Phosphorus(I) iodide: A versatile metathesis reagent for the synthesis of low oxidation state phosphorus compounds

Article abstract of DOI:10.1021/ic060186o

An improved method for the synthesis of cyclic low oxidation state P cations is presented. The only byproducts of the reaction of phosphorus triiodide with chelating phosphines are readily removed, resulting in an easily accessible P^I reagent with an anion that can be readily replaced through salt metathesis chemistry. These P^I salts are surprisingly stable, even in O₂ and moisture, and the origin of this unusual stability is elucidated using density functional theory calculations.

Full text of DOI:10.1021/ic060186o

Inorg. Chem. 2006, 45, 6864−6874
Phosphorus(I) Iodide: A Versatile Metathesis Reagent for the Synthesis
of Low Oxidation State Phosphorus Compounds
Bobby D. Ellis and Charles L. B. Macdonald*
Department of Chemistry and Biochemistry, UniVersity of Windsor, Windsor,
Ontario, Canada N9B 3P4
Received February 1, 2006
An improved method for the synthesis of cyclic low oxidation state P cations is presented. The only byproducts of
the reaction of phosphorus triiodide with chelating phosphines are readily removed, resulting in an easily accessible
PI reagent with an anion that can be readily replaced through salt metathesis chemistry. These PI salts are surprisingly
stable, even in O2 and moisture, and the origin of this unusual stability is elucidated using density functional theory
calculations.
Introduction
for our investigation of low oxidation state p-block chemistry;
because of their relative stability in comparison to other PI
compounds, salts containing “triphosphenium” cations of the
type first described by Schmidpeter and co-workers presented
an ideal starting point.12-15 These cations are related to
phospha-Wittig reagents and other dicoordinate P compounds
investigated by Schmidpeter and others, in which either one
or both phosphines are replaced by anionic groups.16-18
Furthermore, triphosphenium cations 1 are the heavier
analogues of the ubiquitous and inert bis(triphosphoranylide-
ne)ammonium (“PNP”) cations 2 that are used as noncoor-
dinating cations in numerous salts. Such cations are also
isovalent with carbodiphosphoranes 3, which are interesting
ligands for transition metal complexes19 and are
The synthesis and examination of compounds containing
P in low-coordinate bonding environments have been a focus
of modern main group chemistry.1,2 The observations and
conclusions derived from such investigations, in conjunction
with the realization of the close relationship between the
behavior of P and C, have provided for an incredible
advancement in the understanding of the structure, bonding,
and reactivity of P compounds in particular and those of the
p-block elements more generally.3 The preceding studies of
P compounds have typically concentrated on compounds
containing the element in either of the common oxidation
states (3+ or 5+).4 Excluding metal phosphide minerals, the
chemistry of compounds containing P in lower oxidation
states has been largely limited to investigations of transient
phosphinidenes, their transition-metal or Lewis-base adducts,
their cyclic polyphosphine oligomers, other catenated poly-
phosphines, and elemental P.1,2,5-8 The related chemistry of
the heavier pnictogen (Pn) derivatives has been largely
ignored.9-11
(5) Baudler, M.; Glinka, K. Chem. ReV. 1993, 93, 1623-1667.
(6) Mathey, F. Angew. Chem., Int. Ed. Engl. 1987, 26, 275-286.
(7) Lammertsma, K.; Vlaar, M. J. M. Eur. J. Org. Chem. 2002, 1127-
1138.
(8) Lammertsma, K. Top. Curr. Chem. 2003, 229, 95-119.
(9) Driess, M.; Ackermann, H.; Aust, J.; Merz, K.; Von Wullen, C. Angew.
Chem., Int. Ed. 2002, 41, 450-453.
(10) Barnham, R. J.; Deng, R. M. K.; Dillon, K. B.; Goeta, A. E.; Howard,
J. A. K.; Puschmann, H. Heteroat. Chem. 2001, 12, 501-510.
(11) Dikarev, E. V.; Li, B. Inorg. Chem. 2004, 43, 3461-3466.
(12) Schmidpeter, A.; Lochschmidt, S.; Sheldrick, W. S. Angew. Chem.,
Int. Ed. Engl. 1982, 21, 63-64.
(13) Schmidpeter, A.; Lochschmidt, S. Inorg. Synth. 1990, 27, 253-258.
(14) Lochschmidt, S.; Schmidpeter, A. Z. Naturforsch., B: Anorg. Chem.,
Org. Chem. 1985, 40B, 765-773.
(15) Schmidpeter, A. Heteroat. Chem. 1999, 10, 529-537.
(16) Shah, S.; Protasiewicz, J. D. Coord. Chem. ReV. 2000, 210, 181-
201.
We required readily accessible and functionalizable low
oxidation state group 15 reagents as a necessary starting point
* To whom correspondence should be addressed. E-mail: cmacd@
uwindsor.ca.
(1) Dillon, K. B.; Mathey, F.; Nixon, J. F. Phosphorus: The Carbon Copy;
John Wiley & Sons: Chichester, U.K., 1998.
(2) Regitz, M., Scherer, O. J., Eds. Multiple Bonds and Low Coordination
in Phosphorus Chemistry; Georg Thieme Verlag: Stuttgart, 1990.
(3) Bouhadir, G.; Bourissou, D. Chem. Soc. ReV. 2004, 33, 210-217.
(4) We prefer to use a simple, chemically intuitive convention in which
the oxidation state of P is indicated by the number of lone pairs
associated with the atom: P5+ has zero lone pairs, P3+ has one lone
pair, and P+ has two lone pairs.
(17) Schmidpeter, A.; Zwaschka, F. Angew. Chem., Int. Ed. Engl. 1977,
16, 704-705.
(18) Schmidpeter, A.; Burget, G.; Zwaschka, F.; Sheldrick, W. S. Z. Anorg.
Allg. Chem. 1985, 527, 17-32.
(19) Schmidbaur, H. Angew. Chem., Int. Ed. Engl. 1983, 22, 907-927.
6864 Inorganic Chemistry, Vol. 45, No. 17, 2006
10.1021/ic060186o CCC: $33.50
© 2006 American Chemical Society
Published on Web 07/19/2006

Phosphorus(I) Iodide
potential di-Wittig reagents for organic synthesis.20
Instrumentation. NMR spectra were recorded at room temper-
ature in CD2Cl2 solutions on a Bruker Avance 300-MHz spectrom-
eter. Chemical shifts are reported in ppm, relative to external
1
standards (SiMe4 for H and 13C NMR, 85% aqueous H3PO4 for
31P NMR, BF3‚OEt2 for 11B NMR, CFCl3 for 19F NMR, and 0.05
M Ga(NO3)3 for 71Ga NMR). Coupling constant magnitudes, |J|,
are given in hertz. Melting points (Mp) or decomposition points
(Dp) were obtained on samples sealed in glass capillaries under
dry N2 using an Electrothermal melting point apparatus. Elemental
analysis was performed in-house using a PerkinElmer 2400 C, H,
and N analyzer in the Centre for Catalysis and Materials Research,
Department of Chemistry and Biochemistry, University of Windsor,
Windsor, Ontario, Canada.
Theoretical Methods. Calculations were performed with the
Gaussian98 suite of programs.25 Geometry optimizations have been
calculated using DFT, specifically implementing the B3PW91
method [containing Becke’s three-parameter hybrid functional for
exchange (B3, including ca. 20% Hartree-Fock exchange)26
combined with the generalized gradient approximation for correla-
tion of Perdew and Wang (PW91)27] in conjunction with the
6-31+G(d) basis set. The geometries were restricted to the highest
reasonable symmetry, and each stationary point was confirmed to
be a minimum having zero imaginary vibrational frequencies. The
electronic energies of the molecules have been corrected by the
unscaled, zero-point vibrational energy (ZPVE). Single point
energies have been calculated at the B3PW91/6-311+G(3df,2p)//
B3PW91/6-31G(d) level of theory. Population analyses were
conducted using the natural bond orbital (NBO)28 method imple-
mented in Gaussian98.
