Radical synthesis of trialkyl, triaryl, trisilyl and tristannyl phosphines from P4

Article abstract of DOI:10.1039/c0nj00124d

A reaction scheme has been devised according to 3 RX + 3 Ti(iii) + 0.25 P₄ → PR₃ + 3 XTi(iv), wherein RX = PhBr, CyBr, Me₃SiI or Ph₃SnCl, with contrasting results in the case of more hindered RX. The scheme accomplishes the direct radical functionalization of white phosphorus without the intermediacy of PCl₃.

Full text of DOI:10.1039/c0nj00124d

New Journal of Chemistry
An international journal of the chemical sciences
www.rsc.org/njc
Volume 34 | Number 8 | August 2010 | Pages 1493–1784
Themed Issue: Main Group chemistry
ISSN 1144-0546
LETTER
Brandi M. Cossairt and
Christopher C. Cummins
Radical synthesis of trialkyl,
triaryl, trisilyl and tristannyl
phosphines from P4
1144-0546(2010)34:8;1-K

LETTER
www.rsc.org/njc | New Journal of Chemistry
View Article Online
Radical synthesis of trialkyl, triaryl, trisilyl and tristannyl phosphines
from P4wz
Brandi M. Cossairt and Christopher C. Cummins*
Received (in Montpellier, France) 15th February 2010, Accepted 16th March 2010
DOI: 10.1039/c0nj00124d
A reaction scheme has been devised according to 3 RX + 3
Ti(III) + 0.25 P4 - PR3 + 3 XTi(IV), wherein RX = PhBr,
CyBr, Me3SiI or Ph3SnCl, with contrasting results in the case of
more hindered RX. The scheme accomplishes the direct radical
ð1Þ
functionalization of white phosphorus without the intermediacy
of PCl3.
In the course of a previous study of the radical cleavage of
symmetrical 1,4-dicarbonyl compounds by Ti(N[tBu]Ar)3, the
propensity was examined of Ti(N[tBu]Ar)3 to abstract Xꢀ from
halobenzenes.5 This study revealed that the treatment of
Ti(N[tBu]Ar)3 with a stoichiometric amount of PhBr or PhI
effected the conversion to XTi(N[tBu]Ar)3, X-1, rapidly at
room temperature, while the conversion to ClTi(N[tBu]Ar)3
upon treatment with PhCl was considerably slower. Dissolution
of Ti(N[tBu]Ar)3 in neat chlorobenzene and with stirring
overnight at room temperature did effect the complete
conversion to ClTi(N[tBu]Ar)3, however. A radical cyclization
experiment using o-bromophenyl allyl ether as the RX
substrate for Ti(N[tBu]Ar)3 was used to substantiate the
hypothesis that phenyl radicals are indeed generated upon
It is known that P4, white phosphorus, has excellent properties
as a trap for carbon-centered radicals in solution and under
the mild conditions that are typical of organic synthesis. The
most prominent example of this was the demonstration that
phosphonic acids may be prepared from their corresponding
carboxylic acids by way of O-acyl derivatives of N-hydroxy-
2-thiopyridone (Barton PTOC esters).1 The latter provide
carbon-centered radicals in an oxygen-initiated chain reaction,
and these are consumed upon combination with P4 as the
critical P–C bond-forming event; upon oxidative work-up, any
remaining P–P bonds are cleaved and the phosphonic acid,
RP(O)(OH)2, is the end product.2 It is also known that P–P
bonds other than those in P4 may serve as traps for organic
radicals. This has been shown by Sato et al. in a scheme for the
radical phosphination of organic halides, wherein ArX serves
as a source of Arꢀ, which in turn attacks Ph2P–PPh2, yielding
ArPPh2.3
Such a vision for phosphine synthesis via homolytic
substitution at a phosphorus center has also been developed
by Vaillard et al., who employed Me3MPPh2 (M = Si or Sn)
as the phosphorus substrate and RX as the carbon radical
source, together with a radical initiator, to produce RP(O)Ph2
efficiently after an oxidative work-up.4 For our part, we have
previously shown that the three-coordinate Ti(III) complex
Ti(N[tBu]Ar)3 (Ar = 3,5-C6H3Me2) (1) is a potent halogen-
atom abstractor, capable of abstracting Xꢀ (X = Cl, Br, or I)
from various donor molecules at room temperature or below
in aprotic organic media (Fig. 1). In the present work, we
sought to develop a high yielding synthesis of phosphines PR3
from 3 RX and 0.25 P4 using Ti(N[tBu]Ar)3 as a halogen atom
sink (see idealized eqn (1)). Success in this arena would
demonstrate that it is possible to synthesize valuable tertiary
phosphanes, PR3, through the direct functionalization and
complete consumption of P4 by a radical mechanism.
