-4116] appears to be a reliable foundation on which to account for our results. Based on the reported observations, the lone-pair overlap model is extended to through- space ³¹P³¹P coupling, and the model is broadened to encompass metal orbital contributions for coordination complexes. Some of the predictions and consequences of the proposed theory are discussed.

Full text of DOI:10.1021/ja048907a

Published on Web 08/11/2004
“Through-Space” Nuclear Spin-Spin JPP Coupling in
Tetraphosphine Ferrocenyl Derivatives: A 31P NMR and X-ray
Structure Correlation Study for Coordination Complexes
Jean-Cyrille Hierso,*,† Aziz Fihri,† Vladimir V. Ivanov,† Bernard Hanquet,†
Nadine Pirio,† Bruno Donnadieu,‡ Bertrand Rebie`re,† Re´gine Amardeil,† and
Philippe Meunier†
Contribution from the Laboratoire de Synthe`se et d’Electrosynthe`se Organome´talliques,
UMR-CNRS 5188, UniVersite´ de Bourgogne, Faculte´s des Sciences Mirande,
9 AVenue Alain SaVary, BP 47870 F-21078 Dijon Cedex, France, and Laboratoire de Chimie de
Coordination du CNRS, 256, Route de Narbonne, 31077 Toulouse Cedex 4, France
Received February 26, 2004; E-mail: jean-cyrille.hierso@u-bourgogne.fr
Abstract: Herein, we report on 31P31P solution-phase “through-space” nuclear spin-spin coupling constants
(JPP) from a novel family of organometallic tetraphosphine nickel and palladium complexes. These JPP
constants were accurately determined through NMR iterative simulation based on the second-order spectra
obtained for the compounds. The corresponding solid-state X-ray structures of the complexes were
determined, and the “through-space” P‚‚‚P distances are reported. Due to the blocked conformation of the
species in solution, a qualitative and semiquantitative experimental correlation is obtained, which links the
geometric parameters and the intensity of the corresponding P‚‚‚P coupling constant. The lone-pair overlap
theory developed for 19F19F and 15N19F “through-space” couplings in organic compounds [J. Am. Chem.
Soc. 1973, 95, 7747-7752; 2000, 122, 4108-4116] appears to be a reliable foundation on which to account
for our results. Based on the reported observations, the lone-pair overlap model is extended to “through-
space” 31P31P coupling, and the model is broadened to encompass metal orbital contributions for coordination
complexes. Some of the predictions and consequences of the proposed theory are discussed.
t
Scheme 1. Tetraphosphine Fc(P)4 Bu (a) and the Related
Compound 1,1′,2,2′-Tetrakis(diphenylphosphino)-3,3′,4,4′,5,5′-
hexamethylferrocene (b): General View (Top) and View from the
Top (Bottom)
1. Introduction
Interest in the organometallic and coordination chemistry of
multidentate phosphorus-containing organic species has been
ongoing for over 50 years.1,2 New families of useful and
intriguing polyphosphines are developed every year, which
display interesting fundamental structural features as well as
applications.3,4Among these new polyphosphine compounds, our
group has reported on the synthesis,5 coordination chemistry,6
and catalytic properties7 of the conformationally blocked
t
ferrocenyl tetraphosphine Fc(P)4 Bu (1,1′,2,2′-tetrakis(diphen-
ylphosphino)-4,4′-di-tert-butylferrocene, Scheme 1a). These
† Universite´ de Bourgogne.
‡ Laboratoire de Chimie de Coordination du CNRS.
(1) Cotton, F. A.; Hong, B. Prog. Inorg. Chem. 1992, 40, 179.
(2) Hierso, J.-C.; Amardeil, R.; Bentabet, E.; Broussier, R.; Gautheron, B.;
Meunier, P.; Kalck, P. Coord. Chem. ReV. 2003, 236, 143.
(3) For triphosphines, see: (a) Sto¨ssel, P.; Heins, W.; Mayer, H. A.; Fawzi,
R.; Steimann, M. Organometallics 1996, 15, 3393. (b) Sto¨ssel, P.; Mayer,
H. A.; Maichle-Mo¨ssmer, C.; Fawzi, R.; Steimann, M. Inorg. Chem. 1996,
35, 5860. (c) Sto¨ssel, P.; Mayer, H. A.; Auer, F. Eur. J. Inorg. Chem. 1998,
37.
(4) For a tetraphosphine, see: (a) Feurstein, M.; Laurenti, D.; Doucet, H.;
Santelli, M. Chem. Commun. 2001, 43. (b) Laurenti, D.; Feurstein, M.;
Pe`pe, G.; Doucet, H.; Santelli, M. J. Org. Chem. 2001, 66, 1633.
(5) (a) Broussier, R.; Bentabet, E.; Amardeil, R.; Richard, P.; Meunier, P.;
Kalck, P.; Gautheron, B. J. Organomet. Chem. 2001, 637-639, 126. (b)
Broussier, R.; Ninoreille, S.; Bourdon, C.; Blaque, O.; Ninoreille, C.;
Kubicki, M. M.; Gautheron, B. J. Organomet. Chem. 1998, 561, 85.
(6) Bentabet, E.; Broussier, R.; Amardeil, R.; Hierso, J.-C.; Richard, P.; Fasseur,
D.; Gautheron, B.; Meunier, P. J. Chem. Soc., Dalton Trans. 2002, 2322.
(7) Hierso, J.-C.; Fihri, A.; Amardeil, R.; Meunier, P.; Doucet, H.; Santelli,
M.; Donnadieu, B. Organometallics 2003, 22, 4490.
studies have highlighted the efficiency of this triarylphosphine
ligand in conjunction with palladium in Suzuki and Heck cross-
t
coupling reactions.7 The rotational rigidity of Fc(P)4 Bu, which
9
10.1021/ja048907a CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 11077-11087
11077

A R T I C L E S
Scheme 2. Complexes of Nickel and Palladium Synthesized from Fc(P)4 Bua
Hierso et al.
t
a In solution, the complexes 1 and 1b display a paramagnetic temperature-dependent behavior not detected in the 31P NMR spectrum of compounds 2, 2b,
3, and 4.
has not been detected in other analogous bischelating tetra-
plexes, four novel mononuclear and dinuclear nickel complexes
t
t
phosphines (Scheme 1b),5b is suspected to play an important
role in its activity as an auxiliary in palladium-catalyzed
reactions.7,8 In the course of our NMR-mechanistic studies
dealing with the organometallic chemistry of systems involving
[NiCl2{Fc(P)4 Bu}] (1) and [Ni2Cl4{Fc(P)4 Bu}] (2) (and 1b,
2b where bromine atoms substitute chlorines) were synthesized
and fully characterized by X-ray crystallography. These extend
t
on the previously reported [PdCl2{Fc(P)4 Bu}] (3) and [Pd2Cl4-
t
t
Fc(P)4 Bu and palladium allylic species, new palladium com-
{Fc(P)4 Bu}] (4) congeners.
plexes (3 and 4 in Scheme 2) with puzzling 31P NMR spectra
were obtained.7 The most reasonable explanation to account for
the observed NMR patterns was the existence of a variety of
JPP “through-space” nuclear spin-spin couplings not reported
in the coordination chemistry literature up to now.