The unique reactivity of PI and AsI compounds is exempli-
fied by Driess’ synthesis of the first square-planar phospho-
nium and arsonium salts through the action of PI and AsI
sources on Schwartz’s reagent.9,21 Despite such unprec-
edented and impressive results, the behavior and utility of
compounds such as 1 have not been explored or exploited
to any significant extent. To successfully investigate many
types of reactivities that may be exhibited by such salts, the
presence of a robust and nonreactive counteranion is of
critical importance. In this light, we required a suitable source
for PnI ions in which we could readily change the counter-
anion. Thus, we recently reported the synthesis, characteriza-
tion, and initial reactivity studies of new and valuable
phosphorus(I) and arsenic(I) iodide salts.22 Herein we detail
the facile synthesis of the iodide salts of two stable PI sources,
[[1,2-bis(diphenylphosphino)ethane]phosphorus]+ (4) and
[[1,3-bis(diphenylphosphino)propane]phosphorus]+ (5), and
describe the metathesis reactivity of these reagents. In
addition, the structural features, reactivities, and remarkable
stabilities of selected examples of such compounds are
explained using density functional theory (DFT) calculations.
Preparation of [(dppe)P][I] (4[I]) from PI3. A colorless solution
of dppe (5.071 g, 12.728 mmol) in CH2Cl2 was added to a red
solution of PI3 (5.240 g, 12.728 mmol) in CH2Cl2 (100 mL), which
resulted in the formation of a burgundy solution and an orange
precipitate. Volatile components were removed under reduced
pressure, and the crude residue was washed with THF (3 × 30
mL). The solid was redissolved in CH2Cl2 (75 mL) and filtered
through Celite to remove further small amounts of orange precipi-
tate. Volatiles were again removed in vacuo, and after dissolution
in MeCN and slow evaporation of the solvent, a colorless crystalline
material was obtained. Yield: 52% (3.680 g, 6.615 mmol). 31P-
Experimental Section
Reagents and General Procedures. All manipulations were
carried out using standard inert atmosphere techniques. Phosphorus-
(III) iodide, 1,2-bis(diphenylphosphino)ethane (dppe), 1,3-bis-
(diphenylphosphino)propane (dppp), silver trifluoromethanesulfonate
(silver triflate, AgOTf), and potassium hexafluorophosphate were
purchased from Strem Chemicals Inc., and all other chemicals and
reagents were obtained from Aldrich; all reagents were used without
further purification. MeOH was dried over Mg turnings, and all
other solvents were dried on a series of Grubbs’-type columns23
and were degassed prior to use. CD2Cl2 was dried over calcium
hydride. Ag salts are sensitive to light, and precautions were taken
to minimize light exposure in reactions involving such salts. The
compound dppeI4 (6) was synthesized by the slight modification
of a literature procedure.24 The relatively low yields of the iodide
salts 4[I] and 5[I] are a consequence of the extensive washing with
tetrahydrofuran (THF) used to remove the I2; such treatment
dissolves some of the iodide salt and also promotes the formation
of an insoluble orange byproduct, which must be removed by
filtration.
1
1
{1H} NMR: -231.4 (t, JPP ) 453, 1P), 64.4 (d, JPP ) 453, 1P)
1
1
[lit.29 -232.7 (t, JPP ) 453), 62.9 (d, JPP ) 453)]. 13C{1H}
NMR: 30.3 (d, 1JCP ) 42), 126.2 (dd, 1JCP ) 75, 2JCP ) 8), 130.2
(s), 133.0 (s), 134.0 (s). H NMR: 3.54 (d, JHP ) 16, 4H), 7.58
(m, 12H), 7.87 (m, 8H). Dp: 149-150 °C. Anal. Calcd for C26H24-
IP3 (556.295): C, 56.14; H, 4.35. Found: C, 54.92; H, 4.32.
1
2
(25) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, V. G.,
Jr.; Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.;
Daniels, A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.;
Barone, V.; Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo,
C.; Clifford, S.; Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.;
Morokuma, K.; Salvador, P.; Dannenberg, J. J.; Malick, D. K.; Rabuck,
A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz, J.
V.; Baboul, A. G.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz,
P.; Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.;
Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.;
Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Andres, J. L.;
Gonzalez, C.; Head-Gordon, M.; Replogle, E. S.; Pople, J. A. Gaussian
98, revision A.11.1; Gaussian, Inc.: Pittsburgh, PA, 2001.
(26) Becke, A. D. J. Chem. Phys. 1993, 98, 5648-5652.
(27) Perdew, J. P.; Wang, Y. Phys. ReV. B: Condens. Matter 1992, 45,
13244-13249.
(20) Schmidbaur, H. Pure Appl. Chem. 1978, 50, 19-25.
(21) Driess, M.; Aust, J.; Merz, K.; Van Wullen, C. Angew. Chem., Int.
Ed. 1999, 38, 3677-3680.
(22) Ellis, B. D.; Carlesimo, M.; Macdonald, C. L. B. Chem. Commun.
2003, 1946-1947.
(28) Reed, A. E.; Curtiss, L. A.; Weinhold, F. Chem. ReV. 1988, 88, 899-
926.
(29) Boon, J. A.; Byers, H. L.; Dillon, K. B.; Goeta, A. E.; Longbottom,
D. A. Heteroat. Chem. 2000, 11, 226-231.
(23) Pangborn, A. B.; Giardello, M. A.; Grubbs, R. H.; Rosen, R. K.;
Timmers, F. J. Organometallics 1996, 15, 1518-1520.
(24) Schmidt, S. P.; Brooks, D. W. Tetrahedron Lett. 1987, 28, 767-768.
Inorganic Chemistry, Vol. 45, No. 17, 2006 6865

Ellis and Macdonald
Preparation of 4[I] from P2I4. 4[I] can also be prepared from
P2I4 (0.796 g, 1.398 mmol) and dppe (1.114 g, 2.795 mmol) in a
1:2 ratio in a manner similar to the preparation from PI3. Yield:
39% (0.603 g, 1.084 mmol). Identical spectroscopic data are
observed as when 4[I] is prepared from PI3.
mL) followed by filtration through Celite. Slow evaporation of the
solvent afforded a colorless crystalline material. Yield: 69% (0.126
1
g, 0.213 mmol). 31P{1H} NMR: -210.0 (t, JPP ) 424, 1P), 22.7
(d, 1JPP ) 424, 2P). 19F{1H} NMR: -79.1 (s). 13C{1H} NMR: 19.4
1
1
1
(s), 25.5 (d, JCP ) 48), 121.7 (q, JCF ) 321), 126.0 (dd, JCP
)
76, 2JCP ) 11), 130.2 (s), 132.6 (s), 133.9 (s). 1H NMR: 2.57 (m,
2H), 3.04 (m, 4H), 7.47 (m, 8H), 7.65 (m, 12H). Dp: 177-178
°C. Anal. Calcd for C28H26F3O3P3S‚CH2Cl2 (677.419): C, 51.42;
H, 4.17. Found: C, 52.23; H, 4.54. Crystallographic experiments
on crystals of 5[OTf] yielded data of relatively low quality; while
the identities and connectivities of the component ions are
confirmed, the metrical parameters are unreliable. Please refer to
the Supporting Information for details.
Preparation of [(dppp)P][I] (5[I]) from PI3. A colorless
solution of dppp (3.247 g, 7.872 mmol) in CH2Cl2 was added to a
red solution of PI3 (3.241 g, 7.872 mmol) in CH2Cl2 (100 mL),
which produced a reddish-orange solution and an orange precipitate.
The volatile components were removed under reduced pressure,
and the crude residue was washed with THF (3 × 30 mL). The
solid was redissolved in CH2Cl2 (75 mL) and filtered through Celite
to remove further small amounts of orange precipitate. The solvent
was again removed in vacuo, and after dissolution in MeCN and
slow evaporation of the solvent, a colorless crystalline material was
obtained. Yield: 56% (2.531 g, 4.438 mmol). 31P{1H} NMR:
-209.5 (t, 1JPP ) 423, 1P), 22.4 (d, 1JPP ) 423, 2P) [lit.29 -208.9
Preparation of [(dppp)P][PF6] (5[PF6]). To a colorless solution
of 5[I] (0.509 g, 0.892 mmol) in MeOH (10 mL) was added a
colorless solution of [K][PF6] (0.164 g, 0.892 mmol) in MeOH (10
mL), and the mixture became slightly cloudy white. The mixture
was stirred for 30 min. The volatile components were removed
under reduced pressure, and the product was extracted from the
crude reaction mixture using CH2Cl2 (15 mL) followed by filtration
through Celite. Slow evaporation of the solvent produced a colorless
crystalline material. Yield: 53% (0.277 g, 0.471 mmol). 31P{1H}
1
1
(t, JPP ) 423), 23.5 (d, JPP ) 423)]. 13C{1H} NMR: 19.5 (s),
1
1
2
25.0 (d, JCP ) 47), 125.7 (dd, JCP ) 77, JCP ) 12), 130.2 (s),
132.4 (s), 134.0 (s). H NMR: 2.56 (m, 2H), 3.05 (m, 4H), 7.44
1
(m, 8H), 7.65 (m, 12H). Dp: 96-97 °C. Anal. Calcd for C27H26-
IP3 (570.321): C, 56.86; H, 4.60. Found: C, 56.32; H, 4.99.