Department of Chemistry, Massachusetts Institute of Technology,
Cambridge, MA, USA. E-mail: ccummins@mit.edu
w We dedicate this work to the memory of Sir Derek Barton. This
article is part of a themed issue on Main Group chemistry.
z Electronic supplementary information (ESI) available: Additional
synthetic and characterization details. See DOI: 10.1039/c0nj00124d
Fig. 1 Ti(N[tBu]Ar)3 together with various P4-derived phosphanes
and polyphosphorus products. R = tBu; X = Cl, Br, I; R0 = Ph, Mes,
Cy, Ph3Sn; Mes = 2,4,6-Me3C6H2; Dmp = 2,6-Mes2C6H3.
ꢁc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010 New J. Chem., 2010, 34, 1533–1536 | 1533

View Article Online
halogen atom abstraction from PhX by Ti(N[tBu]Ar)3.5,6 On
the basis of this information, together with the knowledge
from recent independent work that Ti(N[tBu]Ar)3 engages in a
negligible reaction with P4,7 we realized that Ti(N[tBu]Ar)3 is
an unusual reducing agent, in that it could be selective for RX
activation in the presence of P4. This is unusual because most
chemical reducing agents capable of Xꢀ abstraction from RX
would not be expected to be selective for this reactivity channel
in the presence of P4. An aspect of this type of special
selectivity in reactions of Ti(N[tBu]Ar)3 has been demonstrated
previously, wherein 7-chloronorbornadiene was treated with a
1 : 1 mixture of Ti(N[tBu]Ar)3 and Mo(N[tBu]Ar)3; in this
instance, Ti(N[tBu]Ar)3 was entirely selective for Cl atom
abstraction, giving ClTi(N[tBu]Ar)3, while exhibiting no
propensity for trapping the 7-norbornadienyl radical, which
was seen to interact selectively with the molybdenum
complex.8 In addition, typical one-electron reducing agents
that might be used for effecting Xꢀ abstraction, e.g.
CoCl(PPh3)3, SmI2 or Cp2TiCl, simply gave no reaction with
substrates such as PhBr.yz
Table 1 Synthesis of PR3 from n(RX + Ti(N[tBu]Ar)3) and 0.25P4 in
benzene solvent at 20 1C
Entry
na
R
X
d
PR /ppmb
Yield (%)c
3
1
2
3
4
5
6
7
8
9
10
3
3.75
5
3
3
3
3
3
3.75
5
Ph
Ph
Ph
Ph
Br
Br
Br
I
Cl
Cl
I
Br
Br
Br
ꢂ4.9
ꢂ4.9
ꢂ4.9
ꢂ4.9
n/a
71
82
95
65
0
96
97
64
77
95
Ph
Ph3Sn
Me3Si
Cy
Cy
Cy
ꢂ325d
ꢂ252
10.5
10.5
10.5
a
Number of equivalents per phosphorus atom. b 31P NMR chemical
shift for the PR3 product referenced to external 85% H3PO4.
c
Phosphorus-based yield of PR3, as determined by 31P NMR spectro-
scopy via integration with respect to an internal standard using a single
d 1
1
= 442 Hz, J117
pulse experiment.
J
= 425 Hz.
119Sn–P
Sn–P
from the crude reaction mixture in 86% yield, while highly
crystalline P(SnPh3)3 can be isolated in 75% yield.