To date, the observation of “through-space” interactions has
been limited to purely organic,9 mostly fluorine-containing,10
organic species as well as correlations based on theoretical
reports.9e,11,12
The conformational rigidity of all of these species allows us
to provide a correlation for phosphorus-containing complexes
between geometric features in the solid state, and “through-
space” spin-spin JPP coupling constants in solution. The present
work aims at completing and extending the current theory that
concerns (JFF) and (JFN) “through-space” coupling in organic
species to the (JPP) spin-spin interaction in coordination
compounds.
2. Experimental Section
To generate a dataset large enough to test our hypothesis
concerning “through-space” interactions in coordination com-
General Considerations. All reactions and workup procedures were
performed under an inert atmosphere of argon using conventional
vacuum-line and Schlenk techniques. Toluene, pentane, hexane, and
THF were degassed and distilled by refluxing over sodium benzophe-
none under argon. Dichloromethane was refluxed on calcium hydride.
CDCl3 and CD2Cl2 were degassed and stored over molecular sieves
t
(8) Fc(P)4 Bu is active as an auxiliary in Pd-catalyzed allylic amination:
Urrutigo¨ıty, M.; Hierso, J.-C.; et al., to be published.
(9) For JPP coupling constants, see: (a) McFarlane, H. E. C.; McFarlane, W.
Polyhedron 1988, 7, 1875. (b) McFarlane, H. E. C.; McFarlane, W.
Polyhedron 1999, 18, 2117. For a JPSe coupling, see: (c) Karac¸ar, A.;
Freytag, M.; Tho¨nnesen, H.; Omelanczuk, J.; Jones, P. G.; Bartsch, R.;
Schmutzler, R. Z. Anorg. Allg. Chem. 2000, 626, 2361. For JPF coupling
constants, see: (d) Hughes, R. P.; Laritchev, R. B.; Williamson, A.;
Incarvito, C. D.; Zakharov, L. N.; Rheingold, A. L. Organometallics 2002,
21, 4873 and ref 10h. For a general review, see: (e) Contreras, R. H.;
Peralta, J. E. Prog. Nucl. Magn. Reson. Spectrosc. 2000, 37, 321.
(10) For leading references on JFF, see: (a) Mallory, F. B. J. Am. Chem. Soc.
1973, 95, 7747. (b) Mallory, F. B.; Mallory, C. W.; Fedarko, M.-C. J. Am.
Chem. Soc. 1974, 96, 3536. (c) Mallory, F. B.; Mallory, C. W.; Ricker, W.
M. J. Am. Chem. Soc. 1975, 97, 4770. (d) Mallory, F. B.; Mallory, C. W.;
Ricker, W. M. J. Org. Chem. 1985, 50, 457. (e) Mallory, F. B.; Mallory,
C. W. J. Am. Chem. Soc. 1985, 107, 4816. (f) Mallory, F. B.; Mallory, C.
W.; Baker M. B. J. Am. Chem. Soc. 1990, 112, 2577. (g) Mallory, F. B.;
Luzik, E. D., Jr.; Mallory, C. W.; Carroll, P. J. J. Org. Chem. 1992, 57,
366. (h) Mallory, F. B.; Mallory, C. W.; Butler, K. E.; Lewis, M. B.; Xia,
A. Q.; Luzik, E. D., Jr.; Fredenburgh, L. E.; Ramanjulu, M. M.; Van, Q.
N.; Francl, M. M.; Freed, D. A.; Wray, C. C.; Hann, C.; Nerz-Stormes,
M.; Carroll, P. J.; Chirlian, L. E. J. Am. Chem. Soc. 2000, 122, 4108 and
references therein.
1
under argon prior to NMR use. Elemental analyses, H (300.13 and
500.13 MHz), 31P (121.49 and 202.46 MHz), and 13C NMR (75.47
and 125.77 MHz), including low-temperature NMR experiments, were
performed in our laboratories (respectively, on Bruker 300 and DRX
500 spectrometers).
(11) For a review, see: (a) Contreras, R. H.; Facelli, J. C. Ann. Rep. NMR
Spectrosc. 1993, 27, 255 and references therein. (b) Diz, A. C.; Contreras,
R. H.; Natiello, M. A.; Gavarini, H. O. J. Comput. Chem. 1985, 6, 647. (c)
Peralta, J. E.; Barone, V.; Contreras, R. H.; Zaccari, D. G.; Snyder, J. P. J.
Am. Chem. Soc. 2001, 123, 9162.
(12) (a) Arnold, W. D.; Mao, J.; Sun, H.; Oldfield, E. J. Am. Chem. Soc. 2000,
122, 12164. (b) Barfield, M.; Della, E. W.; Pigou, P. E.; Walter, S. R. J.
Am. Chem. Soc. 1982, 104, 3550. (c) Nakanishi, W.; Hayashi, S.; Toyota,
S. J. Org. Chem. 1998, 63, 8790.
9
11078 J. AM. CHEM. SOC. VOL. 126, NO. 35, 2004

Coupling in Tetraphosphine Ferrocenyl Derivatives
The synthesis and full characterization of the ligand Fc(P)4 Bu, and
A R T I C L E S
t
procedures). Anal. Calcd for C66H62Br4P4FeNi2 (1471.95): C, 53.9; H,
t
t
1
of the complexes [PdCl2{Fc(P)4 Bu}] (3) and [Pd2Cl4{Fc(P)4 Bu}] (4),
4.2. Found: C, 53.6; H, 3.9. H NMR (CD2Cl2): δ ) 8.60-6.40 (m,
are reported elsewhere.7
40H, Ph), 4.49 (s, 2H, Cp), 4.14 (s, 2H, Cp), 0.78 (s, 18H, Bu).
t
31P{1H}(CD2Cl2): δ ) 41.6 (q, JAB ) 65 Hz). 13C{1H}(CD2Cl2): δ )
t
[NiCl2{Fc(P)4 Bu}], 1. A mixture of 1,1′,2,2′-tetrakis(diphenylphos-
t
t
30.6 (s, 6C, BuCH3), 30.8 (s, 2C, BuCCH3), 65.5 and 69.3 (d, 2C
each, CpCH, JCP ) 8.0 and 9.0 Hz), 119.2 (s, 2C, CpCtBu), 127.0-
135.0 (m, 40C, C6H5). The lack of solubility prevents the unambiguous
phino)-4,4′-di-tert-butylferrocene (855 mg, 0.83 mmol) and NiCl2‚DME
(182 mg, 0.83 mmol) was refluxed in THF (15 mL) for 4 h. From the
cooled reaction mixture, complex 1 was filtered off as an orange-red
precipitate, washed with THF, and dried in vacuo (yield 762 mg, 83%).
An analytically pure sample was prepared by recrystallization of the
crude product from a CH2Cl2/hexane mixture. Anal. Calcd for C66H62-
Cl2P4FeNi (1164.55): C, 68.1; H, 5.4. Found: C, 68.0; H, 5.2. 1H NMR
(CDCl3): δ ) 8.60-6.50 (m, 40H, Ph), 4.73, 4.67, 4.26, 4.21 (s, 1H
2
assignment of most of the quaternary carbons.
