Mechanism of phosphorus-carbon bond cleavage by lithium in tertiary phosphines. An optimized synthesis of 1,2-bis(phenylphosphino)ethane

Article abstract of DOI:10.1021/jo9907336

Conditions influencing the extent of P-C(aryl) vs P-C(alkyl) bond cleavage in the reaction of Ph₂P(CH₂)₂PPh₂ with lithium in THF have been investigated. The results complement and elucidate earlier work; they indicate that the mechanism of P-C bond cleavage in tertiary phosphines of this type involves a thermodynamic equilibrium between P-C(aryl) and P- C(alkyl) cleaved radicals and anions, followed by reaction and stabilization of these as lithium salts. The addition of water to the reaction mixture causes a reestablishment of the cleavage equilibrium prior to the formation of the secondary phosphines. A mechanism involving competitive release of leaving groups as the thermodynamically most stable anion or radical has been proposed. The preparation of (R*, R*)-(±)/(R*, S*)-PhP(H)(CH₂)₂P(H)Ph by this route has been optimized.

Full text of DOI:10.1021/jo9907336

J . Org. Chem. 2000, 65, 951-957
951
Mech a n ism of P h osp h or u s-Ca r bon Bon d Clea va ge by Lith iu m in
Ter tia r y P h osp h in es. An Op tim ized Syn th esis of
1,2-Bis(p h en ylp h osp h in o)eth a n e
J ohn Dogan,† J urgen B. Schulte,† Gerhard F. Swiegers,*,† and S. Bruce Wild‡
Department of Chemistry, University of Wollongong, Wollongong, NSW 2522, Australia, and
Research School of Chemistry, Australian National University, Canberra, ACT 0200, Australia
Received May 3, 1999
Conditions influencing the extent of P-Caryl vs P-Calkyl bond cleavage in the reaction of Ph2P(CH2)2-
PPh2 with lithium in THF have been investigated. The results complement and elucidate earlier
work; they indicate that the mechanism of P-C bond cleavage in tertiary phosphines of this type
involves a thermodynamic equilibrium between P-Caryl and P-Calkyl cleaved radicals and anions,
followed by reaction and stabilization of these as lithium salts. The addition of water to the reaction
mixture causes a reestablishment of the cleavage equilibrium prior to the formation of the secondary
phosphines. A mechanism involving competitive release of leaving groups as the thermodynamically
most stable anion or radical has been proposed. The preparation of (R*, R*)-(()/(R*, S*)-PhP(H)-
(CH2)2P(H)Ph by this route has been optimized.
Sch em e 1
In tr od u ction
Despite its synthetic importance, the mechanism of the
phosphorus-carbon bond cleavage that occurs when a
tertiary phosphine reacts with an alkali metal in tet-
rahydrofuran is poorly understood, with many of the
experimental procedures employed in these reactions
being optimized by trial and error. An important example
concerns the preparation of 1,2-bis(phenylphosphino)-
ethane, 1, from 1,2-bis(diphenylphosphino)ethane, 2.
Early syntheses of 1 are cumbersome and low yielding,1-4
so that the alkali metal cleavage of phosphorus-phenyl
groups5 from commercially available 2, followed by pro-
tonation of the resulting bis(phosphide), is the most
convenient synthesis.6 In the reaction, P-C cleavage is
effected most successfully by lithium to produce PhP(Li)-
CH2CH2P(Li)Ph, 4, although substantial quantities of
LiPPh2, 3, are generated by competitive cleavage of
P-Calkyl bonds.