In a first reaction targeted at generating PPh3, it was found
that the addition of 3 equiv. of PhBr by microsyringe to a
0.04 M solution of 0.25 equiv. P4 containing 3 equiv. of 1 in
benzene resulted in the immediate formation of a bright
orange solution containing BrTi(N[tBu]Ar)3 (Br-1), PPh3
(Ph-2, 71% of the P-containing product) and P2Ph4 (Ph-3,
29% of the P-containing product, Table 1). P2Ph4 is one of the
four possible stable intermediates en route to complete P4
degredation by P4 to give PPh3, and is present in this stoichio-
metric treatment because the trapping of the highly reactive
phenyl radicals is not completely efficient in this system.z In
order to convert the full equivalent of P4 to PPh3, 5 equiv. of
PhBr and Ti(N[tBu]Ar)3 were used, giving 95% conversion
and an isolated yield of 72% (Table 1). We could also
selectively target P2Ph4 by the treatment of 0.25 equiv. of
0.04 M P4 in benzene with 2 equiv. of 1, followed by 2 equiv.
of PhBr, which gave P2Ph4 in 80% yield, with small amounts
of PPh3 and P4Ph4 being observed as well. Evidence for the
intermediacy of P2Ph4 along the reaction pathway was
provided by the use of P2Ph4 itself as a starting material for
PPh3 synthesis.z It was found that PhI could be used in place
of PhBr with similar results, however PhCl did not lead to any
PPh3 or P2Ph4 formation, as Ti(N[tBu]Ar)3 reacts very slowly
with PhCl under these conditions.5
The ability of P4 to act as a radical trap in combination with
the work of Sato and co-workers on the radical phosphination
of aryl halides suggests that P–P bonds, generally, may be
competent radical traps.3 This was found to be the case using
our radical method, opening up the potential for the synthesis
of asymmetric phosphines. The treatment of 0.5 equiv. of
P2Ph4 with 1 equiv. of PhBr, MesBr, CyBr or Ph3SnCl and
1 equiv. of 1 quantitatively produced 1 equiv. of Ph-2
(d 4.9 ppm), P(Ph2)Mes (Mes-4, d 16.0 ppm),11 P(Ph2)Cy
(Cy-4, d 3.4 ppm)12 or P(Ph2)SnPh3 (Ph3Sn-4, d 56.2 ppm,
1J119 = 715 Hz, 1J117 = 682 Hz),13 respectively (Fig. 1).z
Sn–P
Sn–P
This striking attribute of P–P single bond chemistry has great
potential for further synthetic development.
Based on our hypothesis that the radical degradation of the
P4 tetrahedron occurs in a stepwise manner, we thought that it
might be possible to target intermediate structures by tuning
the steric properties of the RX substrate. It was found that the
treatment of 0.25 equiv. of 0.04 M P4 in benzene with
1.5 equiv. 1, followed by 1.5 equiv. of MesBr, gave P3Mes3
(5) as the major product and small amounts of P2Mes4
(Mes-3).14–16 P3Mes3 could be isolated from the reaction
mixture in 61% yield. Increasing the steric pressure further,
we found that the treatment of 0.25 equiv. of 0.04 M P4 with
1.5 equiv. of 1 and 1.5 equiv. of DmpI (Dmp = 2,6-
Mes2C6H3) gave cis,trans-DmpP4Dmp (6) as the exclusive
product, isolated in 78% yield.17z This latter reaction represents
a facile approach for the synthesis of novel substituted tetra-
phosphabicyclobutane molecules directly from P4 in a single
step. Many of the previously reported syntheses of stable
tetraphosphabicyclobutanes involve the coupling of two
substituted diphosphanes18 or the activation of P4 by some
highly designed substrate.17,19–21 Our synthesis is unique in
that a large number of sterically-hindered aryl or alkyl halides
could be employed in a general synthesis.