Preparation of 5[I] from P2I4. 5[I] can also be prepared from
P2I4 (0.454 g, 0.797 mmol) and dppp (0.658 g, 1.594 mmol) in a
1:2 ratio in a manner similar to the preparation from PI3. Yield:
17% (0.156 g, 0.274 mmol). Identical spectroscopic data are
observed as when 5[I] is prepared from PI3.
1
1
NMR: -210.2 (t, JPP ) 424, 1P), -143.8 (sept, JPF ) 711, 1P),
22.6 (d, JPP ) 424, 2P). 19F{1H} NMR: -73.4 (d, JFP ) 711).
1
1
13C{1H} NMR: 19.4 (s), 25.6 (d, JCP ) 48), 125.8 (dd, JCP
)
1
1
76, 2JCP ) 11), 130.2 (s), 132.5 (s), 134.0 (s). 1H NMR: 2.60 (m,
2H), 2.99 (m, 4H), 7.49 (m, 8H), 7.67 (m, 12H). Mp: 185-186
°C. Anal. Calcd for C27H26F6P4 (588.381): C, 55.12; H, 4.45.
Found: C, 54.96; H, 4.71.
Preparation of [(dppe)P][BPh4] (4[BPh4]). To a colorless
solution of 4[I] (0.625 g, 1.124 mmol) in MeOH (10 mL) was added
a colorless solution of [Na][BPh4] (0.384 g, 1.124 mmol) in MeOH
(10 mL), and a white precipitate formed instantly. The precipitate
was collected by filtration and washed with MeOH (5 mL). The
crude product was dissolved in MeCN, and upon slow evaporation
of the solvent, a colorless crystalline material was deposited.
Yield: 98% (0.824 g, 1.100 mmol). 31P{1H} NMR: -235.5 (t, 1JPP
Preparation of [(dppp)P][GaCl4] (5[GaCl4]). To a colorless
solution of 5[I] (0.369 g, 0.647 mmol) in MeOH (10 mL) was added
a colorless solution of [Li][GaCl4] (0.141 g, 0.647 mmol) in MeOH
(10 mL), and the mixture became slightly cloudy white. The mixture
was stirred for 30 min. The volatile components were removed
under reduced pressure, and the product was extracted from the
crude reaction mixture using CH2Cl2 (15 mL) followed by filtration
through Celite. Slow evaporation of the solvent yielded a colorless
crystalline material. Yield: 31% (0.130 g, 0.198 mmol). 31P{1H}
1
) 456, 1P), 64.4 (d, JPP ) 456, 2P). 11B{1H} NMR: -6.9 (s).
1
1
13C{1H} NMR: 29.0 (d, JCP ) 46), 122.4 (s), 125.9 (dd, JCP
)
2
75, JCP ) 9), 126.2 (s), 130.2 (s), 132.8 (s), 134.2 (s), 136.4 (s),
164.4 (q, 1JCB ) 49). 1H NMR: 2.48 (d, 2JHP ) 16, 4H), 6.80 (m,
4H), 6.93 (m, 8H), 7.33 (m, 8H), 7.60 (m, 20H). Mp: 174-175
°C. Anal. Calcd for C50H44BP3 (748.617): C, 80.22; H, 5.92.
Found: C, 77.83; H, 6.68.
1
1
NMR: -209.9 (t, JPP ) 425, 1P), 22.6 (d, JPP ) 425, 2P). 71Ga
NMR: 249 (s). 13C{1H} NMR: 19.6 (s), 25.9 (d, 1JCP ) 47), 125.6
1
2
1
(dd, JCP ) 76, JCP ) 11), 130.3 (s), 132.5 (s), 134.1 (s). H
NMR: 2.60 (m, 2H), 2.98 (m, 4H), 7.50 (m, 8H), 7.64 (m, 12H).
Mp: 195-196 °C. Anal. Calcd for C27H26Cl4GaP3 (654.950): C,
49.51; H, 4.00. Found: C, 49.56; H, 4.22.
Preparation of [(dppp)P][BPh4] (5[BPh4]). To a colorless
solution of 5[I] (0.225 g, 0.395 mmol) in MeOH (10 mL) was added
a colorless solution of [Na][BPh4] (0.135 g, 0.395 mmol) in MeOH
(10 mL), and a white precipitate formed instantly. The precipitate
was collected by filtration and washed with MeOH (5 mL). The
crude product was dissolved in MeCN, and upon slow evaporation
of the solvent, a colorless crystalline material formed. Yield: 71%
(0.214 g, 0.281 mmol). 31P{1H} NMR: -210.2 (t, 1JPP ) 424, 1P),
22.7 (d, 1JPP ) 424, 2P). 11B{1H} NMR: -6.9 (s). 13C{1H} NMR:
Preparation of [(dppp)P][BF4] (5[BF4]). To a colorless solution
of 5[I] (0.183 g, 0.321 mmol) in MeOH (10 mL) was added a
colorless solution of [Na][BF4] (0.035 g, 0.321 mmol) in MeOH
(10 mL), with no noticeable color change. The mixture was stirred
for 30 min. The volatile components were removed under reduced
pressure, and the product was extracted from the crude reaction
mixture using CH2Cl2 (15 mL) followed by filtration through Celite.
Slow evaporation of the solvent afforded a colorless crystalline
material. Yield: 66% (0.113 g, 0.213 mmol). 31P{1H} NMR:
-209.7 (t, 1JPP ) 423, 1P), 22.7 (d, 1JPP ) 423, 2P). 11B{1H} NMR:
-1.3 (s). 19F{1H} NMR: -152.9 (s). 13C{1H} NMR: 19.6 (s), 25.4
1
1
2
19.1 (s), 25.2 (d, JCP ) 47), 122.4 (s), 125.6 (dd, JCP ) 75, JCP
) 9), 126.2 (s), 130.2 (s), 132.3 (s), 134.0 (s), 136.5 (s), 164.6 (q,
1JCB ) 49). H NMR: 2.06 (m, 2H), 2.44 (m, 4H), 6.89 (m, 4H),
1
7.02 (m, 8H), 7.36 (m, 8H), 7.52 (m, 20H). Mp: 168-169 °C.
Anal. Calcd for C51H46BP3 (762.643): C, 80.32; H, 6.08. Found:
C, 79.66; H, 6.39.
Preparation of [(dppp)P][OTf] (5[OTf]). To a colorless solu-
tion of 5[I] (0.176 g, 0.309 mmol) in MeOH (10 mL) was added
a colorless solution of [Ag][OTf] (0.079 g, 0.309 mmol) in MeOH
(10 mL), and a yellow precipitate formed instantly. The volatile
components were removed under reduced pressure, and the product
was extracted from the crude reaction mixture using CH2Cl2 (15
1
1
2
(d, JCP ) 48), 125.9 (dd, JCP ) 77, JCP ) 10), 130.1 (s), 132.6
(s), 133.8 (s). 1H NMR: 2.59 (m, 2H), 3.14 (m, 4H), 7.47 (m, 8H),
7.65 (m, 12H). The results of several single-crystal diffraction
experiments suggest that the crystals contain a mixture of [BF4]
and [I] anions; thus, microanalysis experiments were not undertaken.
Please refer to the Supporting Information for details.