A systematic investigation of lithium-induced P-C
bond cleavage in Ph2P(CH2)nPPh2 (n ) 1-5) has been
carried out, where it was observed that the regioselec-
tivity of the reaction at room temperature in THF
depended on the value of n (Scheme 1).7 For n ) 1,
P-Calkyl cleavage (pathway I) occurred, but for n ) 3-5,
P-Caryl cleavage (pathway II) predominated. In the
reaction of lithium with 2 (n ) 2) a mixture of products
from pathways I and II was obtained; the presumed
intermediate for pathway I, [Ph2PCH2CH2-Li+], was not
observed, nor was Ph2PEt, which would have been formed
by protonation of the anion by the solvent. It was
therefore assumed that the intermediate decomposed to
give Ph2PLi and ethylene, the latter being detected.7
The cleavage reactions are erratic with respect to
induction times and amounts of unreacted starting
material. The diphosphides PhP(Li)(CH2)nP(Li)Ph (n )
2-5) are also highly reactive. For example, the addition
of tert-butyl chloride to the reaction mixtures to remove
the PhLi byproduct resulted in the formation of the
corresponding 1,2-bis(tert-butyl)-substituted diphosphines.7
This contrasts with the behavior of Ph2PLi, which does
not react with tert-butyl chloride.5 Alkylation of the bis-
(phosphides) clearly competes with deprotonation of the
tert-butyl chloride by the phenyllithium. At reduced
temperatures, however, alkylation at phosphorus is
minimized.6,7
* E-mail: g.swiegers@molsci.csiro.au. Current address: CSIRO
Molecular Science, Bag 10, Clayton South, Victoria 3169, Australia.
† University of Wollongong.
‡ Australian National University.
(1) Issleib, K.; Standtke, K. Chem. Ber. 1963, 96, 279.
(2) Issleib, K.; Weichmann, H. Chem. Ber. 1968, 101, 2197.
(3) King, R. B.; Kapoor, P. N. J . Am. Chem. Soc. 1971, 93, 4158.
(4) The title compound is an important starting material for the
synthesis of a variety of ligands employed in homogeneous asymmetric
catalysis; see: Homogeneous Catalysis with Metal Phosphine Com-
plexes; Pignolet, L. H., Ed.; Plenum Press: New York, 1983.
(5) Aguiar, A. M.; Beisler, J .; Mills, A. J . Org. Chem. 1962, 27, 1001
and references therein.
(6) Chou, T.-S.; Tsao, C.-H.; Hung, S. C. J . Org. Chem. 1985, 50,
4329.
(7) Brooks, P.; Gallagher, M. J .; Sarroff, A. Aust. J . Chem. 1987,
40, 1341 and references cited therein.
These data were rationalized in terms of the relative
electronic stabilization of the phosphide ions by the
substituents. Because of its ability to delocalize negative
charge, a phenyl group stabilizes an anion better than a
methyl group, similar to an ethyl group but more poorly
than an n-propyl, n-butyl, or n-pentyl group.7 As an aryl
group is necessary before tertiary phosphines react with
10.1021/jo9907336 CCC: $19.00 © 2000 American Chemical Society
Published on Web 02/03/2000

952 J . Org. Chem., Vol. 65, No. 4, 2000
Dogan et al.
lithium,7 it was suggested that the key process in these
heterogeneous cleavage reactions involves an aryl-medi-
ated association of the phosphine with the metal surface,
followed by a slow step (or steps) leading to bond cleavage
(eq 1).
Sch em e 2
Although this mechanism accounts for P-Calkyl cleav-
age in bis(tertiary phosphines) having n ) 1, it does not
explain P-Caryl cleavage for n ) 3-5 because of the less
stable radical produced. The alternative explanation
involving release of an organic group as an anion is,
however, also inconsistent with the leaving group prefer-
ences.
ESR studies of related reactions have indicated that
some of the R2P-Li+ (R ) Me, n-Pr, n-Bu, Ph) produced
in similar cleavage reactions react with excess lithium
to give R2P2- •(Li+)2, which were detected as the only spin-
active species in solution.8,9 These short-lived radical
dianions apparently propagate by rapid intermolecular
exchange at room temperature since bimolecular or other
termination products have not been observed.9 Low levels
of radical dianions may also be responsible for the
characteristic red color of the bis(arylphosphine)-lithium
reaction mixtures described above.7
From these observations, it became evident that a
means of synthesizing 1 in high yield lay in adjusting
the conditions of the reactions in Scheme 2 so as to
maximize the formation of the bis(phosphide) 4 (pathway
II) and to minimize that of the competing mono(phos-
phide) 3 (pathway I). It has been reported10 that crystal-
line 4 is hydrolyzed by water to give 1 in 68% overall
yield.11 Isolation of the solid on a large scale is trouble-
some, however, so that a high yielding in situ preparation
of 1 from 2 would be desirable. Thus, a recent method of
preparing 1 involved stirring a mixture of 2 and lithium
wire in THF at 0 °C for 7 d and then treating the
resulting mixture with water, which gave 1 in >80%
yield.12 The conditions for the lithiation were a carefully
chosen compromise: the initial mixture was stirred for
long enough to overcome the erratic induction time for
P-Caryl cleavage, while the reduced temperature sup-
pressed the formation of Ph2PLi, 3.
