Homolytic substitution at phosphorus for the synthesis of alkyl and aryl phosphanes

Article abstract of DOI:10.1002/anie.200701650

(Chemical Equation Presented) A transition-metal-free radical phosphonation using Me₃SnPPh₂ and the less toxic Me ₃SiPPh₂ is reported. These readily available reagents react highly efficiently with primary and secondary alkyl radicals. Moreover, aryl radicals and tertiary alkyl radicals are phosphonated with Me ₃SnPPh₂ (see scheme; R = aryl, alkyl, vinyl; X = 1, Br, OC(S)imidazolyl). DFT calculations provide insights into the mechanism of the reaction.

Full text of DOI:10.1002/anie.200701650

Angewandte
Chemie
DOI: 10.1002/anie.200701650
Synthetic Methods
Homolytic Substitution at Phosphorus for the Synthesis of Alkyl and
Aryl Phosphanes**
Santiago E. Vaillard, Christian Mück-Lichtenfeld, Stefan Grimme,* and Armido Studer*
Table 1: Phosphonation of PhX with 1a under various conditions.[a]
During the past thirty years phosphanes and their derivatives
[b]
have received much attention in many fields of chemistry. In
particular in homogeneous catalysis they are highly important
as ligands in various metal complexes. Although different
approaches for the synthesis of phosphanes have been
reported, the development of new methods for their prepa-
ration is still important owing to the lack of generality of the
known methods, the harsh reaction conditions often applied,
and the necessity for expensive metal catalysts.[1] In this
context, radical-based methods appear to be a promising
alternative to polar and transition-metal-mediated reac-
tions.[2]
Entry
X
1a
[equiv]
Initiator
t [h]Yield [%]
1
2
3
4
5
6
I
I
I
I
Br
Cl
1.7
1.7
1.7
4.0
4.0
4.0
AIBN
V-40
V-40
V-40
V-40
V-40
2.5
2.5
41
61
66
87
54
<2
16
16
16
16
[a]All reactions were performed in benzene at 80 8C at 0.083m with
0.15 mol% initiator. [b]Determined by GC analysis.
The reactivity of free radicals towards tricoordinated
phosphorus (PR3) has been studied.[3] With trialkylphosphites
as radical-trapping reagents, efficient P transfer to the radical
only occurs to reactive alkoxyl or s-type vinylic or aryl
radicals.[4] Another paper discusses the radical addition of
(0.083m). For analysis, the product phosphanes were oxidized
with H2O2 to triarylphosphane oxide [Eq. (1)].
ꢀ
stannylated phosphanes to C C triple bonds. The reactive
vinyl radical intermediates undergo homolytic substitution at
phosphorus.[5] Very recently, Oshima et al. published the
reaction of aryl radicals with diphosphane (Ph2P-PPh2)
generated in situ for the synthesis of aryl phosphane
derivatives.[6] Importantly, Ph2P-PPh2 was reported to react
with the less reactive p-type alkyl radicals.[6] Encouraged by
these results we initiated a program to study homolytic
We were delighted to see that reaction of iodobenzene
with reagent 1a (1.7 equiv) and a,a’-azobisisobutyronitrile
(AIBN) delivered after oxidation Ph3PO in 41% yield
(Table 1, entry 1). The yield could be improved to 61% by
switching to V-40 (1,1’-azobis(cyclohexane)-1-carbonitrile) as
initiator (Table 1, entry 2). Extension of the reaction time to
16 h gave Ph3PO in 66% yield (Table 1, entry 3). A further
improvement (87% yield) was achieved upon using a fourfold
excess of 1a (Table 1, entry 4). A lower yield was obtained
with bromobenzene (Table 1, entry 5), while only traces of
Ph3PO were formed with chlorobenzene (entry 6).
substitution[7] at P for C P bond formation. Herein we wish to
ꢀ
report our first results on the mild radical phosphonation with
s- as well as p-type C radicals using stannylated and silylated
phosphanes as radical acceptors. Importantly, transition
metals are not necessary for these reactions. Our results are
further supported by DFT calculations of the reaction
mechanism.