X-ray Crystallography. Crystals were coated in Nujol, Fluoro-
Lube, or Paratone-N, mounted on a glass fiber, and placed in the
6866 Inorganic Chemistry, Vol. 45, No. 17, 2006

Phosphorus(I) Iodide
Table 1. Summary of X-ray Crystallographic Data for Compounds 4[I], 4[BPh4], 5[I], 5[BPh4], 5[PF6], 5[GaCl4], and 6
compd
empirical formula
fw
4[I]
5[I]
4[BPh4]
C50H44BP3
748.22
5[BPh4]
C51H46BP3
762.60
5[PF6]
C27H26F6P4
588.36
5[GaCl4]
C27H26Cl4GaP3
654.91
6
C26H24IP3
556.26
173(2)
0.710 73
triclinic
P1h
C27H26IP3
570.29
173(2)
0.710 73
monoclinic
P21/c
C26H24I4P2
905.99
173(2)
0.710 73
orthorhombic
Aba2
T (K)
173(2)
173(2)
173(2)
173(2)
wavelength (Å)
cryst syst
space group
unit cell dimens
a (Å)
0.710 73
monoclinic
P21/n
0.710 73
monoclinic
P21/c
0.710 73
monoclinic
P21/c
0.710 73
monoclinic
P21/c
10.1179(13)
13.5649(17)
19.688(3)
71.970(1)
81.280(1)
68.866(1)
2394.2(5)
4
10.9315(10)
19.2143(18)
12.9085(12)
90
110.931(2)
90
2532.4(4)
4
1.496
12.0031(8)
17.5742(12)
19.6302(13)
90
100.893(1)
90
4066.3(5)
4
1.223
12.244(3)
17.883(3)
21.258(4)
90
114.957(10)
90
4220.0(15)
4
1.200
12.690(8)
15.736(9)
17.815(8)
90
131.57(3)
90
2662(3)
4
1.468
9.7274(16)
14.815(3)
21.255(3)
90
106.894(6)
90
2930.9(9)
4
1.484
15.2158(8)
15.2876(8)
12.3342(6)
90
90
90
2869.1(3)
4
2.097
4.469
1688
b (Å)
c (Å)
R (deg)
â (deg)
γ (deg)
V (Å3)
Z
D
calcd (g cm-3
)
1.543
1.549
1112
abs coeff (mm-1
F(000)
)
1.467
1144
0.181
1576
0.176
1608
0.342
1208
1.483
1328
θ range for data
collection
1.09-27.50
1.99-27.49
1.57-27.50
1.55-27.50
2.00-27.50
1.70-27.50
2.51-27.49
(deg)
limiting indices
-13 e h e 13
-17 e k e 17
-25 e l e 25
26 843
10 586
0.0330
-13 e h e 14
-24 e k e 24
-16 e l e 16
24 349
5809
0.0229
-15 e h e 15
-22 e k e 22
-25 e l e 25
39 288
9355
0.0350
-15 e h e 14
-20 e k e 23
-27 e l e 27
29 565
9692
0.1148
-16 e h e 16
-20 e k e 19
-22 e l e 21
21 301
5693
0.0654
-12 e h e 12
-19 e k e 19
-26 e l e 26
40 601
6663
0.0442
-19 e h e 19
-19 e k e 19
-15 e l e 15
13 423
3294
0.0245
reflns colld
indep reflns
Rint
data/restraints/param
10586/0/585
1.299
R1 ) 0.0596
5809/0/280
1.033
R1 ) 0.0206
9355/0/529
1.011
R1 ) 0.0435
9692/0/496
0.962
R1 ) 0.0626
5693/0/334
1.199
R1 ) 0.0938
6663/0/316
1.148
R1 ) 0.0505
3294/1/163
1.055
R1 ) 0.0195
GOF on F 2
final R indicesa
[I > 2σ(I)]
wR2 ) 0.1177
R1 ) 0.0694
wR2 ) 0.1240
1.348, -0.914
wR2 ) 0.0507
R1 ) 0.0241
wR2 ) 0.0522
0.653, -0.253
wR2 ) 0.1121
R1 ) 0.0578
wR2 ) 0.1238
0.402, -0.260
wR2 ) 0.1008
R1 ) 0.2144
wR2 ) 0.1460
0.263, -0.202
wR2 ) 0.2195
R1 ) 0.1231
wR2 ) 0.2299
0.629, -0.5
wR2 ) 0.1010
R1 ) 0.0691
wR2 ) 0.1115
0.588, -0.294
wR2 ) 0.0469
R1 ) 0.0202
wR2 ) 0.0473
0.857, -0.360
R indices (all data)
largest difference map
peak and hole
(e Å-3
)
N2 boil-off stream of the Kryoflex low-temperature apparatus.
Reflection data were integrated from frame data obtained using a
hemisphere of scans on a Bruker APEX CCD diffractometer using
Mo KR radiation (λ ) 0.710 73 Å). Diffraction data and unit cell
parameters were consistent with the assigned space groups. Lorentz
and polarization corrections and empirical absorption corrections,
based on redundant data at varying effective azimuthal angles, were
applied to the data sets using SAINTPlus30 and SADABS31 software.
The structures were solved by direct methods using SHELXS32 or
SIR9733 (as implemented in the WinGX package34), completed by
subsequent Fourier syntheses, and refined with full-matrix least-
squares methods against F 2 data using SHELXL.35 All non-H atoms
were refined anisotropically, and all H atoms were placed in
appropriate geometrically calculated positions. Thermal ellipsoid
plots of the molecular structures were generated using SHELXTL.36
The experimental details for each of the diffraction experiments
are listed in Table 1.
phenium salts 1 through the reduction of PCl3 in the presence
of trihydrocarbylphosphines or trisamidophosphines. The
reducing agents used in these reactions were either SnCl2 or
an excess amount of the phosphine ligands (vide infra), and
the resultant salts, containing anions such as [SnCl6]2-,
[AlCl4]-, and [BPh4]-, were separated by fractional crystal-
lization from the byproducts.12,37,38 Some of the byproducts,
such as [ClP(NMe2)3][BPh4], have solubility characteristics
very similar to those of the desired salts; thus, the isolation
of high-purity triphosphenium salts is difficult, and the yields
are relatively small using this approach.
While the triphosphenium cations appear to be a reason-
able source of PI centers, some of the anions obtained using
Schmidpeter’s original methods are not suitable for many
of the reactivity studies that we wish to pursue. For example,
salt containing tetrachloroaluminate anions may be suscep-
tible to degradation by nucleophiles, the salts containing such
metalate anions may also be air- or moisture-sensitive, and
the insolubility of certain triphosphenium salts could hinder
our investigations. For this reason, we sought a convenient
and relatively high-yield synthesis of pnictogen(I) halides
with the hope that such compounds would be useful reagents
for the metathetical preparation of salts with any desired
counteranion.
Results and Discussion
Syntheses of 4[I] and 5[I]. In the 1980s, Schmidpeter and
co-workers reported the synthesis of a variety of triphos-
(30) SAINTPlus; Bruker AXS Inc.: Madison, WI, 2001.
(31) SADABS; Bruker AXS Inc.: Madison, WI, 2001.
(32) Sheldrick, G. M. SHELXS-97; Universitat Gottingen: Gottingen,
Germany, 1997.
(33) Altomare, A.; Burla, M. C.; Camalli, M.; Cascarano, G.; Giacovazzo,
C.; Guagliardi, A.; Polidori, G.; Spagna, R. SIR97; CNR-IRMEC:
Bari, 1997.
(34) Farrugia, L. J. J. Appl. Crystallogr. 1999, 32, 837-838.
(35) Sheldrick, G. M. SHELXL-97; Universitat Gottingen: Gottingen,
Germany, 1997.
(37) Schmidpeter, A.; Lochschmidt, S.; Sheldrick, W. S. Angew. Chem.,
Int. Ed. Engl. 1985, 24, 226-227.
(38) Schmidpeter, A.; Lochschmidt, S. Angew. Chem., Int. Ed. Engl. 1986,
25, 253-254.
(36) Sheldrick, G. M. SHELXTL; Bruker AXS Inc.: Madison, WI, 2001.
Inorganic Chemistry, Vol. 45, No. 17, 2006 6867

Ellis and Macdonald
between the cations and anions; the shortest P-I distance is
4.702(3) Å, and the closest contacts to the I anions are, in
fact, from various H atoms on the dppe ligands; these
contacts are all in excess of 3.055(2) Å. The metrical
parameters of the cation, listed in Table 2, are similar to
those that have been observed in Schmidpeter’s original
hexachlorostannate salt of the [(dppe)P]+ cation, 42[SnCl6].12
For the cation containing P(11), the P-P distances of 2.1305-
(17) and 2.1260(17) Å [the corresponding values for the
cation containing P(21) are 2.1328(17) and 2.1286(16) Å]
compare well with the average distance of 2.133 Å found
between dicoordinate and tetracoordinate P atoms in orga-
nophosphorus compounds reported in the Cambridge Struc-
tural Database (CSD).40 It is noteworthy that these distances
are significantly shorter than typical P-P single-bond lengths
such as those found in Me2P-PMe2 [2.212(1) Å]41 or Ph2P-
PPh2 [2.217(1) Å],42 and they are likewise considerably
shorter than the average distance of 2.215 Å between
tricoordinate and tetracoordinate P atoms for organophos-
phorus compounds in the CSD.40 The unusual shortness of
the P-P bonds in cations such as 4, even when differences
in the covalent radii and coordination number are taken into
account, is generally attributed to the presence of partial
multiple bonding; the nature of the bonding in such cations
is discussed in more detail in the Theoretical Results and
Predictions section. The angle at the dicoordinate P atom
89.35(6)° [88.45(6)°] is notably small because of the
geometric constraints of the dppe chelate.
