Free radical chain reactions of [1.1.1]propellane with three-coordinate phosphorus molecules. Evidence for the high reactivity of the bicyclo[1.1.1]pent-1-yl radical

Article abstract of DOI:10.1021/ja962287z

Three-substituted bicyclo[1.1.1]pent-1-yl radicals (5), generated from additions of radicals to [1.1.1 ]propellane (1), are found to have high propensities to react with three-coordinate phosphorus molecules. For example, the 3-ethylbicyclo[1.1.1]pent-1-yl radical (5d) reacts with (EtO)₃P in a free-radical Arbuzov process to yield dimethyl 3-ethylbicyclo[1.1.1]pent-1-ylphosphonate (13). By contrast, the ethyl radical does not react with (EtO)₃P to yield EtP(O)(OEt)₂. However, the highly reactive phenyl radical yields PhP(O)(OEt)₂. Moreover, attack of the n-pentylbicyclo[1.1.1]pent-1-yl radical (5b) on n-C₅H₁₁P(OMe)₂ results in a free-radical substitution process that gives dimethyl 3-n-pentyl[1.1.1]bicyclopent-1-ylphosphonite (7b) and the primary alkyl radical n-C₅H₁₁·. Thus, 5 has a propensity to bond to three-coordinate phosphorus that is greater than that of a primary alkyl radical and similar to that of phenyl radical. Reaction of 2-benzyl-4-methyl-1,3,2-dioxaphospholane (9) with 5a is found to proceed with predominant inversion of configuration at phosphorus. Furthermore, 3-phenylmethylbicyclo[1.1.1]pent-1-yl radical, 5a (from reaction of benzyl radical with 1), participates in reasonably efficient radical chain reactions with PhCH₂OP(OMe)₂ (11) and PhCH₂P(OMe)₂ (6a) (chain lengths 30-50) to provide structurally novel products of new radical-Arbuzov and radical-substitution reactions. Dimethyl 3-phenylmethylbicyclo[1.1.1]pent-1-ylphosphonate (12) from 11 and dimethyl 3-phenylmethylbicyclo[1.1.1]pent-1-ylphosphonite (7a) from 6a incorporate the bicyclo[1.1.1]pent-1-yl moiety. The high propensity of various 5 to react with three-coordinate phosphorus molecules reflects the highly pyramidal nature of 5 which is accompanied by the increased s-character of the SOMO orbital of 5 and the strength of the phosphorus-carbon bond in the presumed phosphoranyl radical intermediates (3 and 4) formed. Figure 1 depicts the thermodynamics of three classes of free radical substitution and Arbuzov reactions with three-coordinate phosphorus molecules.