concentrated mixture of 2 and lithium in THF was heated
under reflux for 3 h, the solution was filtered, and the
filtrate cooled to room temperature. The crystalline 4‚
4THF that deposited overnight was then quenched with
ammonium bromide to give 1 in 65% yield.
Despite the success of this procedure, it was clear that
the trial-and-error approach hitherto adopted for the
optimization of the reaction conditions required for the
synthesis of 1 was unsatisfactory. Here we present a
detailed investigation employing 31P{1H} NMR, UV-
visible, and ESR spectroscopy to probe the factors
influencing the conversion of 2 into 1.
In our recent work concerning the optical resolution
of the linear tetra(tertiary phosphine) (R*,R*)-(()-1,1,4,7,-
10,10-hexaphenyl-1,4,7,10-tetraphosphadecane,13 we re-
quired considerable quantities of 1 as a starting material.
Although high yields of 1 could be obtained by the
literature method,12 the process was capricious and
reaction times of >7 d were sometimes required. A lower
yielding but less time-consuming and more reliable route
was subsequently developed and reported,13 wherein a
Resu lts a n d Discu ssion
Meth od ology. A series of reactions of the type shown
in Scheme 2 were performed under different conditions
of temperature and concentration, and their progress was
monitored by 31P{1H} NMR spectroscopy. This permitted
the identification of the phosphines and phosphides
present in the reaction mixtures, because, as previously
noted, the presence of small quantities of radical dianions
does not affect the 31P{1H} NMR chemical shifts.7 Each
NMR sample was collected and sealed (under inert
atmosphere) in an NMR tube containing 1 drop of
benzene-d6, which was required to lock the NMR signal.
To correlate the resonances obtained with the species
shown in Scheme 2, each compound and intermediate
was prepared and its 31P{1H} NMR chemical shift was
determined in THF/benzene-d6. The NMR data are given
in Scheme 2 and agree with literature values.7 The
(8) Grim, S. O.; Molenda, R. P. Phosphorus 1974, 4, 189.
(9) Britt, A. D.; Kaiser, E. T. J . Phys. Chem. 1965, 69, 2775.
(10) Anderson, D. M.; Hitchcock, P. B.; Lappert, M. F.; Moss, I. Inorg.
Chim. Acta 1988, 141, 157.
(11) Brooks, P.; Craig, D. C.; Gallagher, M. J .; Rae, A. D.; Sarroff,
A. J . Organomet. Chem. 1987, 323, C1.
(12) Kimpton, B. R.; McFarlane, W.; Muir, A. S.; Patel, P. G.;
Bookham, J . L. Polyhedron 1993, 12, 2525.
(13) Airey, A. L.; Swiegers, G. F.; Willis, A. C.; Wild, S. B. Inorg.
Chem. 1997, 36, 1588.

Mechanism of Phosphorus-Carbon Bond Cleavage
J . Org. Chem., Vol. 65, No. 4, 2000 953
unsymmetrical compounds Ph2P(CH2)2P(Li)Ph and Ph2P-
(CH2)2P(H)Ph were not prepared, but resonances for
them (δ -14.6, -41.9, and -15.45, -49.8, respectively7)
were not evident in any of the experiments. A delay of
50 s between pulses was applied in recording the 31P-
{1H} NMR spectra in order to obtain accurate integra-
tions.
Exp er im en ts. Initial experiments involved the reac-
tion of 2 with lithium as described in ref 13. The addition
of a 10-fold excess of finely divided lithium wire to a
vigorously stirred solution of 2 in THF (0.20 mol L-1) at
20 °C, followed by stirring of the mixture for 2 h, led to
pale red solutions that contained significant quantities
of the bis(phosphide) 4 (ca. 10-50%) and mono(phos-
phide) 3 (ca. 10-40%), in addition to 2 (ca. 10-80%).