In a similar manner, the room temperature reaction of PI3
with dppp in dichloromethane also produces a reddish-orange
solution with orange precipitate. Following stirring for 2 h
and removal of the solvent, the crude product can be rinsed
with THF to remove I2. Extraction in dichloromethane affords
colorless solutions that yield a colorless solid upon concen-
tration. Signals in the multinuclear NMR spectra of the solid
are attributable to the cation [(dppp)P]+ 5. The microcrys-
talline solid is air- and moisture-stable and can be recrystal-
lized from MeCN to obtain material suitable for single-crystal
X-ray diffraction. The salt crystallizes in the space group
P21/c, the molecular structure is shown in Figure 2, and
relevant metrical parameters are listed in Table 2. As
observed in 4[I], the P-P bond distances in 5[I] are shorter
than the typical single bonds and are similar to those
observed in the only two other salts containing cations of
this type, namely, 52[SnCl6] and 5[AlCl4].29,43 The P-P-P
angle of 97.77(2)° is wider than the corresponding angle in
4[I] as one would expect for the six-membered ring produced
by the larger and more flexible dppp chelate.
Figure 1. Thermal ellipsoid plot (30% probability surface) of 4[I]. H atoms
are omitted for clarity.
Because the reduction of PI3 should be easier than that of
any of the lighter phosphorus trihalides, we chose to
concentrate on the generation of iodide salts. In this vein,
the room-temperature reaction of PI3 with dppe in dichlo-
romethane (eq 1) rapidly produces a burgundy solution. It
is important to note that the use of the non-nucleophilic
solVent is necessary to aVoid excessiVe formation of byprod-
ucts. After 2 h of stirring, the volatile components were
removed under reduced pressure and the solid was washed
rapidly with THF to remove the resultant I2.39 It must be
noted that while the formation of the cation is quantitative
according to 31P NMR spectra of the reaction mixture, the
concentration of the reaction mixture and, particularly, the
washing of the crude solid with THF to remove the I2 results
in the formation of an orange byproduct and reduces the
isolated yield. Redissolution of the crude solid in dichlo-
romethane and filtration to remove any of the insoluble
orange byproduct affords pale yellow solutions that deposit
colorless solid upon concentration. Multinuclear NMR
spectra of the solid exhibit signals that are attributable to
the cation [(dppe)P]+ 4. Negative-ion electrospray mass
spectrometry experiments and microanalytical data for the
product are consistent with the proposed composition as an
iodide salt and not a triiodide salt. It should be noted that
4[I] is stable in air and can be washed with water in the
open atmosphere.
The proposed composition of the salt is also supported
by the results of an X-ray crystallographic investigation.
Whereas the precipitated salts of 4[I] are generally micro-
crystalline materials that are unsuitable for single-crystal
diffraction studies, larger crystals of 4[I] may be obtained
by recrystallization from MeCN. The salt crystallizes in the
space group P1h with two independent ion pairs in the
asymmetric unit, as illustrated in Figure 1. One notable
feature of the structure is the lack of any close interactions
While the production of PI derivatives from the direct
reaction of phosphorus trihalides with various phosphine
donors has been demonstrated unambiguously using 31P
(40) Allen, F. H. Acta Crystallogr., Sect. B: Struct. Sci. 2002, B58, 380-
388.
(41) Mundt, O.; Riffel, H.; Becker, G.; Simon, A. Z. Naturforsch., B: Chem.
Sci. 1988, 43, 952-958.
(42) Dashti-Mommertz, A.; Neumuller, B. Z. Anorg. Allg. Chem. 1999,
625, 954-960.
(43) Deng, R. M. K.; Goeta, A. E.; Dillon, K. B.; Thompson, A. L. Acta
Crystallogr., Sect. E: Struct. Rep. Online 2005, E61, m296-m298.
(39) EI-MS. Calcd for I2 [M]+: m/z 253.81. Found: m/z 253.61.
6868 Inorganic Chemistry, Vol. 45, No. 17, 2006

Phosphorus(I) Iodide
Table 2. Selected Metrical Parameters for Compounds 4[I], 4[BPh4], 5[I], 5[BPh4], 5[PF6], and 5[GaCl4]
4[I]
5[I]
4[BPh4]
5[BPh4]
5[PF6]
5[GaCl4]
P(1)-P(2)
2.1263(17) [2.1223(18)]
2.1315(18) [2.1240(18)]
1.821(5) [1.820(5)]
1.826(5) [1.837(5)]
1.520(7) [1.500(7)]
2.1318(6)
2.1203(6)
1.8068(16)
1.8112(16)
1.533(2)
2.1166(6)
2.1293(6)
1.8164(17)
1.8430(16)
1.521(2)
2.1224(13)
2.1326(14)
1.807(3)
1.804(3)
1.523(4)
2.122(2)
2.113(2)
1.811(6)
1.815(6)
1.521(9)
1.526(9)
97.19(9)
2.1276(12)
2.1211(12)
1.815(3)
1.816(3)
1.532(5)
P(2)-P(3)
P(1)-C(5) [or -C(6)]
P(3)-C(4)
C(4)-C(5)
C(5)-C(6)
1.531(2)
97.77(2)
1.523(4)
93.98(5)
1.530(4)
93.76(5)
P(1)-P(2)-P(3)
88.37(7) [89.46(7)]
86.52(2)
nium salt during the reaction of PI3 with either of the
chelating ligands when the reactions are conducted in
dichloromethane. Under such conditions, significant amounts
of iodophosphonium iodide byproducts are only observed if
the initial reaction mixtures are allowed to stir for extended
periods. In contrast, significant amounts of the iodophos-
phonium byproduct are generated when the reactions are
performed in donor solvents such as THF or Et2O. Given
the foregoing, we believe that the experimental observations
suggest that the iodophosphonium iodide byproduct is
attributable to the reaction of 2 equiv of I2 and dppe.48 In
this vein, we have verified that the reaction of 2 equiv of I2
with dppe occurs in dichloromethane and produces color
changes that are consistent with those observed after the
formation of 4[I] in reactions containing excess dppe. To
confirm the NMR assignment, a mixture of dppe and 2 equiv
of I2 was concentrated to obtain yellow crystals that were
suitable for analysis by single-crystal X-ray diffraction; the
molecular structure of 6 is depicted in Figure 3.
Figure 2. Thermal ellipsoid plot (30% probability surface) of 5[I]. H atoms
are omitted for clarity.
NMR methods,29 the fact that these reactions proceed in such
a manner is surprising in terms of both the known chemistry
of P and that of its heavier congeners. For example, the
reaction of equimolar amounts of PMe3 and PBr3 results in
the formation of the Lewis acid-base adduct Me3P f PBr3,42
which is perhaps a more predictable result.44 Similarly, the
reaction of arsenic trihalides with mono- and bidentate
phosphine or arsine donors also results in the formation of
acid-base adducts in certain cases45 or in the formation of
the arsenic analogues of 1 in other instances.10,46,47 In this
light, the rapid and spontaneous nature of the reactions of
PI3 with the chelating donors outlined above is perhaps
unexpected and is certainly dependent on the stoichiometries
and conditions used for the reaction.
In 1985, Schmidpeter and associates demonstrated that the
acyclic triphosphenium salt [(Ph3P)2P][AlCl4] is obtained by
the reaction of PCl3 with 3 equiv of PPh3 in the presence of
AlCl3; the byproduct isolated from the reaction was [ClPPh3]-
[AlCl4], which indicates that 1 equiv of PPh3 is oxidized
during the course of the reaction.37 Likewise, in 1986, the
same researchers showed that the reaction of PCl3 with 3
equiv of P(NMe2)3 in the presence of [Na][BPh4] proceeds
in a similar manner to yield [((Me2N)3P)2P][BPh4] and [ClP-
(NMe2)3][BPh4].38 More recently, the spectroscopic inves-
tigations of Dillon and associates illustrated the spontaneous
nature of the reduction of PX3 (X ) Cl, Br, I) to PI cations
in the absence of auxiliary reducing agents.29 In each of these
cases, the completely reasonable suggestion has been that a
phosphine ligand acts as the redox couple for the reduction
of PIII to PI in that it is oxidized to the corresponding
halophosphonium halide (PIII to PV).