Cristallographic Data for Compounds 1b and 2. For compound
1b, data were collected on a Nonius Kappa CDD (Mo KR) diffracto-
meter at 110 K. The structure was solved by a Patterson search program
and refined by full-matrix least-squares methods based on F2 using
SHELX97 with the help of the WinGX program (Universite´ de
Bourgogne). For compound 2, data were collected at 180 K on a IPDS
STOE diffractometer (Mo KR). The structure was solved by Direct
Methods and refined by full-matrix least-squares methods based on F2
using SHELX97 with the help of the WinGX program (LCC Toulouse).
t
each, Cp), 0.81, 0.71 (s, 9H each, Bu). 31P{1H}(CDCl3, 298 K): δ )
31.8, 27.9 (m, very broad, P-Ni each), -18.4 (m, very broad, pendant
PPh2), -25.9 (s, pendant PPh2). 31P{1H}(CD2Cl2, 193 K): δ ) 33.8
(dd, JP1P4 ) 78 Hz, JP1P2 ) 23 Hz), 29.8 (d, JP4P1 ) 78 Hz), -21.7 (d,
JP2P1 ) 23 Hz), -26.5 (s). 13C{1H}(CDCl3): δ ) 30.8 and 31.4 (s, 1C
each, tBuCCH3), 31.8 and 31.9 (s, 3C each, tBuCH3), 66.2 and 70.1 (s,
3. Results and Discussion
2
1C each, CpCH,), 71.5 and 76.3 (d, 1C each, JCP ) 4.5 Hz, CpCH),
81.4 (d, 2C, 1JCP ) 24 Hz, CpCP), 89.2 (m, 2C, CpCP), 108.8 (s, 1C,
CpCtBu), 118.6 (s, 1C, CpCtBu), 127.0-137.0 (m, 40C, C6H5), 137.7,
138.6, 139.6, and 141.5 (4 m, 2C each, ipso-C6H5).
With the aim of providing a more accurate interpretation of
the NMR data for compounds 3 and 4 as well as the
corresponding free Fc(P)4 Bu, we simulated the spectra.5a,7 The
t
t
[Ni2Cl4{Fc(P)4 Bu}], 2. A solution of 150 mg (0.145 mmol) of
t
simulations for Fc(P)4 Bu and for complex 3 are depicted in
t
Fc(P)4 Bu in toluene (20 mL) was added dropwise to a suspension of
Figure 1. The corresponding numerical data are collected in
Table 1, which summarizes the coupling constants obtained after
iterative calculations.13 As compared to a first-order interpreta-
tion previously reported,5a,7 these simulations afford coupling
constants with a slight deviation (0.1-3 Hz).
NiCl2‚DME (62 mg, 0.28 mmol) in THF (20 mL). The mixture was
stirred and heated at 90 °C for 15 h, leading to the formation of a deep
red precipitate. The solvent was removed in vacuo, and the red solid
was extracted with dichloromethane. After CH2Cl2 removal, the product
was washed twice with 5 mL of hexane. Compound 2 was crystallized
at -18 °C from a CH2Cl2/pentane mixture (yield 132 mg, 72%). Anal.
Calcd for C66H62Cl4P4FeNi2 (1294.15): C, 61.3; H, 4.8. Found: C,
The same simulation procedure was applied to the newly
synthesized mononuclear and dinuclear nickel complexes 1 and
2 (analogous to 3 and 4, see Scheme 2). The bromide analogue
1b as well as the dinuclear complex 2 (Table 2) were structurally
characterized by single-crystal X-ray, thus allowing a correlation
1
59.9; H, 4.6. H NMR (CD2Cl2): δ ) 8.60-6.90 (m, 40H, Ph), 4.45
(s, 2H, Cp), 4.11 (s, 2H, Cp), 0.70 (s, 18H, tBu). 31P{1H}(CD2Cl2): δ
) 33.3 (q, JAB ) 76.3 Hz). 13C{1H}(CD2Cl2): δ ) 30.6 (s, 6C, tBuCH3),
t
2
31.0 (s, 2C, BuCCH3), 65.1 and 69.0 (d, 2C each, CpCH, JCP ) 9.0
t
between solid-state structures (3,7 4,7 Fc(P)4 Bu,5a 1b, 2) and
1
and 12.0 Hz), 81.8 and 91.6 (dd, 2C each, CpCP, JCP = 48 Hz and
JPP coupling constants.
2JCP = 39 Hz), 119.2 (s, 2C, CpCtBu), 125.5, 126.6, 128.4, 129.4 (d,
3.1. NMR Spectroscopic Studies. Both chloride and bromide
analogues of the mononuclear 1 and 1b and the dinuclear 2
and 2b display essentially identical 31P NMR patterns. As
compared to their palladium analogues 3 and 4, their solution
NMR spectra reveal significant differences.7 A noticeable
difference is the magnitude of the JPP coupling constants
between homoannular phosphorus when compared to the
dinuclear complexes [Ni2Cl4{Fc(P)4 Bu}], 2, and [Pd2Cl4{Fc-
(P)4 Bu}], 4. While both of these spectra confirm the presence
1
2C each, 49 < JCP < 56 Hz, ipso-C6H5), 127.4 (d), 127.8 (d), 129.9,
130.2, 130.4 (d), 131.3, 131.8 (d), 132.2, 133.1 (d), 134.0 (d), 134.7
2
(d) (40C, for the doublets: 9 < JCP < 18 Hz, C6H5). Because of the
increased solubility of complex 2 in CD2Cl2, and in contrast to 1b, 2b,
and 4, the quaternary carbons CP of the Cp rings could be unambigu-
ously assigned.
3
t
[NiBr2{Fc(P)4 Bu}], 1b. A mixture of 1,1′,2,2′-tetrakis(diphenyl-
phosphino)-4,4′-di-tert-butylferrocene (94 mg, 0.091 mmol) and NiBr2‚
DME (28 mg, 0.091 mmol) was stirred in dichloromethane (60 mL) at
ambient temperature for 2 h. The solution quickly turned purple. After
the solution was filtered, the solvent was removed in vacuo. Upon
evaporation and standing at room temperature, compound 1b crystal-
lized (120 mg, yield 90%), yielding single crystals suitable for X-ray
studies. Anal. Calcd for C66H62Br2P4FeNi (1253.45): C, 63.2; H, 5.0.
Found: C, 63.3; H, 4.9. 1H NMR (CD2Cl2): δ ) 8.60-6.40 (m, 40H,
t
t
of two pairs of isochronous phosphorus (P1/P2 and P3/P4 in
Scheme 3), the signals for 2 appear at 33.73 and 32.83 ppm,
JPP ) 76.3 Hz, as an A2B2 quartet, and the signals for 4 appear
as a quasi doublet of doublet (almost an A2X2 pattern, with
signals centered at 44.86 and 42.15 ppm, JPP ) 12.0 Hz). The
discrepancy between the data obtained in solution from 31P
t
Ph), 4.49, 4.14 (m, 2H each, Cp), 0.79, 0.78 (s, 9H each, Bu).
31P{1H}(CD2Cl2, 298 K): δ ) 39.2, 34.3 (m, very broad, P-Ni each),
-23.2 (m, very broad, pendant PPh2), -25.7 (s, pendant PPh2).
31P{1H}(CD2Cl2, 193 K): δ ) 39.0 (dd, JP1P4 ) 67.0 Hz, JP1P2 ) 31.2
Hz), 33.9 (d, JP4P1 ) 67.0 Hz), -26.5 (d, JP2P1 ) 31.2 Hz), -28.0 (s).