Solutions prepared in this way became slightly warm
when the lithium was added; bubbles of ethene were not
evident when 3 was the major product. The amounts of
3 and 4 in the mixtures increased over 3 d, but in an
erratic manner. Thus, after 3 d, yields of 3 and 4 were
20-90%. For reactions that were essentially complete
within 2 h of the addition of lithium, the proportions of
3 and 4 did not change with continued stirring for 3 d.
Moreover, heating of these mixtures to 40 °C or cooling
them to 0 °C for several hours did not result in significant
changes in the relative proportions of 3 and 4.
F igu r e 1. ESR spectrum at 20 °C of a solution of 4 prepared
by the addition of a dilute THF solution of 2 to a suspension
of lithium in THF at 0 °C (g ) 2.051).
a degassed solution of THF containing a trace of water
was added dropwise to the solution of 4 at 0 °C, the peak
at 421 nm was selectively diminished in intensity,
consistent with the higher reactivity of the radical
dianion.
ESR Sp ectr oscop y. The ESR spectrum of a solution
of freshly prepared 4 exhibited a symmetrical pattern of
peaks, suggesting the presence of a single spin-active
species (Figure 1). The spectrum consists of at least 33
triplets, each separated by 0.75 G, and displays a
monomodal, Gaussian-like distribution of intensities. The
hyperfine splitting for each triplet is 0.25 G.
Previous investigations of R2P2- •(Li+)2 indicated 31P
hyperfine splittings of 11-12 G (R ) Me, n-Pr, n-Bu) and
Con cen tr a tion Dep en d en ce. Experiments at con-
centrations of 2 within the range 0.02-0.20 M and with
4.3- to 10-fold excesses of lithium produced similar
quantities of 3 or 4, within experimental error. For
solutions below 0.02 M in 2, accurate NMR determina-
tions of species concentrations could not be achieved.
1
8 G (R ) Ph), with H splittings of 2.2-2.4 G (R-alkyl; R
) n-Pr, n-Bu), 0.2-0.3 G (â-alkyl; R ) n-Pr, n-Bu), 2 G
(o- and p-phenyl; R ) Ph), and 0.8 G (m-phenyl; R )
Ph).8,9 Li7-induced splitting is absent from these spectra.
The 33 major peaks in the ESR spectrum of a freshly
prepared solution of 4 can therefore be assigned to
hyperfine splitting by the m-hydrogen atoms of a phenyl
group, that is, am(1) ) 0.75 G. The triplets into which each
of the major peaks are split arise from splittings by the
protons on the â-carbon of the 1,2-ethano group (aâ-alkyl
) 0.25 G). Because the 33 triplets are equally spaced,
the splittings induced by the remaining atoms are
multiples of 0.75 G. The monomodal intensity distribu-
tion of the peaks suggests overlapping between the
outermost resonances of adjacent triplets created by the
m-phenyl H atoms. Thus, the hyperfine splittings caused
by either the o- or p-hydrogen atoms of the phenyl group
having the most radical character, or by the R H atoms
of the bridging 1,2-ethano group, are 1.5 G. In the
absence of detailed experiments, which were beyond the
scope of this investigation, it was not possible to establish
the a values for these and the other atoms because the
31P hyperfine splittings could not be determined by simple
inspection (as was the case in previous work8,9). Iterative
simulations using likely a values in the radical PhP(Li)-
CH2CH2(Ph)P2- •(Li+)2 produced spectra that closely re-
sembled the one reproduced in Figure 1, however. By
contrast, the spectrum in Figure 1 cannot be due to
Ph2- •(Li+)2 or Ph2P2- •(Li+)2.