Figure 3. Thermal ellipsoid plot (30% probability surface) of 6. H atoms
are omitted for clarity.
The dppeI4 complex crystallizes in the space group Aba2,
and the individual molecules are linked to one another
through a series of long (>3.0 Å) H-I intermolecular H
bonds. The metrical parameters of the individual adducts are
similar to those of other phosphine-iodine adducts. For
example, the P-I bond length of 2.4712(1) Å and the I-I
While a PIII to PV redox couple is certainly plausible, the
31P NMR spectra of the reaction mixtures described in this
work do not show signals attributable to the iodophospho-
(44) Burford, N.; Ragogna, P. J. J. Chem. Soc., Dalton Trans. 2002, 4307-
4315.
(45) Hill, N. J.; Levason, W.; Reid, G. J. Chem. Soc., Dalton Trans. 2002,
1188-1192.
(46) Gamper, S. F.; Schmidbaur, H. Chem. Ber. 1993, 126, 601-604.
(47) Ellis, B. D.; Macdonald, C. L. B. Inorg. Chem. 2004, 43, 5981-5986.
(48) Godfrey, S. M.; Kelly, D. G.; McAuliffe, C. A.; Mackie, A. G.;
Pritchard, R. G.; Watson, S. M. J. Chem. Soc., Chem. Commun. 1991,
1163-1164.
Inorganic Chemistry, Vol. 45, No. 17, 2006 6869

Ellis and Macdonald
distance of 3.1482(4) Å in 6 are comparable to the corre-
sponding P-I [2.481(4) Å] and I-I [3.16(2) Å] distances
found for Ph3P-I2.48 While the structural features of the new
adduct are unexceptional, the crystallographic investigation
confirms the composition of 6, which is a material that is
soluble in dichloromethane and that produces a signal at 51
ppm in the 31P{1H} NMR spectrum. Most importantly, this
result confirms that 6 is not formed in the preparations of
4[I] or 5[I] using the procedures described above.
Given the preceding observations and arguments, it is
possible that the experimental observation of byproducts of
the form [X-PR3]+ (X ) Cl, Br) in the reduction of PX3
results from the transient in situ formation of X2. However,
because Cl2 and Br2 are much more potent oxidizing agents
than I2 and will react more rapidly with PR3 to form cations
with stronger P-X bonds, it is unlikely that the intermediate
formation of X2 would be observed experimentally under
conditions similar to those used for the synthesis of 4[I]. In
any event, in terms of being able to separate the PI salt from
the byproducts, the use of PI3 as the P source in such
preparative experiments allows one to obtain both greater
yields and products of higher purity.
It is worth noting that insight into the mechanism of the
formation of phosphorus(I) iodide in the process outlined in
eq 1 is provided by the solution behavior of the phosphorus
triiodide starting material itself. Variable-temperature 31P
NMR experiments reveal that, in solution, PI3 is in equilib-
rium with P2I4 and I2. We postulate that a further dispro-
portionation to oligomeric (P-I)n may occur to provide the
“P-I” fragments that are complexed by the dppe or dppp
chelates. In accord with the results of previous investigations
of the interaction of Lewis bases with halophosphines, this
additional disproportionation is undoubtedly facilitated or
enabled by the presence of the chelating ligands.49 While
we have, as yet, only obtained circumstantial evidence for
the formation of (P-I)n,50 the existence of (P-Cl)6 and (P-
Br)6 at low temperature has been reported,51 and our
investigations of the As analogue of 4[I] reveal that the
oxidative cleavage of the dppe ligand produces an ion
containing an (As-I)6 ring from the formal release of “As-
I” fragments.47 Regardless of the intermediacy of (P-I)n, we
have verified that the reaction of P2I4 with appropriate
amounts of either dppe or dppp yields 4[I] or 5[I] with the
concomitant elimination of I2.
salts, respectively. It must be noted that the use of solvents
other than methanol for such reactions generally gives
products in significantly lower yield and of lower purity.
Following extraction into dichloromethane, the colorless
solution of the desired salt is filtered from the metal iodide
byproduct and subsequent removal of all volatile compounds
under reduced pressure provides colorless solid salts in high
yield.
NMR studies of the salts demonstrate the integrity of either
the [(dppe)P]+ or the [(dppp)P]+ cation in every case and
suggest that there is little interaction between the anions and
cations in solution. Although, in some instances, there are
minor changes in the chemical shift of the PI center
depending on the identity of the anion in the salt, the
differences from the chemical shift observed for the iodide
salt are small in any case. Interestingly, salts containing air-
stable counteranions such as tetraphenylborate or triflate
appear to be stable for at least several hours upon exposure
to air either as solids or as solutions in CD2Cl2. It should be
emphasized that this behavior is in stark contrast to other PI
compounds (e.g., phosphinidenes) or dicoordinate PIII com-
pounds (e.g., phosphaalkenes, iminophosphines, phosphe-
nium salts). Examples of reactions that illustrate the need
for robust and nonreactive counteranions for the systematic
study of the chemistry of PI cations, such as ligand
replacement chemistry,52 will be detailed in future publica-
tions.
Each of the salts formed by the anion exchange reactions
was characterized completely by multinuclear NMR and
other suitable analytical methods; in the cases of the
tetraphenylborate, triflate, hexafluorophosphate, and tetra-
chlorogallate salts, crystals suitable for examination using
X-ray crystallography were also obtained by recrystallization
from either acetonitrile or dichloromethane.
The tetraphenylborate salt 4[BPh4] crystallizes in the space
group P21/c; the molecular structure is depicted in Figure 4,
and pertinent metrical parameters are assembled in Table 2.
The solid-state arrangement of the salt consists of discrete
cations and anions, and there are no unusually short contacts
between any of the ions. The metrical parameters for the
anion correspond to those anticipated for an undistorted
tetraphenylborate anion, and the metrical parameters for the
cation are entirely consistent with those found in 4[I].
Crystals of suitable quality for analysis by single-crystal
X-ray diffraction were likewise obtained for the tetraphe-
nylborate salt 5[BPh4], which crystallizes in the space group
P21/c. The molecular structure is depicted in Figure 5, and
relevant metrical parameters are compiled in Table 2. Once
again there are no unusually close contacts between the cation
and anion. The metrical parameters of the tetraphenylborate
Anion Exchange. As anticipated, the I- anion in 4[I] or
5[I] is readily replaced using metathesis reactions with a
variety of alkali-metal or Ag salts. As depicted in eq 2, the
reactions of 4[I] or 5[I] with [Na][BPh4], [Ag][OTf], [K]-
[PF6], [Li][GaCl4], or other suitable salts in methanol proceed
rapidly at room temperature to produce the desired tetraphe-
nylborate, triflate, hexafluorophosphate, or tetrachlorogallate
(49) Kirsanov, A. V.; Gorbatenko, Z. K.; Feshchenko, N. G. Pure Appl.
Chem. 1975, 44, 125-139.
(50) Preliminary results suggest that (PI)n is the primary component of the
insoluble orange precipitate formed in these reactions. Further
investigations to elucidate the structure and reactivity of this material
are currently underway.
(51) Baudler, M.; Grenz, D.; Arndt, U.; Budzikiewicz, H.; Feher, M. Chem.
Ber. 1988, 121, 1707-1709.
(52) Ellis, B. D.; Dyker, C. A.; Decken, A.; Macdonald, C. L. B. Chem.
Commun. 2005, 1965-1967.
6870 Inorganic Chemistry, Vol. 45, No. 17, 2006

Phosphorus(I) Iodide
Figure 6. Thermal ellipsoid plot (30% probability surface) of 5[GaCl4].
H atoms are omitted for clarity.
Figure 4. Thermal ellipsoid plot (30% probability surface) of 4[BPh4]. H
atoms are omitted for clarity.
Figure 7. Thermal ellipsoid plot (30% probability surface) of 5[PF6]. H
atoms are omitted for clarity.
Figure 5. Thermal ellipsoid plot (30% probability surface) of 5[BPh4]. H
atoms are omitted for clarity.
anion are consistent with those expected, and those of the
cation are similar to those observed in the structure of 5[I].