The reaction of the readily prepared Me3SnPPh2 (1a)[8,9]
with phenyl halides was studied first (Table 1). The phospho-
nations were conducted in sealed tubes in benzene at 808C
We next studied the scope and limitations of the radical
ꢀ
P C bond formation. Electron-withdrawing (see 2b–d, 2 f)
and also electron-releasing groups (see 2a,e) in the o and p
position are tolerated in the reaction with reactive aryl
radicals. Good yields of the corresponding triarylphosphane
oxides 3a–f were obtained (59–79%, Table 2, entries 1–6).
Vinyl radicals efficiently react with Me3SnPPh2 as shown for
the transformation of 2-bromopropene (Table 2, entry 7, !
3g, 76%). We then focused on less reactive p-type radicals.
Pleasingly, the primary and secondary alkyl iodides 2h,i, 1-
bromoundecane, and tert-butyl bromide were readily phos-
phonated in high yields (!3h (79%), 3i (94%), 3j (54%), 3k
(83%), Table 2, entries 8, 10, 12, and 14). Other typical radical
precursors such as thionocarbonate 2l and phenyl selenide
2m can be transformed to the phosphane oxides 3i and 3h in
good yields (Table 2, entries 16 and 17). Reducing the amount
of reagent 1a from 4 equiv to 1.5 equiv resulted in slightly
[*]Dr. S. E. Vaillard, Dr. C. Mück-Lichtenfeld, Prof. Dr. S. Grimme,
Prof. Dr. A. Studer
Organisch-Chemisches Institut
Westfälische Wilhelms-Universität
Corrensstrasse 40, 48149 Münster (Germany)
Fax: (+49)281-83-36523
E-mail: grimmes@uni-muenster.de
studer@uni-muenster.de
[**]The Alexander von Humboldt Foundation is acknowledged for
supporting our work (stipend to S.E.V). A.S. thanks Novartis
Pharma AG for financial support (Novartis Young Investigator
Award).
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2007, 46, 6533 –6536
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
6533

Communications
Table 2: Radical phosphonations using reagent 1a.[a]
Given the toxicity of organotin compounds[11] and diffi-
culty in removal of tin by-products from the reaction mixture,
we turned our attention to the silylphosphane 1b, which is
commercially available and can be efficiently prepared from
readily available starting materials.[12,13]
We switched to n-heptane as a solvent, because silyl
radicals are known to react with benzene and would therefore
be prevented from propagating the chain.[14] To our delight,
the reaction of pentyl iodide with Me3SiPPh2 (4 equiv) and V-
40 (0.25 equiv) in heptane (0.1m) at 1008C for 72 h provided
after oxidation 3h in 86% yield (Table 3, entry 1). A slightly
Entry Radical
precursor
R
X
Product
(Yield [%])[b]
1
2
3
4
5
6
7
8
2a
2b
2c
2d
2e
2 f
2g
2h
2h
2i
2i
2j
2j
2k
2k
2l
p-MeO-C6H4
p-NC-C6H4
p-F3C-C6H4
p-Cl-C6H4
o-MeO-C6H4
o-MeO2C-C6H4
I
I
I
I
I
I
Br
I
I
I
I
Br
Br
Br
Br
3a (73)
3b (79)
3c (75)
3d (72)
3e (59)
3 f (73)
3g (76)
3h (79)
3h (64)
3i (94)
3i (82)
3j (54)
3j (33)
3k (83)
3k (64)
=
CH3C CH2
n-pentyl
n-pentyl
9[c]
10
11[c]
12
13[c]
14
15[c]
16
17
cyclohexyl
cyclohexyl
n-undecyl
n-undecyl
tert-butyl
tert-butyl
cyclohexyl
n-pentyl
Table 3: Radical phosphonations using reagent 1b.[a]
Entry
Radical
precursor
R
X
Product
(Yield [%])[b]
1
2
3
4
5
6
7
8
2h
2h
2h
2i
2i
2i
n-pentyl
n-pentyl
n-pentyl
cyclohexyl
cyclohexyl
cyclohexyl
n-undecyl
cyclohexyl
I
I
I
I
I
I
Br
3h (86)
3h (4)[c]
3h (74)[d]
3i (70)
OC(S)imidazolyl 3i (57)
SePh 3h (60)
2m
[a]All reactions were performed in benzene at 80 8C at 0.083m using a
fourfold excess of 1a and 16 mol% of V-40. [b]Yield of isolated product.