Tem p er a tu r e Dep en d en ce. When the reaction was
carried out at 40 °C (0.20-0.02 M 2, 10-fold Li), larger
proportions of 3 were obtained (>30%).6 At 0 °C, solutions
enriched in the bis(phosphide) 4 were obtained.6 Because
temperature was clearly a critical factor, a solution of 2
in THF was added dropwise to a cooled (0 °C) and
vigorously stirred suspension of lithium in THF.6,14 After
1 h, stirring was stopped, and the mixture was allowed
to warm to room temperature. The resulting solution
contained 4 and 2 only. The reaction mixture was then
heated under reflux for 2 h, after which the sample dis-
played the NMR resonance of 4 only. The results were
reproducible within the concentration range examined,
provided that the addition of 2 was carried out slowly
and in sufficient local dilution. When the reaction was
complete, unreacted lithium was removed by filtration
of the hot mixture. Yellow crystals of PhP(Li)CH2CH2P-
(Li)Ph‚4THF, 4‚4THF, separated from the cold fil-
trate.10,11
UV-visible Sp ectr oscop y. The UV-visible spectrum
of the reaction mixture at completion contained absorp-
tions at 366 nm (yellow) and 421 nm (red). The 366 nm
band was due to PhP(Li)CH2CH2P(Li)Ph, 4, which was
verified by recording the spectrum of 4‚4THF under
similar conditions. Because the major byproduct of the
reaction, PhLi, is colorless,15 the 421 nm absorption is
most likely due to a radical dianion, as previously
assumed.7 The known radical dianion Ph2P2- •(Li+)2, 6,
absorbs intensely at 429 nm in THF.9 Furthermore, when
Hyd r olysis of 4. It has been reported that PhLi de-
composes in THF with a half-life of ca. 12 h at 25 °C and
45 min at 65 °C.16 Thus, boiling of a solution of freshly
prepared 4 in THF for 2-3 h should have been sufficient
to decompose PhLi, leaving 4 for protonation to 1.5
The hydrolysis of a solution of 4 was not straightfor-
ward, however. The dropwise addition of water to a
(14) Bautista, M. T.; Earl, K. A.; Maltby, P. A.; Morris, R. H.;
Schweitzer, C. T. Can. J . Chem. 1994, 72, 548.
(15) Schlosser, M.; Landberger, V. J . Organomet. Chem. 1967, 8,
193.
(16) Gilman, H.; Gaj, B. J . J . Org. Chem. 1957, 22, 1165.

954 J . Org. Chem., Vol. 65, No. 4, 2000
Dogan et al.
solution of 4 at 20 °C produced a mixture containing 5
(10-50%) and 1 (50-90%), according to NMR spectros-
copy. When the hydrolysis was performed at 40 °C, 5 was
produced in greater quantities (>40%). When carried out
at 0 °C, the reaction afforded 1 in quantitative yield.
Thus, the bis(phosphide) 4 is converted into the bis-
(secondary phosphine) 1 in high yield at ca. 0 °C. At
higher temperatures, increasing amounts of 4 were
converted into 5. Similar results were obtained when the
solution of 4 in THF was allowed to stand for 24 h prior
to hydrolysis.
also appears to be more stable in THF in the presence of
4. Alone, it forms a dimer that slowly reacts with THF.19
The addition of water to a solution of 4 in THF causes
a reequilibration of the P-Caryl and P-Calkyl cleavage
products as indicated by the dramatic changes in the
proportions of products 1 and 5 at different reaction
temperatures. Thus, 5, which usually arises by pathway
I, can be obtained from 4 of pathway II if the hydrolysis
is carried out at elevated temperatures. The species
present in the reestablished equilibrium are the same
as those formed in the original cleavage reaction, because
in both cases pathway II is favored at 0 °C, pathway I is
favored at 40 °C, and pathways I and II are favored at
20 °C. The temperature dependence of this equilibrium
indicates it to be thermodynamic in nature. It appears
that the compound responsible for the reestablishment
of the equilibrium is derived from the reaction of the red
species (λmax 421 nm) with water, because 1 is formed
when water is added to a solution of PhLi and 4 at 40
°C, but 5 is formed in the presence of the red species.
The ESR and UV-visible data, in conjunction with data
from related systems, indicate that the red species is the
radical anion PhP(Li)CH2CH2(Ph)P2- •(Li+)2, 7. In com-
mon with other radical anions,20 7 will be a powerful
reducing agent and readily protonated by water. The
most likely product of the hydrolysis of 7, and the only
one that could reestablish the cleavage equilibrium, is
the phosphinyl radical PhP(Li)CH2CH2(Ph)P•, which
would be formed in a two-electron transfer from 7 to
water (eq 5). One-electron transfer from 7 would produce
H• and 4, the latter already being present in the reaction
mixture and known to be unaffected by temperature.