13C{1H}(CD2Cl2): δ ) 29.4 and 31.0 (s, 3C each, tBuCH3), δ obscured
for tBuCCH3, 65.5, 69.3, 71.8, 74.5 (m, 1C each, CpCH), 127.0-135.0
(m, 40C, C6H5). Due to a lack of solubility, the quaternary carbons
were not detected.
t
NMR between the mononuclear palladium [Pd2Cl4{Fc(P)4 Bu}],
3, and the mononuclear nickel complexes 1 or 1b is even much
more striking. The four nonequivalent phosphorus nuclei of the
mononuclear compound 3 give four well-defined multiplets
(Figure 1b). This suggests a 1,2-homoannular biligate coordina-
t
tion of the tetradentate ligand Fc(P)4 Bu to the palladium center
(δ ) 41.69, 39.24 ppm). The remaining other two homoannular
phosphorus nuclei are noncoordinated (δ ) -22.59, -26.34
ppm). In the case of the mononuclear nickel complex 1 (Figure
t
[Ni2Br4{Fc(P)4 Bu}], 2b. A mixture of 1,1′,2,2′-tetrakis(diphenyl-
phosphino)-4,4′-di-tert-butylferrocene (940 mg, 0.91 mmol) and NiBr2‚
DME (560 mg, 1.8 mmol) was stirred and refluxed in dichloromethane
(40 mL) for 15 h. The solvent of the red-purple suspension was removed
in vacuo to yield compound 2b (1.06 g, yield 80% after workup
(13) The spin systems were simulated using “g-NMR” software (Adept Scientific
v.-5.0). The line width was individually fitted to the experimental spectrum
(see Table 1).
9
J. AM. CHEM. SOC. VOL. 126, NO. 35, 2004 11079

A R T I C L E S
Hierso et al.
t
Figure 1. (a) Experimental and simulated 31P NMR spectra of the ferrocenyl tetraphosphine Fc(P)4 Bu. (b) Experimental and simulated 31P NMR spectra
of the mononuclear palladium complex 3.
2), four different phosphorus signals are observed: two at low-
field (δ ) 31.84, 27.91 ppm) and two at high-field (δ ) -18.45,
-25.87 ppm). At ambient temperature, the three signals at lower
field are very broad (w1/2 ) 164 Hz), suggesting either a
fluxional behavior or a paramagnetic nickel center. The possible
fluxional behavior that can be anticipated might be an equilib-
rium of the {NiCl2}-moiety between two sites, leading to a fast
intramolecular exchange from 1,2- to 1,1′-phosphorus chelating
ligation. As a consequence, the resulting NMR would be an
average spectrum between a diamagnetic square-planar complex
and a paramagnetic tetrahedral complex.14,15 To test this
hypothesis, low-temperature 31P NMR experiments were carried
out. As shown in Figure 2, cooling a solution of 1 in CD2Cl2
transforms the three broad phosphorus signals into clearly
9
11080 J. AM. CHEM. SOC. VOL. 126, NO. 35, 2004

Coupling in Tetraphosphine Ferrocenyl Derivatives
A R T I C L E S
Table 1. Correlation of 31P NMR JPP Coupling Constants and X-ray Crystallographic Data (“Through-Space” Distances)
tetradentate compound
P‚‚‚P distances (Å)/
coupling constant (Hz)
Fc(P)4tBu
PdFc(P)4tBu, 3
Pd2Fc(P)4tBu, 4
NiFc(P)4tBu, 1b
Ni2Fc(P)4tBu, 2
3.644(3)/31b
5.397(3)/c
4.429(3)/c
6.740(3)/c
3.002(3)/67b
3.765(3)/c
4.987(8)/c
5.382(8)/c
5.373(9)/c
6.577(8)/c
2.950(7)/76.3
3.067(7)/76.3
a
d P1‚‚‚P2/JPP
3.728(2)/59.8
4.861(2)/c
4.861(2)/c
6.633(2)/c
3.364(2)/74.5
3.364(2)/74.5
3.842(1)/24.0
4.698(2)/c
4.440(1)/6.4
6.292(2)/c
3.035(2)/15.7
3.515(2)/14.3
4.985(5)/c
5.372(5)/c
5.351(4)/c
6.582(5)/c
3.113(5)/13.0
3.018(5)/13.0
a
d P1‚‚‚P3/JPP
a
d P4‚‚‚P2/JPP
a
d P4‚‚‚P3/JPP
a
d P1‚‚‚P4/JPP
a
d P2‚‚‚P3/JPP
b
a Values from simulated spectra at (0.1 Hz, except for 3 at (0.5 Hz and 1b at (1 Hz. JPP values are given at -80 °C due to the broadening from the
paramagnetic influence observed at ambient temperature (Figure 2); the data are from compound 1b, that displays a temperature-dependent 31P NMR spectrum
similar to that of 1.17 c Coupling constant null or <0.5 Hz. The numbering scheme corresponds to that displayed in Scheme 3.
Scheme 3. Representation of the Complexes and Phosphorus Numbering (Px) Used in the Discussion
identifiable multiplets. This phenomenon is fully reversible upon
warming to room temperature. Only very weak shifts of the
signals are observed (e4 ppm) between 298 and 193 K. If two
magnetically different species (a paramagnetic tetrahedral and
one diamagnetic square-planar complex) were in equilibrium,
we expect a significant change in chemical shift upon cooling,
which is clearly not the case.
The alternative explanation implies a weakly bonding inter-
action between the phosphorus P2 and the nickel center to yield
a 4 + 1 coordination. This leads to a structural distortion of the
square-planar (diamagnetic) environment to afford an elongated
pseudo-five-coordinate geometry (see Scheme 2, top). Compa-
rable elongated five-coordinate geometries have been reported
for nickel(II)/trisphosphine complexes.16 Furthermore, it explains
why the P2 signal appears at high-field (as compared to Ni-
bonded P1 and P4) and yet is broadened. This weak interaction
between the phosphorus P2 and the paramagnetic nickel center
(leading to a broadened signal) at ambient temperature disap-
pears upon cooling to 193 K, to yield a sharp 31P NMR signal,
indicative of a square-planar diamagnetic Ni(II) coordination
(Figure 2).17
of the mononuclear nickel compound 1b (as well as the
dinuclear nickel complex 2) was carried out.
3.2. Crystallographic Studies. The X-ray structure obtained
for 1b (Figure 3) is consistent with the NMR spectra observed
in solution at low temperature for the mononuclear nickel
complexes 1 and 1b. The nickel center lies in a slightly distorted
square-planar environment, and while two phosphorus atoms
are bonded to Ni, two other homoannular phosphorus atoms
remain pendant.