[c]Reaction conducted with 1.5 equiv of 1a.
3i (9)[c]
3i (58)[d]
3j (41)
2j
2l
OC(S)imidazolyl
3i (44)
diminished but still good yields, as shown for the reactions
with alkyl iodides and bromides (Table 2, entries 9, 11, 13, and
15).
To prove the radical nature of the process, iodide 4 was
reacted with Me3SnPPh2 to afford the cyclization/phospho-
nation product 5 in 72% yield [Eq. (2)]. Moreover, reaction
[a]All reactions were performed in n-heptane at 1008C at 0.1m using a
fourfold excess of 1b and 25 mol% of V-40 for 72 h. [b]Yield of isolated
product. [c]Reaction was conducted without V-40. [d]Reaction con-
ducted with 1.5 equiv of 1b.
lower yield was obtained for the phosphonation of cyclohexyl
iodide (Table 3, entry 4). The reactions conducted without
initiator afforded the desired compounds in less than 10%
yield supporting a radical mechanism (Table 3, entries 2 and
5). Even with 1.5 equiv of reagent 1b good yields are obtained
(Table 3, entries 3 and 6). Bromides and thionocarbonates are
also acceptable as radical precursors (Table 3, entries 7 and
8). Unfortunately, tert-butyl bromide and iodobenzene were
not phosphonated under optimized conditions.
To gain further insight into the mechanism of the reaction
we conducted density functional theory (DFT) calculations.
We used the PBE functional[15] together with a basis set of
triple zeta quality (TZVP)[16] for geometry optimizations. The
energies were then recalculated with the double-hybrid
density functional B2-PLYP[17] adding additional polarization
functions (TZVPP). An empirical correction accounting for
long-range dispersion interactions was added during all the
calculations (“PBE-D” and “B2-PLYP-D”).[18]
of 5-hexenyl bromide under the same conditions provided the
radical-cyclization product 6b in 85% yield along with
uncyclized phosphane oxide 6a (6%) [Eq. (3)]. From the
ratio of 6a to 6b the rate constant for the trapping of a
primary alkyl radical with Me3SnPPh2 at 808C is calculated to
be 3 ” 105 mꢀ1 sꢀ1 by applying radical-clock methodology.[10] To
our surprise, reaction of 1a (4 equiv) with iodobenzene
without initiator provided after oxidation triphenylphos-
The reaction of three different radicals (phenyl, ethyl and
tert-butyl) with phosphanes 1a and 1b was found to proceed
by a two-step mechanism [Eq. (4)]. The radical intermediate 7
ꢀ
phane oxide in 30% yield. Probably, thermal Sn P bond
homolysis occurs eventually initiating the radical-chain reac-
tion. Indeed, reaction of 5-hexenyl bromide with Me3SnPPh2
(1.5 equiv) in benzene without initiator at 808C for 16 h
afforded 6b (39%) and 6a (2%), clearly showing that
thermal initiation occurs. Hence, Me3SnPPh2 can act as a
radical phosphonation reagent and also as a radical initiator.