Direct abstraction of H+ from water by 7 would produce
PhP(Li)CH2CH2(Ph)PH- • (Li+) or PhP(H)CH2CH2(Ph)-
PH- •(Li+) and OH-; none of these species could have been
present in the initial cleavage equilibrium, and subse-
quent oxidation of the phosphinyl radical would lead
directly to the desired product 1 without reestablishment
of the equilibrium. Thus, the reaction of 7 with water
releases PhP(Li)CH2CH2(Ph)P•, which reestablishes the
cleavage equilibrium before final protonation by water
to give secondary phosphines 1 and/or 5. Accordingly, the
phosphinyl radical must also have participated in the
original cleavage reaction and equilibration of P-Caryl
and P-Calkyl products.
Hyd r olysis of 4 in th e Absen ce of P h P (Li)-
CH2CH2(P h )P 2- •(Li+)2. The influence of the species
absorbing at 421 nm in the hydrolysis reaction was
investigated by observing the reaction of water with a
solution of 4‚4THF in THF. When 4‚4THF was dissolved
in THF, the solution exhibited an absorption at 366 nm.
This solution produced pure 1 when treated with water,
regardless of the reaction temperature (0, 20, 40 °C). The
addition of 2-4 equiv of PhLi to the solution immediately
prior to hydrolysis did not affect the outcome (the
concentrations of the PhLi-containing solutions employed
in these experiments were determined prior to use).17
From these observations it was concluded that the
species displaying the absorption at 421 nm was respon-
sible for the rearrangement leading to 5 when solutions
of 4, generated in situ, were hydrolyzed at >0 °C.
Clea va ge of 2‚2BH3. The adduct 2‚2BH3 was pre-
pared by heating 2 under reflux with 4 equiv of BH3.SMe2
in THF.18 The adduct was isolated as an air-stable,
microcrystalline solid by evaporation of the solvent and
chromatography of the residue on silica with dichlo-
romethane as eluant. Characterizations of the product
1
by 31P{1H} NMR (CDCl3, δ 19.0) and H NMR (CDCl3, δ
7.8-7.3 (m, 20 H, ArH), 2.38 (d, 4 H, CH), 0.72 (br s, 6H,
BH)) spectroscopy and electrospray mass spectrometry
were consistent with the formulation 2‚2BH3.
No reaction was observed when a mixture of 2‚2BH3
in boiling THF (0.2 M) and lithium (10-fold excess) was
stirred for 1 h, followed by a further 12 h at room
temperature. Thus, 2‚2BH3 is considerably less reactive
toward lithium than 2.
In flu en ce of Im p u r ities. To investigate the effect of
sodium impurities in the lithium employed (99.9% lithium
wire containing 0.01% sodium), experiments were per-
formed using lithium pieces (98% Li containing 0.5-1.0%
sodium) and lithium dispersion in mineral oil (99.95%
Li). No notable changes in product yields or induction
times were observed.
Su m m a r y. The cleavage of 2 by lithium in THF is
dependent on temperature during the initial P-C bond
cleavage and the later protonation step. An equilibrium
appears to exist between the products of P-Caryl cleavage
and those of P-Calkyl cleavage in these steps (Scheme 2);
pathway I is relatively favored at 40 °C, and pathway II
predominates at 0 °C. Thus, by performing the cleavage
at 0 °C, P-Calkyl cleavage is minimized and the yield of
1 increased.
It has been suggested7 that small amounts of short-
lived Ph2- •(Li+)2 may be generated from PhLi and lithium
in tertiary phosphine-lithium mixtures; in this case the
Ph• radicals produced during hydrolysis would reestablish
the equilibrium. The alternative products, PhLi (one-
electron reduction) or (C6H6)- • Li+ (protonation), cannot
reestablish the equilibrium.
This rationalization supports the assertion7 that a
number of steps are involved in the cleavage of tertiary
phosphines by lithium. In particular, it is suggested that
the initial reaction in the cleavage of 2 involves the
phosphinyl radical PhP(Li)CH2CH2(Ph)P• in addition to,
or instead of, Ph•.