This synthesis of phosphines from P4 and a burst of radicals
was found not to be limited to aryl substituents. The treatment
of a 0.04 M solution of 0.25 equiv. P4 with 5 equiv. of 1 and
5 equiv. of CyBr resulted in formation of PCy3 (Cy-2) as the
exclusive P-containing product (Table 1). The use of less than
5 equiv. of CyBr resulted in mixtures of P2Cy4 (Cy-3) and
Cy-2, much like what was seen for PhBr. When the radicals
produced were longer lived, it was possible to obtain stoichio-
metric conversion of P4 to the trisubstituted phosphine. For
instance, treatment of a 0.04 M solution of 0.25 equiv. P4 with
3 equiv. of 1 and 3 equiv. of Me3SiI or Ph3SnCl resulted in the
clean and quantitative formation of known phosphines
P(SiMe3)3 or P(SnPh3)3, respectively, as the sole products
(Table 1, Fig. 1).9,10 The P(SiMe3)3 produced here was easily
separated from the reaction co-products by vacuum transfer
In terms of recycling the titanium by-products from these
syntheses, it is worth noting that X-1 (X = I, Br, Cl) are
cleanly reduced back to Ti(N[tBu]Ar)3 by reduction with
Na/Hg amalgam.22,23 This ability to easily recycle the titanium
by-products generates a closed cycle for the synthesis of
ꢁc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010
1534 | New J. Chem., 2010, 34, 1533–1536

View Article Online
trisubstituted phosphines from P4. One might begin to con-
template a catalytic cycle using this system; however, the
reduction of XTi(N[tBu]Ar)3 is slow and P4 is itself susceptible
to reduction to Na3P by the Na/Hg amalgam under such
conditions. As such, other halogen atom abstractors are
currently being screened as potential entry points into the
catalytic generation of trisubstituted phosphines from P4 by
this radical trapping method.z
Ti(N[tBu]Ar)3 and 5 equiv. (126 mg, 0.807 mmol) of BrC6H5.
Again, over the course of a minute, the originally green
reaction mixture took on a bright orange color. The reaction
mixture was analyzed by 1H, 13C and 31P NMR spectroscopy.
Using OPPh3 (26 ppm) as an internal standard, a single pulse
31P NMR experiment showed 98% conversion to PPh3
(s, ꢂ4.9 ppm). GC-MS analysis confirmed this assignment.
A screening of the reaction stoichiometry showed 5 equiv. of
Ti(N[tBu]Ar)3 and 5 equiv. BrC6H5 were necessary for the
complete conversion of P4 to PPh3; when fewer equivalents
were used, small amounts of P2Ph4 were observed. When the
optimized conditions were scaled up 10-fold, PPh3 was
isolated by repeated crystallizations at ꢂ35 1C in Et2O in
72% yield (304 mg).
The present day synthesis of many organophosphorus
compounds is a multi-step process in which P4 is first
chlorinated to generate PCl3.24 PCl3 is then functionalized
via salt elimination reactions with appropriate Grignard or
organolithium reagents, or with an organohalide and a harsh
reducing agent.24 For example, the industrial method for PPh3
preparation is based on the high temperature reaction of
chlorobenzene with PCl3 in the presence of molten sodium.25
Manufacturers of organophosphorus compounds have
recognized that the direct functionalization of white phos-
phorus is one of the major challenges in this field.25,26 New
studies are needed to work out alternative direct routes to
organophosphorus compounds that avoid the chlorination of
white phosphorus. Strides have been made with regard to the
electrosynthesis of trisubstituted phosphines directly from
P4,27 but facile solution methods are lacking. It is our hope
that this work will inspire renewed interest in the use of P–P
bonds as efficient radical traps and will eventually lead to a
robust catalytic system for the synthesis of organophosphorus
compounds directly from white phosphorus. Meanwhile, the
syntheses reported herein represent novel methodologies for
the direct functionalization of P4 and will themselves be the
subject of further investigation.
These optimized conditions of 0.04 M P4 (0.25 equiv.),
benzene and 5 equiv. of RX/Ti(N[tBu]Ar)3 were effective for
both PPh3 and PCy3 syntheses. For P(SiMe3)3 and P(SnPh3)3,
the same conditions were used, but with only 3 equiv.