ꢀ
with a tetracoordinate phosphorus atom and an elongated P
M bond (M = Si, Sn) is stable, at least under our model
conditions (gas phase). The direct homolytic substitution can
thus be disregarded.[19]
6534
www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 6533 –6536

Angewandte
Chemie
Table 4 reports the energies of the first (DEadd) and second
step (DEdiss) of the homolytic substitution. B2-PLYP-D
predicts the intermediates 7 to be about 1–5 kcalmolꢀ1 less
stable than PBE-D does, but the overall reaction energy is
Table 4: Radical substitution energies[a] for different alkyl radicals.
Entry Phosphane Radical (R) dint
DEadd
DEdiss
(P M) [kcalmolꢀ1
[pm][b]
]
[kcalmolꢀ1
]
ꢀ
1
2
3
4
5
6
Me3SnPPh2 Ph
Me3SiPPh2 Ph
Me3SnPPh2 Et
Me3SiPPh2 Et
Me3SnPPh2 tBu
Me3SiPPh2 tBu
308.0 ꢀ28.2 (ꢀ29.2) 4.2 (7.2)
246.8 ꢀ16.6 (ꢀ19.9) 5.7 (9.6)
313.3 ꢀ12.5 (ꢀ15.8) 5.0 (7.8)
252.7
307.9
244.6
ꢀ0.7 (ꢀ5.8) 5.5 (9.4)
ꢀ7.2 (ꢀ9.3) 5.5 (8.5)
3.5 (ꢀ0.4) 6.9 (11.3)
[a]B2-PLYP-D/TZVPP//PBE-D/TZVP, including PBE-D/TZVP zero-point
ꢀ
vibrational energies (in brackets: PBE-D/TZVP energies). [b]M
distance in intermediate 7; calculated distances in phosphanes 1a and
1b are 257.3 pm (1a, M=Sn) and 230.3 pm (1b, M=Si).
P
similar. The formation of 7 is exothermic (Table 4, entries 1–
5), except for R = tBu and M = Si where DEadd is 3.5 kcal
molꢀ1 (entry 6). This is probably the reason why in the
experiment tert-butyl radical was not able to replace SiMe3 in
Figure 1. Spin densities (PBE-D/TZVP) for radical intermediates 7 with
R=Ph, M=Sn (top) and R=Ph, M=Si (bottom) (isosurface val-
ue=+0.005 a.u.).
1b in
a homolytic substitution reaction. Interestingly,
although ethyl radical addition to 1b is only slightly exother-
mic (Table 4, entry 4), good experimental yields were
obtained for the reaction of the pentyl radical with 1b. For
all radicals investigated, the addition reaction is more
exothermic for the Sn compound as compared to the
phosphorus with a trialkylsilyl radical as a leaving group.
Importantly, expensive transition metals are not necessary to
conduct these reactions. The results of our experimental
investigations are further supported by DFT calculations of
the reaction mechanism.
ꢀ
corresponding Si P derivative.
For reactions with 1a, the dissociation of the radical
complex liberating the Me3Sn radical requires less energy
than is obtained in the addition step. However, for the
homolytic substitution of the Me3Si radical with the ethyl and
tert-butyl radicals the overall process is endothermic. As
expected, the energy required for dissociation is 2–3 kcal
molꢀ1 less for the Sn derivatives (M = Sn) than for the
corresponding Si intermediates, reflecting the higher intrinsic
radical stability of SnMe3 versus SiMe3. This difference has
also an effect on the geometry and electronic structure of the
Received: April 14, 2007
Published online: July 18, 2007
Keywords: density functional calculations ·
.
homolytic substitutions · radical reactions · silicon · tin
[1]I. P. Beletskaya, M. A. Kazankova, Russ. J. Org. Chem. 2002, 38,
1447.
[2]D. Leca, L. Fensterbank, E. Lacôte, M. Malacria, Chem. Soc.
Rev. 2005, 34, 858.