In a similar manner, the tetrachlorogallate and hexafluo-
rophosphate salts of 5, 5[GaCl4], and 5[PF6] also gave
crystals suitable for crystallographic analysis. Both salts
crystallized in the monoclinic space group P21/c. Figures 6
and 7 show the molecular structures, and pertinent metrical
parameters are summarized in Table 2. The anions in both
salts exhibit metrical parameters that are unexceptional and
that do not suggest any strong interactions with the cations;
the cations are again consistent with those observed in the
iodide and tetraphenylborate salts described above. The salt
5[PF6] is also interesting in that it is an example of a single
compound containing P atoms in two-, four-, and six-
coordinate environments.
Figure 8. Model cation systems 7-18 investigated using DFT methods.
) 1,2-bis(dimethylphosphino)ethane, 10], amine (NH3, 11;
NMe3, 12; en ) ethylenediamine, 13; tmeda ) tetrameth-
ylethylenediamine, 14), or both phosphine and amine (PH3/
NH3, 15; PMe3/NMe3, 16; H2P(CH2)2NH2, 17; Me2P(CH2)2-
NMe2, 18) donor groups to the PI center. For completeness,
these model compounds include both acyclic and cyclic
examples and are depicted in Figure 8. A summary of
important calculated values is presented in Table 3.
Theoretical Results and Predictions. In an effort to
explain the remarkable stability of the univalent P cations
that we have examined experimentally, we performed a series
of DFT calculations. To assess the importance of the nature
of the donor ligands on the stabilities of the complexes, the
model compounds were chosen to contain either phosphine
[PH3, 7; PMe3, 8; dhpe ) 1,2-bis(phosphino)ethane, 9; dmpe
Inorganic Chemistry, Vol. 45, No. 17, 2006 6871

Ellis and Macdonald
Table 3. Summary of Computational Results for Compounds 7-18
corrected energy,
EHOMO
(eV)
ELUMO
(eV)
P-P
(Å)
P-N E-P-E occ Pb
q(P)c
(au)
P-P
P-N
P-P
P-N
P(3p)
(%)f
model sym
Etotala (au)
(Å)
(deg)
(3p)
WBId
WBId
BOe
BOe
7
8
9
C2V
C2V
C2
C2
C2V
C2
C2
C2
Cs
-1027.159 94
-1262.911 80
-1104.550 33
-1261.723 64
-454.013 14
-689.625 46
-531.371 69
-688.450 28
-740.586 20
-976.267 98
-817.959 86
-975.084 67
-10.852 -5.919 2.153
-9.342 -4.126 2.150
-10.405 -5.775 2.152
-9.406 -4.207 2.149
-9.700 -5.787
97.0
105.0
87.8
90.2
93.9
107.3
86.3
90.6
97.2
105.9
87.1
90.2
1.721 -0.22 1.0979
1.729 -0.32 1.0971
1.673 -0.22 1.1014
1.664 -0.27 1.1070
0.8328
0.8166
0.8579
0.9031
86.398
85.976
83.949
83.328
10
11
12
13
14
15
16
17
18
1.925
1.944
1.931
1.926
1.947
1.938
1.948
1.924
1.789
1.800
1.725
1.764
0.31
0.30
0.30
0.33
0.6274
0.5836
0.6106
0.5877
0.4947 97.535
0.4483 96.993
0.4760 97.578
0.4693 96.318
-8.746 -4.231
-9.352 -5.530
-8.748 -4.809
-10.404 -6.376 2.142 1.934
-9.097 -4.837 2.151 1.930
-10.014 -5.894 2.142 1.934
-9.160 -4.968 2.144 1.935
0.08 1.1438 0.6433 0.8556 0.4878 89.736
0.02 1.1311 0.6027 0.8034 0.4855 90.046
0.08 1.1296 0.6330 0.8601 0.4770 88.740
0.06 1.1339 0.6041 0.9032 0.4660 87.596
C1
C1
C1
a Etotal ) Ecalcd + ZPVE. b NBO population of the 3px orbital on the dicoordinate P atom. c NBO charge on the dicoordinate P atom. d NBO Wiberg bond
indices for the specified bonds. e NBO atom-atom overlap-weighted natural atomic orbital bond order for the specified bonds. f NBO NLMO percentage
contribution of the 3px orbital on the dicoordinate P atom to the “lone-pair” or delocalized “lone-pair” orbital.
The calculated metrical parameters in models 7-10 are
in reasonably good agreement with those that have been
determined experimentally given the simplified nature of the
models and attest to the viability of the chosen models.
Although the calculated P-P distances tend to be slightly
longer than those observed in the crystallographic experi-
ments outlined above, the values are still much shorter than
those of typical single P-P bonds.53 In contrast, the models
employing amine substitutents (11-14) have computed bond
lengths that are much longer than typical P-N single bonds
imaginary frequencies and are thus not true minima on their
potential energy surfaces. In the literature, triphosphenium
cations are most typically depicted using drawings such as
1b, in which the dicoordinate P atom has one “lone pair” of
electrons and is thus a PIII center. Model 1b also suggests
the presence of one double bond delocalized between the
dicoordinate P center and the P atoms of the phosphine
substituents and could thus account for the short P-P
distances observed experimentally. A third Lewis-type
structure that is sometimes used to depict triphosphenium
cations is a bis-ylidic model illustrated by 1c, in which the
dicoordinate P atom bears two “lone pairs” of electrons and
is thus able to be described as a PI center. An alternative
but equivalent depiction of 1c is 1d, which may be used to
emphasize that such molecules can be considered as coor-
dination compounds of PI cations.
54
[cf. 1.800(4) Å for O3PNH3 and 1.696(3) Å for PhHN-
PPh255]. Although such long P-N distances may appear odd,
it must be emphasized that the covalent radius of PI will be
significantly larger than those of PIII or PV, which are
contained in the compounds containing the “typical” bonds.
The model compounds containing both phosphine and amine
donors (15-18) exhibit calculated P-P and P-N bond
lengths that are comparable to those of the corresponding
homoleptic models. The calculated P-P-P angles for the
cyclic models 9 and 10 correlate well with those observed
experimentally for salts of 4. Similarly, the calculated angles
for the acyclic triphosphenium cations 7 and 8 are also in
good agreement with reported experimental values.37
While Lewis-type drawings should certainly not be over-
interpreted, such simple models can provide insight into the
structural properties and potential reactivity of compounds
such as 1. In this vein, we examined the frontier orbitals of
model compounds 7-18 in an effort to gain insight into the
real electronic structure of triphosphenium cations and to
assess the appropriateness of models such as 1b and 1c.
Drawings of the relevant frontier orbitals for the most realistic
cyclic models 10, 14, and 18 are presented in Figure 9, and
pertinent details of the NBO analysis are listed in Table 3.
Perhaps the most important overall feature that is apparent
in Figure 9 is the general similarity of the appearance of
two highest occupied molecular orbitals (HOMO-1 and
HOMO) in each model regardless of the composition of the
stabilizing ligand. In every case, the HOMO-1 orbital is of
a-type symmetry and is consistent with a “lone pair” orbital
on the dicoordinate P atom in the E-P-E plane (where E
) N or P depending on the model). Similarly, the HOMO
in each of the models is a b-type symmetry (for the C2
models) orbital that is almost exclusively composed of the
atomic 3px orbital on the dicoordinate P atom. The HOMOs
The potentially ambiguous nature of the electronic struc-
ture of triphosphenium cations is emphasized by the possible
canonical drawings that may be used to depict such
compounds, as illustrated by structures 1a-d. Drawing 1a
suggests the presence of a PV center bearing no “lone pairs”
of electrons and exhibiting a linear, allene-like PdPdP
geometry. Because every structurally characterized triphos-
phenium compound has a strongly bent P-P-P moiety,
drawing 1a is not an instructive or reasonable depiction of
such cations. Accordingly, if computational models such as
7 are constrained to D3d symmetry, they exhibit several
(53) Hinchley, S. L.; Morrison, C. A.; Rankin, D. W.; Macdonald, C. L.;
Wiacek, R. J.; Voigt, A.; Cowley, A. H.; Lappert, M. F.; Gundersen,
G.; Clyburne, J. A.; Power, P. P. J. Am. Chem. Soc. 2001, 123, 9045-
9053.
(54) Cameron, T. S.; Chan, C.; Chute, W. J. Acta Crystallogr., Sect. B:
Struct. Crystallogr. Cryst. Sci. 1980, B36, 2391-2393.
(55) Wetzel, T. G.; Dehnen, S.; Roesky, P. W. Angew. Chem., Int. Ed.
1999, 38, 1086-1088.