As compared to the molecular structure of the palladium
analogue 3,7 the cyclopentadienyl planes show a smaller
deviation from an eclipsed conformation: 9.7(6)° for 1b versus
17.1(6)° for 3 (mean value of dihedral angles defined by C(1i)-
CNT(1)-CNT(2)-C(2i), C(1i) and C(2i) being atoms of both
Cp rings). An additional difference is the heteroannular phos-
phorus-phosphorus internuclear distances (see Table 1 and
Scheme 3): for 1b, the P2 is closer to P1 and P4, while P3 is
significantly further from P1 and P4 (d P1‚‚‚P3 ) 5.397(3) Å
against 4.698(2) Å for 3, d P4‚‚‚P3 ) 6.740(3) Å against
6.292(2) Å for 3). Thus, in 1b, one of the pendant phosphorus
atoms (P3) is clearly isolated from the first coordination sphere
of the nickel.18 These results are in full agreement with the
following:
At 193 K, the 31P NMR spectrum of the mononuclear
complex 1 can be described as a first-order ABMX spin system
with detectable coupling constants between P1/P4 and P1/P2. In
contrast to P2, its P3 neighbor appears at all temperatures as a
sharp singlet. This suggests that this donor is fully decoupled
(chemically and magnetically) from the rest of the system.
To confirm the structural assumptions deduced from the low-
temperature NMR evidence, the X-ray structural determination
i) the 31P NMR spectrum of 1b which shows, at all considered
temperatures, the most shielded signal for P3 as a singlet which
does not undergo any paramagnetic influence (similar to the
phenomenon depicted in Figure 2 for 1), and
ii) the possibility, at the higher temperature and in solution,
of an elongated pseudo-five-coordinate complex formation
involving the atom P2 (Scheme 3, middle).
(14) Angulo, I. M.; Bouwman, E.; Lok, S. M.; Lutz, M.; Mul, W. P.; Spek, A.
L. Eur. J. Inorg. Chem. 2001, 1465.
(15) 1,1′-P(Ph2) bound nickel halide complexes of ferrocenyl diphosphine ligands
display a tetrahedral geometry: (a) Butler, I. R.; Cullen, W. R.; Kim, T.-
J.; Rettig, S. J.; Trotter, J. Organometallics 1985, 4, 972. (b) Casellato,
U.; Ajo, D.; Valle, G.; Corain, B.; Longato, B.; Graziani, R. J. Crystallogr.
Spectrosc. Res. 1988, 18, 583.
Comparing the homoannular P‚‚‚P distances, we note a
significant difference for the d P2‚‚‚P3 between the free ligand
t
Fc(P)4 Bu, and the corresponding mononuclear palladium and
nickel analogues 3 and 1b (3.364(2), 3.515(2) Å vs 3.765(3)
(16) Benelli, C.; Di Vaira, M.; Noccioli, G.; Sacconi, L. Inorg. Chem. 1977,
16, 182 and references therein. Longer metal-ligand distances (radial
expansion) are predicted and found in five-coordinate high-spin complexes,
as compared to low-spin complexes.
(18) The wide exo-open angles P(3)-C(35)-C(36) ) 130.0(5)° and P(4)-
C(36)-C(35) 127.0(5)° (see Figure 3) should also be noted, while in the
t
(17) The 31P NMR spectrum at -80 °C for 1b is available as Supporting
Information.
other X-ray structures of 2, 3, 4, and Fc(P)4 Bu, the corresponding angles
display values e120°.
9
J. AM. CHEM. SOC. VOL. 126, NO. 35, 2004 11081

A R T I C L E S
Hierso et al.
Figure 2. Variable-temperature 31P NMR monitoring of the mononuclear nickel complex 1.
Å, respectively). Interestingly, the only system which displays
no P2 P3 coupling constant is the nickel complex 1b.
The X-ray structure obtained for 2 (Figure 4) and compared
to that obtained for the congener dinuclear palladium complex
47 indicates quasi similar structures for 2 and 4. The Cp planes
show deviations from an eclipsed conformation of 12.2(6)° and
13.1(3)°, respectively, and the phosphorus-phosphorus het-
eroannular internuclear distances are equal in the range of (0.02
Å (see Table 1). The need to rationalize the JPP results
summarized in Table 1 led us to propose the theoretical model
described in the following.
designated as “through-space” coupling (also called nuclear
spin-spin coupling via nonbonded interactions) has been
analyzed by Mallory and co-workers using a simple but powerful
perturbation model. The “through-space” couplings result from
overlap interactions between lone-pair orbitals on the two
crowded elements. For instance, in the particular constrained
geometry of the compounds sketched in Scheme 4 (top), the
C-F/C-F, or C-F/C-N bonds, are coplanar and approximately
parallel. As a consequence, the nonbonding distances d F‚‚‚F,
or d F‚‚‚N, are short. In this orientation, the two lone-pair
orbitals experience a σ-directed overlap. As displayed in Scheme
4 (bottom), the overlap between these lone-pair orbitals is
expected to afford an in-phase and out-of-phase combination.
As both orbitals are occupied (two-orbital, four-electron interac-
tion), no stabilization is observed. However, it provides an
adequate pathway for transmitting spin information between the
3.3. Model for Nonbonded Spin-Spin Coupling Interac-
tions. Organic compounds containing pairs of fluorine atoms
(or nitrogen/fluorine atoms) that are intramolecularly crowded
against one another exhibit large 19F19F (or 19F15N) nuclear spin
10
coupling constants JFF or JFN
.
This phenomenon commonly
9
11082 J. AM. CHEM. SOC. VOL. 126, NO. 35, 2004

Coupling in Tetraphosphine Ferrocenyl Derivatives
A R T I C L E S
Table 2. Crystal Data, Data Collection, and Structure Refinements for Compounds 1b and 2
1b‚3CHCl3
2‚CH Cl2
2
formula
C66H62Br2P4FeNi‚3CHCl3
C67H64Cl6P4FeNi2
1379.03
180(2)
formula weight
temperature, K
crystal color
1611.52
110(2)
purple
red
crystal system
space group
monoclinic
P21/n
monoclinic
P21/n
a, Å
b, Å
c, Å
12.1064(2)
40.2368(7)
14.5029(3)
93.320(1)
7052.8(2)
4
1.518
2.078
14.3226(19)
22.814(3)
21.159 (3)
104.850(14)
6682(15)
4
1.371
1.146
2840
â, deg
volume, Å3
Z
calculated density, mg m-3
absorption coeff., mm-1
F(000)
3264
θ range for data collection
reflections collected/unique
R(int)
refinement method
goodness-of-fit on F2
final R indices [I > 2σ(I)]
R indices (all data)
residuals, e Å-3
2.47-26.06°
2.5-45.0°
26 546/7872
0.2225
26 492/12 660
0.0931
full-matrix least-squares on F2
1.014
R1 ) 0.0691, wR2 ) 0.1278
R1 ) 0.1668, wR2 ) 0.1583
1.390, -1.557
0.826
R1 ) 0.0840, wR2 ) 0.1915
R1 ) 0.2161, wR2 ) 0.2609
0.726, -0.424
Figure 4. Plot of the molecular structure of 2. Selected bond lengths (Å)
and angles (deg): Fe-CNT(1) 1.687, Fe-CNT(2) 1.671, Ni(1)-P(3)
2.161(6), Ni(1)-P(4) 2.147(5), Ni(2)-P(1) 2.160(5), Ni(2)-P(2)
2.182(6), Ni(1)-Cl(1) 2.187(5), Ni(1)-Cl(2) 2.195(5), Ni(2)-Cl(3)
2.194(5), Ni(2)-Cl(4) 2.201(5); CNT(1)-Fe-CNT(2) 173.49, P(1)-Pd-
(1)-P(3) 84.40(10), P(3)-Pd(1)-Cl(2) 89.87(10), P(1)-Pd(1)-Cl(1)
91.48(10), Cl(2)-Pd(1)-Cl(1) 93.70(10), P(2)-Pd(2)-P(4) 87.41(10),
P(4)-Pd(2)-Cl(4) 91.00(10), P(2)-Pd(2)-Cl(3) 90.95(11), Cl(3)-Pd(2)-
Cl(4) 90.45(11), C(23)-C(24)-P(1) 118.1(7), C(13)-C(14)-P(2)
116.5(7), C(24)-C(23)-P(3) 113.4(7), C(14)-C(13)-P(4) 118.8(8).