The products of the cleavage, lithium diphenylphos-
phide (3) and the bis(phosphide) 4, are stable to heat,
however. The major side-product of the reaction, PhLi,
Mech a n ism . On the basis of these observations, a
mechanism can be proposed for the lithium-induced
cleavage of 2 in THF. The cleavage step comprises the
(17) Furniss, B. S.; Hannaford, A. J .; Smith, P. W. G.; Tatchell, A.
R. In Vogel’s Textbook of Practical Organic Chemistry, 5th ed.;
Longman Scientific and Technical, New York, 1989; pp 1170-1171.
(18) Imamoto, T.; Oshiki, T.; Onozawa, T.; Kusumoto, T.; Sato, K.
J . Am. Chem. Soc. 1990, 112, 5244.
(19) West, P.; Waack, R. J . Am. Chem. Soc. 1967, 89, 4395.
(20) For information on reductive electron transfer and proton
abstraction by radical anions, see: Holy, N. L. Chem. Rev. 1974, 74,
243 and references therein.

Mechanism of Phosphorus-Carbon Bond Cleavage
J . Org. Chem., Vol. 65, No. 4, 2000 955
Sch em e 3
establishment of an equilibrium involving phosphinyl and
phenyl/alkyl radicals and anions, followed by the ir-
reversible formation of lithium phosphides. Scheme 3
summarizes the process, with the equilibria in the center
of the scheme representing the processes occurring at
each P atom.
When the lithium cleavage is performed at 0 °C,
pathways IIa and IIb are possible for each P center. Our
data indicates IIa as the pathway for the reestablishment
of the cleavage equilibrium during hydrolysis.
way IIa and/or IIb. Following the irreversible formation
of the bis(phosphide) 4 and/or mono(phosphide) 3, it is
clear that these species undergo further reaction with
lithium to form the radical dianions 7 and 6, respectively
(eqs 2 and 3),8,9 with which they are in equilibrium
(Scheme 4). Neither process can alter the proportions of
3 and 4 already present in solution, nor can the presence
of Ph2- •(Li+)2 or Ph•. These data are therefore consistent
with the observation that the proportions of the phos-
phides 3 and 4 formed in the initial reaction are unaf-
fected by subsequent changes in temperature.
The addition of water to the reaction mixture results
in the formation of Ph2P•, 8, and/or PhP(Li)CH2CH2PPh•,
9, as shown in eqs 4 and 5, and perhaps Ph• from the
hypothetical Ph2- •(Li+)2.20 These species reestablish the
cleavage equilibrium, which is shown in the central
bracket of Scheme 3. Although the equilibrium applies
to a small proportion of the phosphines in solution, its
products are unstable and highly reactive. In the ab-
sence of lithium with which to form the more stable
lithium phosphides, the reestablishment of the cleavage
equilibrium occurs by a series of rapid intermolecular
exchange reactions (eqs 6 and 7). The proportions of 3
and 4 in the solution readjust according to the popula-
tions of the species included in the cleavage equilibrium
at the temperature at which the initial water is added.
Further addition of water results in secondary phos-
phines (eqs 8-11).
When the reaction is performed at more elevated
temperature, pathways Ia and Ib are also possible for
each P center. Each of these options produces different
byproducts. If both P centers form phosphinyl radicals
(pathway Ia) in the cleavage, the byproduct LiCH2CH2-
Li will abstract 2H+ from solvent molecules to give CH3-
CH3 (as observed in the cleavage of Ph2PCH2PPh2 by
lithium7). If one P center forms a radical (pathway Ia)
and the other a phosphide ion (pathway Ib) during the
cleavage, LiCH2CH2• will be produced. This species may
form CH3CH2CH2CH3 by dimerization and reaction with
solvent molecules. If both P centers form phosphide ions
•
•
(pathway Ib), the CH2CH2 generated will be liberated
as ethylene. Ethylene has, in fact, been observed as a
byproduct in the reaction, so Ib appears to be an
important pathway.7 However, because the cleavage
involves a thermodynamic equilibria, pathway Ia must
also be possible.7
When lithium reacts with 2 in THF at 20 °C, path-
way Ib (and possibly Ia) therefore competes with path-
A mechanism involving competitive release of a leaving
group as the thermodynamically most-stable radical or

956 J . Org. Chem., Vol. 65, No. 4, 2000
Dogan et al.