(stoichiometric) of RX/Ti(N[tBu]Ar)3. Starting with 50 mg
of P4, P(SiMe3)3 was isolated by vacuum transfer in 86% yield
(348 mg) and P(SnPh3)3 was isolated in 75% yield (1.30 g) by
repeated recrystallization from Et2O. For the synthesis of
P3Mes3 and cis,trans-DmpP4Dmp, the same conditions were
used, but with only 1.5 equiv. of RX/Ti(N[tBu]Ar)3. P3Mes3
was isolated by repeated crystallization from Et2O in 61%
yield starting from 50 mg of P4. cis,trans-DmpP4Dmp was
isolated by repeated crystallization from Et2O in 78% yield
starting from 50 mg of P4.
In order to use P2Ph4 as the starting material for PPh3
synthesis, the same reaction protocol and conditions could be
used. The treatment of a 0.04 M solution of P2Ph4 (5 mg,
0.014 mmol, 0.5 equiv.) with Ti(N[tBu]Ar)3 (93 mg, 0.16 mmol,
1 equiv.) followed by BrPh (60 mg, 0.16 mmol, 1 equiv.)
resulted in a rapid color change from green to orange upon
stirring. The reaction mixture was analyzed by 1H, 13C and
31P NMR spectroscopy. Using OPPh3 (26 ppm) as an internal
standard, a single pulse 31P NMR experiment showed 97%
conversion to PPh3 (s, ꢂ4.9 ppm). Similar results were found
when 0.5 equiv. P2Ph4 was treated with 1 equiv. of MesBr,
CyBr or Ph3SnCl, which produced 1 equiv. of P(Ph2)Mes
(ꢂ16.0 ppm), P(Ph2)Cy (ꢂ3.4 ppm) or P(Ph2)SnPh3
(ꢂ56.2 ppm, 1J119 = 715 Hz, 1J117 = 682 Hz), respectively,
We gratefully acknowledge the US National Science
Foundation (grant CHE-719157) and Thermphos International
for support.
Experimental
A representative protocol for the reaction between
Ti(N[tBu]Ar)3, RX (RX = PhBr, MesBr, DmpI, CyBr, Me3SiI
and Ph3SnCl) and P4: the synthesis of PPh3
Ti(N[tBu]Ar)3 (279 mg, 0.484 mmol) was added to a 0.04 M
solution of P4 in benzene (5 mg total P4, 0.040 mmol). BrC6H5
(76 mg, 0.484 mmol) was then added to the reaction mixture at
room temperature by a microlitre syringe. Over the course of a
minute, the originally green reaction mixture took on a bright
Sn–P
Sn–P
each in greater than 95% yield.
Notes and references
1
orange color. The reaction mixture was analyzed by H, 13C
y It is well documented that SmI2 is capable of slowly reducing aryl
halides in the presence of HMPA, however the slow rate of this
transformation did not allow for a radical synthesis of trisubstituted
phosphines.
and 31P NMR spectroscopy. Using OPPh3 (26 ppm) as an
internal standard, a single pulse 31P NMR experiment showed
71% conversion to PPh3 (s, ꢂ4.9 ppm) with the balance being
made up of P2Ph4 (ꢂ14 ppm). GC-MS analysis confirmed this
assignment. Solvent screening (benzene, toluene, THF, Et2O,
n-hexane) and concentration screening (0.01, 0.02, 0.03, 0.04
and 0.05 M P4) indicated that these conditions were optimal
for the conversion of 0.25 equiv. P4 to 1 equiv. PPh3 using
3 equiv. Ti(N[tBu]Ar)3 and 3 equiv. PhBr.
z The fate of the untrapped radicals is unknown.
1 D. H. R. Barton and J. Zhu, J. Am. Chem. Soc., 1993, 115,
2071–2072.
2 D. H. R. Barton and R. A. V. Embse, Tetrahedron, 1998, 54,
12475–12496.
3 A. Sato, H. Yorimitsu and K. Oshima, J. Am. Chem. Soc., 2006,
128, 4240–4241.
4 S. Vaillard, C. Muck-Lichtenfeld, S. Grimme and A. Studer,
¨
Angew. Chem., Int. Ed., 2007, 46, 6533–6536.