quantitatively estimated for a family of 1,8-difluoronaphthalene
through an exponential relation between the ab initio calculated
internuclear distances and the JFF coupling observed.10h
Figure 3. Plot of the molecular structure of 1b. Selected bond lengths (Å)
and angles (deg): Fe-CNT(1) 1.688, Fe-CNT(2) 1.687, Ni-P(1)
2.187(2), Ni-P(2) 2.173(2), Ni-Br(1) 2.3548(11), Ni-Br(2) 2.3379(11);
CNT(1)-Fe-CNT(2) 177.77, P(2)-Ni-P(1) 87.01(7), P(1)-Ni-Br(2)
89.63(6), P(2)-Ni-Br(1) 89.05(6), Br(2)-Ni-Br(1) 93.33(4), P(3)-
C(35)-C(36) 130.0(5), P(4)-C(36)-C(35) 127.0(5), P(1)-C(2)-C(3)
114.4(4), P(2)-C(3)-C(2) 117.0(5).
On the basis of this excellent work exclusively dealing with
organic species and (FF) nuclear spin coupling, we propose to
extend this model to phosphorus nuclear 31P-31P spin-spin
“through-space” interactions and to complete it with the case
of metallo-organic species.
coupled nuclei. The magnitude of JFF depends on the extent to
which the two lone-pair orbitals interact due to their overlap.
This model, initially mainly qualitative,10a,g has led to a
breakthrough when the so-called overlap interaction has been
3.3.1. The Problem of “Through-Space” versus “Through-
Bond” Nuclear Coupling in the Case of Multidentate
9
J. AM. CHEM. SOC. VOL. 126, NO. 35, 2004 11083

A R T I C L E S
Hierso et al.
Scheme 4. Bonding and Anti-Bonding Orbitals Generated by the
Overlap of Two Lone-Pair Orbitals on Intramolecularly Crowded
Nitrogen- and Fluorine-Containing Compounds
3.3.2. The Requirements for “Through-Space” Nuclear
Spin-Spin Coupling. Empirical requirements for the detection
of true “through-space” coupling constants can be derived from
the last 30 years of work concerning mainly (FF),10a,b,f,h (NF),10e,g
(HF), or (CF)10c,d spin-spin coupling, and occasionally
(PX),9a-d,10h with X ) P, Se, F. The first requirement is trivial
but involves in practice some “architectural” difficulties. To
easily and directly detect a coupling constant when the nuclei
are of same nature, they have to be anisochronous. The molecule
must be dissymmetric (see Scheme 4, top left). That is mainly
why so few “through-space” (PP) couplings have been reported
to date.19,20 The second requirement is that the nuclei must be
in close proximity in solution. A way to achieve this is to
provide backbone rigidity: this is the case of Mallory and co-
workers’ aromatic-fluorine species.10 That is the case, as well,
of the polymetallic catenanes-containing bridging phosphorus
reported by Rheingold and Fountain.19 This conformational
rigidity is also evoked by Ito et al. in their diphosphine bis-
(ferrocenyl) compounds.20 In our case as well, the tetraphosphine
species described here have a locked-conformation in solution
even at temperatures above 70 °C.5a,7 A last requirement invoked
by Mallory is the presence of two lone-pair atomic orbitals for
mutual overlap.
3.3.3. Extension of the Lone-Pair Overlap Interaction
Model and Involvements. In the case of the ferrocenyl
t
tetraphosphine Fc(P)4 Bu (Scheme 5, top), the spatial proximity
of the two heteroannular phosphorus P1 and P2 atoms (Scheme
3) and their lone-pair spatial orientation (ascertained by X-ray
diffraction study5a) certainly lead to an 3sp3-3sp3 orbital overlap
of the same nature as those assumed for 2p-2p orbitals in F/F
pairs or for 2p-2sp2 orbitals in F/N pairs (Scheme 4). In line
with Mallory’s theory,10 the overlap of phosphorus lone-pair
atomic orbital might be formulated as generating two molecular
orbitals occupied by four electrons. That overlap interaction does
not lead to net chemical bonding between the phosphorus atoms
but ensures the transmittal of nuclear spin information. As has
been detailed for fluorine atoms,10a the overlap between the
electronic clouds generates significant Fermi contact (FC)
interactions12a between the P nuclei.
Phosphinoferrocene Compounds. In the general approach used
to describe the origin of nuclear spin coupling constants, the
three terms of eq 1 are distinguished:
tb(π)
tb(σ)
JXX′ ) tsJXX′
+
J
+
J
(1)
XX′
XX′
ts
The coupling constant term
J
is due to “through-space”
XX′
interaction, and the two other terms tbJXX′ arise from “through-
bond” interaction (σ- and π-transmitted components).10h,11a The
common scheme to experimentally evaluate strong JXX′ is to
ts
study systems where the shortest “through-bond” pathway is
long enough (at least four bonds) that one could reasonably
assume that “through-bond” contributions to nuclear coupling
are negligible in magnitude. In the case of the tetraphosphine
ferrocenyl derivatives presented here, the shortest “through-
bond” interactions between phosphorus couples can be consid-
In agreement with this simple, but powerful model, recent
theoretical work has shown that a neat distinction should be
made between the “through-space” transmission of the FC term
and of the paramagnetic spin-orbit (PSO) terms.21a,b The
predominant role of the Fermi contact has been confirmed by
density functions interpretation,21c some specifically addressing
the problem of “through-space” JFF coupling.12a In the present
3
ered as JPP′ coupling constants for homoannular phosphorus,
4
and JPP′ for heteroannular phosphorus couples (for instance,
P1-C-Fe-C-P2 or P1-C-Fe-C-P3, Scheme 3). The 31P
NMR spectra of all of the studied compounds display no
detectable values for JP1P3 (entry 2 in Table 1). This suggests
J
PP coupling, as demonstrated in the case of F‚‚‚F interactions,
the magnitude of the coupling between two crowded phosphorus
ts
4
that both JP1P3 and “through-bond” JP1P3 are too small to be
detected. In contrast, JP1P2 shows values ranging from 0 to 60
Hz (entry 1 in Table 1), while there is no valid argument to
(19) As a rare example, see: Rheingold, A. L.; Fountain, M. E. Organometallics
1984, 3, 1417. In this work, a 2JP‚‚‚P of 125.4 Hz was found and correlated
to the internuclear “through-space” distance d(P‚‚‚P) ) 2.715(3) Å; the
authors underline the lack of sufficient NMR and crystallographic data to
demonstrate the extent and nature of correlations between d(P‚‚‚P) and
4
account for the fact that the “through-bond” JP1P2 coupling
should be different from JP1P3 found (because the “through-
2
2JP‚‚‚P. It might be added that, in JP‚‚‚P cases, the relative contribution of
4
ts
tb
J
and
J
is difficult to estimate.