ꢀ
intermediates 7. The P Sn bond in 7 (M = Sn) is more
ꢀ
elongated (about 20%) than the P Si bond in 7 (M = Si),
[3]a) W. G. Bentrude in The Chemistry of Organophosphorous
Compounds, Vol. 1 (Ed.: F. R. Hartley), Wiley, Chichester, 1990,
p. 531; b) W. G. Bentrude, Acc. Chem. Res. 1982, 15, 117.
[4]X.-Y. Jiao, W. G. Bentrude, J. Am. Chem. Soc. 1999, 121, 6088;
X.-Y. Jiao, W. G. Bentrude, J. Org. Chem. 2003, 68, 3303; B.
Ding, W. G. Bentrude, J. Am. Chem. Soc. 2003, 125, 3248.
[5]T. N. Mitchell, H.-J. Belt, J. Organomet. Chem. 1990, 386, 167.
[6]A. Sato, H. Yorimitsu, K. Oshima, J. Am. Chem. Soc. 2006, 128,
4240.
[7]Reviews on homolytic substituitions in main-group chemistry:
C. H. Schiesser, L. M. Wild, Tetrahedron 1996, 52, 13265; D.
Crich, Helv. Chim. Acta 2006, 89, 2167.
[8]K. Jones, M. F. Lappert, J. Organomet. Chem. 1965, 3, 295.
[9]For the use of Me 3SnPPh2 in transition-metal chemistry, see:
R. A. Rossi, S. E. Martꢀn, Coord. Chem. Rev. 2006, 250, 575.
[10]D. Griller, K. U. Ingold, Acc. Chem. Res. 1980, 13, 317; M.
Newcomb, Tetrahedron 1993, 49, 1151.
relative to bond lengths of the starting phosphanes 1a and 1b.
Figure 1 depicts the geometry and the spin density of the
intermediates 7. It is apparent that the spin density (the
difference between a and b electron densities) in the
intermediate 7 is more localized on Sn than on Si if the two
intermediates are compared. The excess spin density on the
phosphorus atom is found in the equatorial position of a
distorted trigonal-prismatic coordination sphere.
In conclusion we have described a highly efficient radical
phosphonation of alkyl and aryl radicals using readily
available Me3SnPPh2. Moreover, phosphonations of primary
and secondary alkyl radicals can also be performed using the
less toxic commercially available Me3SiPPh2. To our knowl-
edge, this is the first report on a homolytic substitution at
Angew. Chem. Int. Ed. 2007, 46, 6533 –6536
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
6535

Communications
[11]For reviews on tin-free radical chemistry, see: P. A. Baguley, J. C.
[16]A. Schꢁfer, C. Huber, R. Ahlrichs, J. Chem. Phys. 1994, 100,
5829.
Walton, Angew. Chem. 1998, 110, 3272; Angew. Chem. Int. Ed.
1998, 37, 3072; A. Studer, S. Amreim, Synthesis 2002, 835.
[12]S. E. Turney, J. K. Stille, J. Org. Chem. 1987, 52, 748.
[13]G. Fritz, P. Scheer, Chem. Rev. 2000, 100, 3341.
[14]S. J. Cole, J. N. Kirwan, B. P. Roberts, C. R. Willis, J. Chem. Soc.
Perkin Trans. 1 1991, 103. See also: C. Chatgilialoglu, K. U.
Ingold, J. C. Scaiano, J. Am. Chem. Soc. 1983, 105, 3292.
[15]J. P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 1996, 77,
3865.
[17]S. Grimme, J. Chem. Phys. 2006, 124, 034108.
[18]S. Grimme, J. Comput. Chem. 2004, 25, 1463; S. Grimme, J.
Comput. Chem. 2006, 27, 1787. s6(B2-PLYP-D) = 0.55. The
correction values (detailed dispersion energies) are listed in
the Supporting Information.
[19]Indeed, phosphoranyl radicals generated upon addition of
reactive radicals to trialkylphosphites have been experimentally
detected, see Ref. [3b].
6536
www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 6533 –6536