Full text of DOI:10.1021/ja962287z

1388
J. Am. Chem. Soc. 1997, 119, 1388-1399
Free Radical Chain Reactions of [1.1.1]Propellane with
Three-Coordinate Phosphorus Molecules. Evidence for the
High Reactivity of the Bicyclo[1.1.1]pent-1-yl Radical
Kevin P. Dockery and Wesley G. Bentrude*
Contribution from the Department of Chemistry, UniVersity of Utah, Salt Lake City, Utah 84112
ReceiVed July 5, 1996X
Abstract: Three-substituted bicyclo[1.1.1]pent-1-yl radicals (5), generated from additions of radicals to [1.1.1]-
propellane (1), are found to have high propensities to react with three-coordinate phosphorus molecules. For example,
the 3-ethylbicyclo[1.1.1]pent-1-yl radical (5d) reacts with (EtO)3P in a free-radical Arbuzov process to yield dimethyl
3-ethylbicyclo[1.1.1]pent-1-ylphosphonate (13). By contrast, the ethyl radical does not react with (EtO)3P to yield
EtP(O)(OEt)2. However, the highly reactive phenyl radical yields PhP(O)(OEt)2. Moreover, attack of the
n-pentylbicyclo[1.1.1]pent-1-yl radical (5b) on n-C5H11P(OMe)2 results in a free-radical substitution process that
gives dimethyl 3-n-pentyl[1.1.1]bicyclopent-1-ylphosphonite (7b) and the primary alkyl radical n-C5H11•. Thus, 5
has a propensity to bond to three-coordinate phosphorus that is greater than that of a primary alkyl radical and
similar to that of phenyl radical. Reaction of 2-benzyl-4-methyl-1,3,2-dioxaphospholane (9) with 5a is found to
proceed with predominant inversion of configuration at phosphorus. Furthermore, 3-phenylmethylbicyclo[1.1.1]-
pent-1-yl radical, 5a (from reaction of benzyl radical with 1), participates in reasonably efficient radical chain reactions
with PhCH2OP(OMe)2 (11) and PhCH2P(OMe)2 (6a) (chain lengths 30-50) to provide structurally novel products
of new radical-Arbuzov and radical-substitution reactions. Dimethyl 3-phenylmethylbicyclo[1.1.1]pent-1-ylphos-
phonate (12) from 11 and dimethyl 3-phenylmethylbicyclo[1.1.1]pent-1-ylphosphonite (7a) from 6a incorporate the
bicyclo[1.1.1]pent-1-yl moiety. The high propensity of various 5 to react with three-coordinate phosphorus molecules
reflects the highly pyramidal nature of 5 which is accompanied by the increased s-character of the SOMO orbital of
5 and the strength of the phosphorus-carbon bond in the presumed phosphoranyl radical intermediates (3 and 4)
formed. Figure 1 depicts the thermodynamics of three classes of free radical substitution and Arbuzov reactions with
three-coordinate phosphorus molecules.
Introduction
addition to phosphorus (reactions 1a, 2a), have been character-
ized by ESR spectroscopy.10b,d,e A key process undergone by
three-coordinate phosphorus molecules on reaction with radicals
is substitution (eq 1) via R scission (path 1b) of the phosphoranyl
The strained hydrocarbon [1.1.1]propellane (1) and the highly
pyramidal 3-substituted bicyclo[1.1.1]pent-1-yl radical (2),
generated by addition of X• to 1, have been the subject of
numerous structural and chemical studies in recent years.1,2
Propellane 1 is of special interest because of the thermodynamic3
and kinetic7 ease of additions of free radicals to it. The tertiary
alkyl product radical 2 has attracted attention because of its
highly pyramidal structure and consequent potentially increased
reactivity.8 Thus, 2 might be expected to form relatively strong
bonds and feature reduced steric demands compared to those
of unconstrained tertiary alkyl radicals. Comparisons of the
reactivity of radical 2 with those of unstrained primary,
secondary, and tertiary alkyl radicals are of particular interest.
radical intermediate 3. Another is the addition of a radical (path
2a) followed by â scission (path 2b) of an alkoxy group already
attached to phosphorus. Reaction 2 has been termed a free-
radical Arbuzov reaction,10 a name particularly appropriate when
R• is an alkyl radical, although this reaction is also known for
R• ) Me2N•11 and Ph•.12
The rate constants for oxidative addition of radicals to three-
coordinated phosphorus, the ease of the ensuing R and â scission
reactions of 3 and 4, and the overall thermodynamic ease of
reactions 1 and 2 are correlated by the relative strengths of the
new bonds to phosphorus formed in intermediates 2 and 3 and
The reactions of free radicals with three-coordinate phos-
phorus molecules present an attractive diagnostic system for
the study of radical reactivity. The majority of these reactions
undoubtedly proceed through intermediate phosphoranyl radicals
(3 and 4),10 a great number of which, formed via oxidative
(1) For recent reviews on the chemistry of small ring propellanes and
the formation of bridgehead free radicals, see: (a) Kaszynski, P.; Michl, J.
In The Chemistry of the Cyclopropyl Group; Rappoport, Z., Ed.; Wiley:
New York, 1995; Vol. 2, p 773. (b) Szeimies, G. In AdVances in Strain in
Organic Chemistry; Halton, B., Ed.; JAI Press Ltd.: Greenwich, 1992; Vol.
2, p 1. (c) Walton, J. C. Chem. Soc. ReV. 1992, 21, 104. Wiberg, K. B.
Angew. Chem., Int. Ed. Engl. 1986, 25, 312. (d) Kaszynski, P.; Michl, J. In
AdVances in Strain in Organic Chemistry; Halton, B., Ed.; JAI Press:
London: 1995; Vol. 4, p 283. (e) Tobe, Y. In Carbocylic Cage Com-
pounds: Chemistry and Applications; Osawa, E., Yonemitsu, O., Eds.;
VCH: New York, 1992; p 125. (f) Wiberg, K. B. Chem. ReV. 1989, 89,
975. (g) Ginsburg, D. In The Chemistry of the Cyclopropyl Group;
Rappoport, Z., Ed.; Wiley: New York, 1987; Vol. 1, p 1193.
X Abstract published in AdVance ACS Abstracts, January 15, 1997.
S0002-7863(96)02287-1 CCC: $14.00 © 1997 American Chemical Society