PP′
PP′
(20) As a rare example, see: Sawamura, M.; Hamashima, H.; Sugawara, M.;
Kuwano, R.; Ito, Y. Organometallics 1995, 14, 4549. In this work, a
7JP‚‚‚P of 22.0 Hz has been found; unfortunately, no X-ray structural data
are reported. This value was found for the only dissymmetric derivative
(containing PPh2 and P(p-Tol)2 moieties) of a family of diphosphine bis-
(ferrocenyl) compounds.
(21) (a) Soncini, A.; Lazzeretti, P. J. Chem. Phys. 2003, 118, 7165. (b) Wu, A.;
Gra¨fenstein, J.; Cremer, D. J. Phys. Chem. 2003, 107, 7043. (c) Soncini,
A.; Lazzeretti, P. J. Chem. Phys. 2003, 119, 1343.
bond” environments of P2 and P3 are identical). Thus, due to
the particular geometry of these tetraphosphine species, it can
be ascertained that heteroannular coupling is exclusively arising
from “through-space” interactions and that “through-bond” 4tbJPP′
contributions are nondetectable. The case of homoannular 3JPP′
is more delicate and is discussed in section 3.3.3.
9
11084 J. AM. CHEM. SOC. VOL. 126, NO. 35, 2004

Coupling in Tetraphosphine Ferrocenyl Derivatives
A R T I C L E S
Scheme 5. Extension of the Lone-Pair Overlap Interaction Model
to Phosphorus Atoms (Left), and Schematic Diagram Applied to
Coordination Complexes Involving a Transition Metal Orbital
Contribution (Right)
tions. From the 20 results of heteroannular PP′ coupling con-
stants observed (entries 1-4), it clearly appears that the values
of JPP′ present a strong inverse dependence on P‚‚‚P distances.
Whatever the compound, for P‚‚‚P distances above 4.90 Å, no
detectable JPP′ is observed. On the other hand, for distances
below 4.00 Å, JPP′ > 20 Hz are systematically obtained.
No heteroannular phosphorus coupling constants are detected
for the dinuclear metal complexes 2 and 4. As none of the
phosphorus atoms possess a lone-pair for overlap interactions,
a “through-space” interaction for such a situation would require
an overlap between two bonding orbitals (rather than a lone-
pair and a bonding orbital). Such a situation awaits experimental
demonstration. In line with the absence of orbital interactions,
the phosphorus center distances lie in the range between 4.98
and 6.58 Å.
More specifically, the phosphorus distances (d P4‚‚‚P2) are
comparable in the nickel complex 1b and in the related
palladium complex 3 (∼4.4-4.3 Å within the range of standard
deviation). However, the palladium complex 3 exhibits a JP2P4
) 6.4 Hz, whereas the nickel complex 1b displays no such
heteroannular “through-space” coupling. This is consistent with
the proposed theoretical extended-formulation involving a metal-
orbital contribution. As the extent of the palladium valence
orbital is larger than its nickel analogue, the extent of an
effective overlap with the lone-pair orbital is larger, thus yielding
a detectable “through-space” coupling constant between the
phosphorus nucleus. The proposed model allows one to predict
that, in an analogous case involving a platinum mononuclear
species, for instance, at a similar distance (d P‚‚‚P), a coupling
constant P2P4 would most probably be detected.
that each bear a lone-pair depends on the effectiveness of the
overlap interactions between the two lone-pair orbitals.
Interestingly, in the mononuclear Pd and related Ni complexes
3 and 1b, one of the two “through-space” interacting phosphorus
lone-pairs is involved in P-M bonding interactions. Thus, the
“through-space” interaction involves a three-center system
composed of P-M (M ) Ni, Pd) and P. Scheme 5 (right)
presents an important corollary to the model developed for
fluorinated purely organic compounds. For simplicity, only the
three localized orbitals involved in the “through-space” interac-
tion are sketched in Scheme 5. The metal-ligand bond between
P1 and M is considered as coming from the σ-overlap between
3.4.2. Homoannular Coupling in Complexes. In the case
of homoannular coupling (entries 4 and 5), the situation is more
3tb
complicated because JPP′ is the sum of two terms tsJPP′
+
J .
PP′
The contribution from “through-bond” 3tbJPP′ could be nonneg-
ts
ligible (or even of inverse algebraic sign relative to
J
).11c
PP′
Moreover, an eventual nonbonding interaction between two
filled orbitals would no longer occur via an interaction of mainly
axial nature (as described for heteroannular phosphorus sp3
orbitals) but more of lateral symmetry, and thus of weaker
overlap effectiveness. The results in Table 1 (entries 4 and 5)
suggest that the contribution of a nonbonded lateral σ-type
interaction between the homoannular phosphorus atoms for the
palladium complexes 3 and 4 would be very weak because, for
distances ranging from 3.019 to 3.515 Å, very similar coupling
constants of moderate intensity are found (∼14 Hz). We thus
propose that, in the case of the palladium species, the homoan-
nular 31P31P coupling constants are mostly “through-bond” in
nature. On the contrary, for the nickel complexes, a dependence
of JPP′ on the P‚‚‚P internuclear distances is observed (for d
P‚‚‚P ) 3.765 Å, JPP′ ) 0 Hz, and for d P‚‚‚P ) 3.002 Å,
JPP′ ) 67 Hz). The JPP′ ) 0 Hz for a P‚‚‚P ) 3.765 Å and the
very strong 31P31P coupling constants at short internuclear
distances lead us to assume that, in the case of the nickel
compounds 1b and 2, the “through-bond” contribution to the
overall coupling is weak as compared to the “through-space”
pathway.23
3
a 3sp and a dz orbital.22 To account for the fact that nuclear
spin information is transmitted between the phosphorus atoms,
two filled molecular orbitals have been constructed which
incorporate a contribution from the metal. The qualitative orbital
ordering as well as the symmetry of the interacting orbitals allow
this schematic view. Consequently, to observe “through-space”
nuclear spin coupling, two lone-pair orbitals are not required.
One lone-pair orbital with an appropriate orientation can interact
with a bonding electron pair shared between a phosphorus and
a metal and thus transmit “through-space” nuclear spin P‚‚‚P
information.
2
In light of the above model, the results summarized in Table
1 are examined below.
3.4. Discussion of the Model Applied to Tetraphosphine
Ferrocenyl Derivatives.
3.4.1. Heteroannular Coupling in Metal Complexes. As
was explained in section 3.3.2, the heteroannular phosphorus
coupling constants exclusively arise from nonbonded interac-
t
3.4.3. 31P31P Coupling in the Free Ligand Fc(P)4 Bu.
According to the extended overlap interaction formulation, the
free ligand Fc(P)4 Bu should be treated differently from the
coordination complexes thereof. The X-ray structure of the
(22) For simplicity, we ignore the complete treatment of the molecular orbitals
of the coordination complexes, including all of the metal and ancillary ligand
orbitals.
t
9
J. AM. CHEM. SOC. VOL. 126, NO. 35, 2004 11085

A R T I C L E S
Hierso et al.
to 76 Hz) as well as a wide range of distances (from 2.9 to 4.4
Å). For comparison, Figure 5 (bottom) displays the fitted 19F19F
versus d F‚‚‚F data previously reported by Ibrom and Ernst
(R2 ) 0.987 with the distances computed with MM2).25 We
note for phosphorus (due to larger valence orbitals) detectable
coupling constants corresponding to longer “through-space”
distances. We note also a less steep exponential inverse
dependence on internuclear distances, which we attribute to a
less efficient orbital overlap.