(56) Schaftenaar, G.; Noordik, J. H. J. Comput.-Aided Mol. Des. 2000,
14, 123-134.
6872 Inorganic Chemistry, Vol. 45, No. 17, 2006

Phosphorus(I) Iodide
noted that, in main group or organic compounds, back-
bonding interactions of the type depicted in Figure 10 are
more commonly described using the term “hyperconjuga-
tion”.
Direct evidence for the presence of back-bonding interac-
tions is found in the results of the NBO population analysis.
In the absence of a back-bonding interaction, the population
of the 3px orbital on the PI center should be roughly two
electrons. In all of the models containing phosphine ligands,
back-bonding allows for the effective partial oxidation of
the PI center, which is reflected in the decreased population
of the 3px orbital on the dicoordinate P center, as shown in
Table 3. A comparable decrease in the 3px population is not
observed in the case of the models containing the corre-
sponding amine ligands in which the orbital population
remains at a value of almost 2. The magnitude of the
interaction is also confirmed by the results of natural
localized MO (NLMO) calculations, which provide a mea-
sure of the delocalization of the “lone pair” orbitals. As listed
in Table 3, these calculations show virtually no delocalization
in the amine-substituted models and significant delocalization
in the phosphine-containing models. Furthermore, the NBO
analysis indicates that the back-bonding delocalization
stabilizes the model compound [(H3P)2P]+ 7 by around 187
kJ mol-1, while the much-less-favorable interaction in the
model [(H3N)2P]+ 11 provides only 66 kJ mol-1 of stabiliza-
tion. For the cations containing both phosphine and amine
substituents, there is a decrease in the orbital population,
although less than that observed for models containing two
phosphine donors.
In addition to the differences in orbital occupancies and
energies, the presence or absence of the back-bonding
interaction is clearly manifested in the shape of the HOMO
for each of the model cations, as depicted in Figure 9. In
the case of models containing phosphine ligands, there are
perturbations of the lobe derived primarily from the 3px
orbital on the PI center toward the phosphine substituents to
produce orbitals that are consistent with the presence of a
π-bonding interaction between the P atoms. The much
smaller magnitude of the back-bonding in the amine-
substituted models is further emphasized by the presence of
a nearly undistorted 3px orbital on PI in the HOMO of the
model depicted in Figure 9.
Figure 9. MOLDEN56 depictions of the HOMOs and LUMOs for the
model cations 10, 14, and 18.
Figure 10. Orbital interactions between the empty P-R anti-bonding
orbitals and a filled 3px orbital on PI allowing for back-bonding from the
dicoordinate PI to the phosphine substituents.
thus appear almost identical with those of water and suggest
that the ylidic model 1c (or 1d) is the most appropriate
drawing in terms of the overall electronic structure.
Despite the general similarity in the appearance of the
occupied orbitals of the model compounds depicted in Figure
9 and those of the other model compounds listed in Table 3,
there are significant differences between them in terms of
their energies. In particular, whereas the energy of the lowest
unoccupied MO (LUMO) remains roughly similar in the
corresponding models regardless of whether they contain
amine or phosphine ligands, the energies of the HOMOs are
considerably lower for models containing phosphine ligands.
For example, in the series of models 9, 13, and 17, the
energies of the LUMOs are similar: -5.775, -5.530, and
-5.894 eV, respectively. In sharp contrast, the energies of
the corresponding HOMOs are significantly different, with
the values being -10.405, -9.352, and -10.014, respec-
tively. Thus, it is apparent that the presence of phosphine
donor ligands on the dicoordinate P center stabilizes the
HOMO and increases the HOMO-LUMO separation in
comparison to when amine donors are used.
Given the foregoing, an explanation for the remarkable
stability of triphosphenium cations is almost certainly related
to the stabilization of the HOMO. An explanation is provided
by the understanding that, in contrast to amines, phosphine
ligands may act as back-bonding acceptors, as they are
known to do in compounds containing phosphines ligated
to transition metals. In phosphine ligands, the empty π-type
P-R antibonding orbitals (these have e symmetry in the C3V
point group of the free PR3 ligand) are of the correct
symmetry and appropriate energy to accept electron density
from the filled 3px orbital on the PI center (Figure 10). Such
a back-bonding interaction lowers the energy of the MO thus
formed and stabilizes the electron-rich center. It should be
In addition to the effects on the shape and composition of
the HOMO, the greater magnitude of back-bonding in the
phosphine-substituted model compound [(H3P)2P]+ (7) is
highlighted by the structural consequences of preventing the
back-bonding interactions. When the important acceptor
orbitals on the phosphine ligands are removed using the
NBOdel method, the optimized P-P bond distance increases
to 2.350 Å from the 2.149 Å distance in the unperturbed
model; this is a change of more than 0.2 Å. The comparable
change in the amine-substituted analogue is less than 0.1 Å
(from 1.919 to 2.012 Å). For the mixed-ligand model system
[(H3P)P(NH3)]+ (15), the P-P bond distance changes from
2.141 to 2.364 Å, while the P-N bond distance only changes
from 1.926 to 2.022 Å when back-bonding interactions are
prevented. Thus, both structural features and energy values
Inorganic Chemistry, Vol. 45, No. 17, 2006 6873

Ellis and Macdonald
demonstrate that the magnitude of the back-bonding is clearly
more pronounced for the phosphine ligands, as one would
expect. In this light, the Lewis-type model 1b certainly
provides insight into the stability of triphosphenium cations.
In addition to the effects on the occupied orbitals, it is
notable that the composition of the LUMO is also highly
dependent on the nature of the ligands, as shown in Figure
9. For model 10, the LUMO is localized primarily on the
phosphine substituents, whereas for 14, the orbital is
principally on the dicoordinate P atom (the MO resembles a
3py atomic orbital) and, for 18, the orbital is mainly localized
over the two P atoms. Thus, the calculations suggest that
the nature of the ligand not only affects the stability of the
E-P-E framework but also should have a large impact on
the reactivity of the coordinated PI center. Overall, these
computational results suggest that there are many possible
ways to tune the reactivity of such compounds through the
judicious choice of stabilizing ligand. While there is much
more experimental work to be done in this regard, we have
already demonstrated that an understanding of the frontier
MOs can be used to control the oxidation chemistry of some
analogous AsI compounds.47
Overall, examinations of the electronic structure of several
model compounds suggest that the two HOMOs in triphos-
phenium cations correspond to “lone pairs” of electrons that
are primarily localized on the dicoordinate P atom; in this
light, the ylidic description (1c) of the bonding in such
molecules is the most appropriate. However, the HOMO in
triphosphenium cations is rendered significantly more stable
by back-bonding interactions with the adjacent phosphine
ligands so, while ylidic depictions such as 1c most accurately
describe the electronic structure, a canonical structure such
as 1b provides insight into the remarkable stability of these
compounds.
Conclusions
High-purity iodide salts of cations containing PI centers
are readily prepared and isolated in good yield by the base-
induced spontaneous elimination of I2 from PI3. The facile
exchange of the I- anion in 4[I] or 5[I] using metathesis
reactions allows for the tailoring of reactivity and solubility
of univalent P salts and will allow for further investigations
into the unique chemistry and potential uses for this easily
accessible reagent. In contrast to most other PI compounds,
many of these salts are stable even in the presence of water
and air. The remarkable stability of these PI salts is largely
attributable to the back-bonding interactions between the
ligand and the PI center, and such an understanding may be
used to tune the behavior of these reagents.
Acknowledgment. We thank the Natural Sciences and
Engineering Research Council (NSERC, Canada), the Canada
Foundation for Innovation, the Ontario Research and De-
velopment Challenge Fund, and the Ontario Innovation Trust
for financial support. B.D.E. thanks the Government of
Ontario and NSERC for postgraduate scholarships. We also
thank Michelle Carlesimo and Kevin A. Renaud for some
synthetic work, the CCMR (Windsor) for elemental analyses,
and Dr. R. J. Letcher’s laboratory (GLIER) and Dr. J. R.
Green (Windsor) for obtaining the mass spectra.
Supporting Information Available: X-ray crystallographic data
in CIF format for compounds 4[I], 4[BPh4], 5[I], 5[BPh4], 5[OTf],
5[BF4/I], 5[PF6], 5[GaCl4], and 6 are provided. Cartesian coordinates
for the geometry optimizations and a summary of calculation results
are tabulated. This material is available free of charge via the
Internet at http://pubs.acs.org.
IC060186O
6874 Inorganic Chemistry, Vol. 45, No. 17, 2006