5 T. Agapie, P. L. Diaconescu, D. J. Mindiola and C. C. Cummins,
Organometallics, 2002, 21, 1329–1340.
In order to convert all of the P4 to PPh3, the reaction was
repeated using a 0.04 M solution of P4 (5 mg total P4,
0.040 mmol, 0.25 equiv.), 5 equiv. (465 mg, 0.807 mmol) of
ꢁc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010 New J. Chem., 2010, 34, 1533–1536 | 1535

View Article Online
6 D. P. Curran and M. J. Totleben, J. Am. Chem. Soc., 1992, 114,
6050–6058.
18 E. Niecke, R. Rger and B. Krebs, Angew. Chem., Int. Ed. Engl.,
1982, 21, 544–545.
7 B. M. Cossairt and C. C. Cummins, J. Am. Chem. Soc., 2009, 131,
15501–15511.
19 M. B. Power and A. R. Barron, Angew. Chem., Int. Ed. Engl., 1991,
30, 1353–1354.
8 T. Agapie, P. L. Diaconescu and C. C. Cummins, J. Am. Chem.
Soc., 2002, 124, 2412–2413.
20 J. P. Bezombes, P. B. Hitchcock, M. F. Lappert and J. E. Nycz,
Dalton Trans., 2004, 499–501.
9 H. H. Karsch, F. Bienlein, T. Rupprich, F. Uhlig, E. Herrmann
and M. Scheer, in Phosphorus, Antimony, Arsenic, and Bismuth,
ed. H. H. Karsch, Verlag, 1996, pp. 58–64.
21 O. J. Scherer, T. Hilt and G. Wolmershauser, Organometallics,
1998, 17, 4110–4112.
22 J. C. Peters, PhD thesis, Massachusetts Institute of Technology,
1998.
23 J. C. Peters, A. R. Johnson, A. L. Odom, P. W. Wanandi,
W. M. Davis and C. C. Cummins, J. Am. Chem. Soc., 1996, 118,
10175–10188.
24 J. Emsley, The 13th Element: The Sordid Tale of Murder,
Fire, and Phosphorus, John Wiley & Sons, Inc., New York,
2000.
25 R. Engel, Synthesis of Carbon Phosphorus Bonds, CRC Press,
Boca Raton, 2nd edn, 2004.
26 J. Emsley and D. Hall, The Chemistry of Phosphorus:
Environmental, Organic, Inorganic, Biochemical, and Spectroscopic
Aspects, Harper and Row, London, 1976.
¨
10 V. G. Engelhardt, Z. Anorg. Allg. Chem., 1972, 387, 52–60.
11 J. F. Blount, D. Camp, R. D. Hart, P. C. Healy, B. W. Skelton and
A. H. White, Aust. J. Chem., 1994, 47, 1631–1639.
12 S. O. Grim and W. McFarlan, Nature, 1965, 208, 995–996.
13 G. Engelhardt, P. Reich and H. Schumann, Z. Naturforsch., B: Anorg.
Chem., Org. Chem., Biochem., Biophys., Biol., 1967, 22, 352–353.
14 C. Frenzel and E. Hey-Hawkins, Phosphorus, Sulfur Silicon Relat.
Elem., 1998, 143, 1–17.
15 X. Li, D. Lei, M. Y. Chiang and P. P. Gaspar, J. Am. Chem. Soc.,
1992, 114, 8526–8531.
16 X. Li, S. I. Weissman, T.-S. Lin, P. P. Gaspar, A. H. Cowley and
A. I. Smirnov, J. Am. Chem. Soc., 1994, 116, 7899–7900.
17 A. R. Fox, R. J. Wight, E. Rivard and P. P. Power, Angew. Chem.,
Int. Ed., 2005, 44, 7729–7733.
27 Y. H. Budnikova, D. G. Yakhvarov and O. G. Sinyashin,
J. Organomet. Chem., 2005, 690, 2416–2425.
ꢁc
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2010
1536 | New J. Chem., 2010, 34, 1533–1536