Further improvements of the current model26 should take into
account the nature of the metal, as well as the weak “through-
bond” coupling for homoannular phosphorus 3JPP′ values.11c As
was previously noted for JFF coupling constants,25 phosphorus
nuclear spin-spin “through-space” interactions are detected for
distances longer than the sum of van der Waals radii for
phosphorus (3.6 Å).27 While the semiquantitative experimental
formulation presented herein is very convenient (as it presents
an intuitive correlation between “geometric parameters” and
“through-space” nucleus spin-coupling), a more significant
correlation should probably be expressed in terms of orbitals
overlap, taking into account the shape and the relative diffuse
character of the electronic clouds. Consequently, further im-
provements will include quantitative orbital overlap effectiveness
derived from molecular orbital calculations.
Figure 5. (Top) Plot of JPP′ against d P‚‚‚P using data from Table 1 related
to the 1b, 2, and 3 metal complexes. The data are fitted by the exponential
relationship defined by eq 2. (Bottom) Same data as in graph (b), as
compared to data extracted from ref 25 in graph (a) related to the “through-
space” JFF′ coupling in organic compounds (curve-fitting program Excel
2000).
4. Summary
Both from a theoretical and from an experimental point of
view, the analysis of high-resolution NMR parameters is a
critical issue. A deeper understanding of the relationship between
coupling constants and molecular structure could greatly broaden
future applications of high-resolution spectroscopy for the
elucidation of molecular structures.28 Besides the theoretical
computational studies on nuclear spin-spin transmission,
concerning specifically “through-space” coupling constants, the
previous works have focused on fluorine- and nitrogen-
containing organic compounds with atoms having two lone-
pair orbitals interacting. The present paper provides experimental
t
Fc(P)4 Bu ligand5a indicates that the P1 and P2 lone-pairs clearly
point toward each other, suggesting a significant overlap and
thus coupling pathway, in line with Mallory’s formulation. In
the case discussed above, the overlap between the bonding
M-P1 orbital and the P4 lone-pair is much less directional, and
thus less effective, leading to smaller “through-space” coupling
constants. This is consistent with the strong nonbonded coupling
constant JPP′ ) 60 Hz observed for a distance d P‚‚‚P ) 3.73
t
Å in Fc(P)4 Bu, while at such internuclear distances in the metal
complexes, the corresponding JPP′ values range between 20 and
30 Hz.
(23) Opposite results are found with palladium and nickel complexes for
distances of about 3 Å. An explanation might be proposed if is taken into
account that a contribution from metal is expected also when a lateral
“through-space” coupling process is involved. Thus, while the valence
orbitals of Ni are more contracted as compared to Pd orbitals, the Ni orbitals
are much less diffuse, and their contribution to the nonbonded coupling
should be, as well, “less diffuse”. The consequence would be that, at short
P internuclear distances, a nonbonded lateral orbital interaction would be
more effective for Ni species than for Pd species as soon as metal orbitals
are involved.
(24) In the reference work on fluorine-containing compounds (for instance, ref
10h), the dependence of JFF on the extent to which the two lone-pair orbitals
interact because of their overlap is directly proportional to the energy
difference ∆E(σFF/σFF*) (Scheme 4). In turn, this predicted dependence of
JFF on the extent of the lone-pair overlap suggests that JFF should fall off
exponentially with the distance between the two fluorines d F‚‚‚F.
(25) Ernst, L.; Ibrom, K. Angew. Chem., Int. Ed. Engl. 1995, 34, 1881. In this
communication devoted to ”through-space” JFF coupling constants in
difluorocyclophane, the authors underline the discrepancy between the
experimental values obtained for JFF as a function of d F‚‚‚F and some
previous theoretical reports.
(26) One should be aware, as well, of the solvent dependence of the “through-
space” coupling constant. In the present work, however, this is of no concern
as CDCl3 and CD2Cl2 were used throughout and similar values were
consistently obtained. For a relevant study on the solvent effects in
nonbonded JCF and JHF, see: Mele, A.; Vergani, B.; Viani, F.; Meille, S.
V.; Farina, A.; Bravo, P. Eur. J. Org. Chem. 1999, 187, 7.
(27) (a) Bondi, A. J. Phys. Chem. 1964, 68, 441. (b) Pauling, L. The Nature of
the Chemical Bond; Cornell University Press: Ithaca, NY, 1945.
(28) For convincing examples of structural determination using nonbonded spin-
spin couplings, see: (a) Kimber, B. J.; Feeney, J.; Roberts, G. C. K.;
Birdsall, B.; Griffiths, D. V.; Burgen, A. S. V.; Sykes, B. D. Nature 1978,
271, 184-185. (b) Albe´niz, A. C.; Casado, A. L.; Espinet, P. Organo-
metallics 1997, 16, 5416.
3.4.4. Quantitative Correlation of the Distance Depen-
dence of 31P31P Coupling in Coordination Complexes. We
have constructed the plot displayed in Figure 5 (top) using the
data collected in Table 1. For the reasons outlined above, the
t
data from the ligand Fc(P)4 Bu and from the homoannular
phosphorus couplings in the palladium complexes 3 and 4 were
excluded. On the other hand, the data from homoannular
phosphorus in the nickel species, which do depend on phos-
phorus distances, were included. The data points were plotted
JPP′ versus d P‚‚‚P. The best fit was obtained for an exponential
curve expressed by eq 2, in which JPP′ is in units of hertz and
d P‚‚‚P is in units of angstroms.24
JPP′ ) (9691.9)e-1.6102d
(2)
P‚‚‚P
The regression coefficient for this curve, R2 ) 0.975, falls short
of providing a definitive quantitative demonstration of the
extended lone-pair overlap interaction model. However, this
curve is overall consistent with the reported results concerning
19F19F nonbonded interactions.10h,12a,25 It should be emphasized
that the fitted data all correspond to experimentally obtained
values, which span a wide range of coupling constants (from 6
9
11086 J. AM. CHEM. SOC. VOL. 126, NO. 35, 2004

Coupling in Tetraphosphine Ferrocenyl Derivatives
A R T I C L E S
evidence and numerical values, potentially useful both to
theoretical and to experimental chemists.25 It provides a rationale
to the variety of coupling constants observed between phos-
phorus atoms which are the result of nonbonded interaction.
To the best of our knowledge, it is shown for the first time that
only one lone-pair orbital which interacts with a bonding orbital
can transfer the 31P31P nuclear spin information “through-space”
between two phosphorus nuclei in a coordination complex.
help. We also thank the “Conseil Re´gional de Bourgogne” for
a postdoctoral grant to V.V.I.
Supporting Information Available: Simulated 31P NMR
spectra for 1b, 2, and 4, and full X-ray structural data for 1b
and 2 (CIF). This material is available free of charge via the
Internet at http://pubs.acs.org.
Acknowledgment. We thank Thomas R. Ward, Anthony
Vion, Sylviane Royer, and Genevie`ve Delmas for their kind
JA048907A
9
J. AM. CHEM. SOC. VOL. 126, NO. 35, 2004 11087