Sch em e 4
anion also explains the incidence of P-Calkyl vs P-Caryl
cleavage in the reactions of Ph2P(CH2)nPPh2 (n ) 1-5)
with lithium.7 The leaving group can then separate as
the alkyl radical when n ) 1 but as the phenyl anion
when n ) 3-5. The nature of the bond cleavage therefore
does not depend on leaving group preferences as radicals
OR anions but rather on the relative thermodynamic
stabilities of the radical and anionic species involved.
Con clu sion s
The cleavage of 1,2-bis(diphenylphosphino)ethane by
lithium in THF involves a rapid equilibrium between
P-Caryl and P-Calkyl cleaved anions and radicals, followed
by their reaction with lithium and stabilization as lithium
salts. The temperature dependence of the cleavage
indicates that the equilibrium is thermodynamic in
nature. The addition of water causes a reestablishment

Mechanism of Phosphorus-Carbon Bond Cleavage
J . Org. Chem., Vol. 65, No. 4, 2000 957
THF (360 mL) was cooled to 0 °C and vigorously stirred as a
solution of 1,2-bis(diphenylphosphino)ethane (80 g; 0.2 mol)
in THF (480 mL) was added over 4 h. After completion of the
addition, the mixture contained crystals of 4‚4THF. The
reaction mixture was warmed to room temperature and heated
under reflux for 2 h. The hot solution was separated from
excess lithium with use of a cannula and cooled to 0 °C over 1
h. The mixture was then stirred vigorously as aqueous THF
(20% water, 50 mL) was added over 1 h. The brown reaction
mixture was decolorized by the aqueous THF, and lithium
hydroxide was deposited. The THF was removed in vacuo, and
water (400 mL) and diethyl ether (200 mL) were added to
the residue. Three extractions of the aqueous phase with
similar quantities of diethyl ether followed. The combined
extracts were dried (MgSO4), filtered, and evaporated in vacuo
to afford the product as a pale yellow, air-sensitive oil that
of the cleavage equilibrium prior to formation of the
secondary phosphines with further water. The data col-
lected in this investigation suggest a mechanism involv-
ing competitive separation of the leaving groups as the
thermodynamically most-stable radical or anion. This
mechanism is consistent with previous investigations
involving homologous bis(tertiary phosphines).
Exp er im en ta l Section
All reactions and manipulations were performed under an
inert atmosphere using the Schlenk technique. THF was
distilled off sodium/benzophenone ketyl. 31P{1H} NMR spectra
were recorded at 20 °C on a Varian VXR 300S spectrometer
operating at 299.96 MHz and referenced to external aqueous
H3PO4 (85%). UV-visible spectra were obtained using a Cary
5E spectrometer and ESR spectra on a Bruker ER 200 D-SRC
spectrometer. 1,2-Bis(diphenylphosphino)ethane, lithium wire
(99.9%; sodium content ca. 0.01%), PhLi (1.8 M in diethyl
ether/cyclohexane), and BH3‚SMe2 were purchased from Ald-
rich. Lithium diphenylphosphide and diphenylphosphine were
prepared by literature methods,5 as were PhP(Li)(CH2)2P(Li)-
Ph and PhP(H)(CH2)2P(H)Ph.13
1
was pure according to 31P{1H} and H NMR spectroscopy. The
product distilled as a colorless oil having bp 140 °C (0.07
mmHg) [lit.13 140 °C; 0.07 mmHg]; 40 g (81% yield). 31P{1H}
1
NMR (C6D6): δ -46.43, -46.59 (R*, R*-(()/R*, S*). H NMR
(C6D6): δ 1.80-2.10 (br m, 4 H, CH2), 3.94 (br m, 1 H, PH),
4.62 (br m, 1 H, PH), 7.27 (br m, 6 H, ArH), 7.48 (br m, 4 H,
ArH).
Ack n ow led gm en t. G.F.S., J .D., and J .B.S. thank
the Australian Research Council for support of this
work.
(R*, R*)-(()/(R*, S*)-1,2-Bis(p h en ylp h osp h in o)eth a n e
(1). A mixture of finely divided lithium wire (13.8 g; 2 mol) in
J O9907336