Free Radical Chain Reactions of [1.1.1]Propellane
J. Am. Chem. Soc., Vol. 119, No. 6, 1997 1389
those in the reactants and products.10a-d,f,13,14 Thus, radicals
that form relatively strong bonds to phosphorus (alkoxy, thiyl,
and phenyl radicals) add to trialkyl phosphites, (RO)3P, ir-
reVersibly with second-order rate constants10a of 3 × 108-5 ×
109 M-1 s-1. By contrast ESR studies17 have demonstrated that
methyl radicals add reVersibly to a series of (RO)3P (reaction
2a, R• ) Me•) and only give â scission (reaction 2b) when R′•
is the relatively stable benzyl radical. The efficiency of reaction
1 depends on both the ability of X• to form a bond to
phosphorus, at least reversibly, and the relative energies of the
P-Z and P-X bonds.
In the present study a comparison is made of the abilities of
the 3-substituted bicyclo[1.1.1]pent-l-yl radical (2 and 5) and
the radicals ethyl, isopropyl, and benzyl to undergo free radical
Arbuzov (eq 2) and substitution (eq 1) processes with certain
three-coordinate phosphorus molecules. This research is based
on the known facile additions of a variety of radicals to 1 to
give intermediate 5 (eq 3).1-3,7 The important conclusion of
these studies is that 3-substituted bicyclo[1.1.1]pent-1-yl radicals
haVe a propensity to undergo reactions 1 and 2 that is greater
than that of primary radicals and closer to that of phenyl
radicals. Furthermore, because of its high reactivity and ease
of formation from 1, the 3-substituted bicyclo[1.1.1]pent-1-yl
radical (5) participates in two new free-radical chain reactions
that yield novel insertion products (7a-c, 9, and 12) containing
the bicyclo[1.1.1]pent-1-yl structural moiety.15
(2) For recent papers on the free radical chemistry of [1.1.1]propellane,
see: (a) Scaiano, J. C.; McGarry, P. F. Tetrahedron Lett, 1993, 34, 1243.
(b) Kaszynski, P.; Friedli, A. C.; Michl, J. J. Am. Chem. Soc. 1992, 114,
601. (c) Obeng, Y. S.; Laing, M. E.; Friedli, A. C.; Yang, H. C.; Wang, D.;
Thulstrup, E. W.; Bard, A. J.; Michl, J. J. Am. Chem. Soc. 1992, 114, 9943.
(d) Kaszynski, P.; McMurdie, N. D.; Michl, J. J. Org. Chem. 1991, 56,
307. (e) Wiberg, K. B.; Waddell, S. T. J. Am. Chem. Soc. 1990, 112, 2194.
(f) Sadovaya, N. K.; Blokhin, A. V.; Surmina, L. S.; Tyurekhodzhaeva, M.
A.; Koz’min, A. S.; Zefirov, N. S. IzV. Akad. Nauk SSSR, Ser. Khim. 1990,
2451. (g) Sadovaya, N. K.; Blokhin, A. V.; Tyurekhodzhaeva, M. A.;
Grishin, Yu.K.; Surmina, L. S.; Koz’min, A. S.; Zefirov, N. S. IzV. Akad.
Nauk SSSR., Ser. Khim. 1990, 716. (h) Kaszynski, P.; Friedli, A. C.;
McMurdie, N. D.; Michl, J. Mol. Cryst. Liq. Cryst. 1990, 191, 193. (i)
Hassenrueck, K.; Murthy, G. S.; Lynch, V. M.; Michl. J. J. Org. Chem.
1990, 55, 1013. (j) McGarry, P. F.; Johnston, L. J.; Scaiano, J. C. J. Am.
Chem. Soc. 1989, 111, 3750. (k) McGarry, P. F.; Johnston, L. J.; Scaiano,
J. C. J. Org. Chem. 1989, 54, 6133. (l) Murthy, G. S.; Hassenrueck, K.;
Lynch, V. M.; Michl, J. J. Am. Chem. Soc. 1989, 111, 7262. (m) Bunz, U.;
Polborn, K.; Wagner, H.-U.; Szeimies, G. Chem. Ber. 1988, 121, 1785. (n)
Kaszynski, P.; Michl, J. J. Am. Chem. Soc. 1988, 110, 5225. (o) Kaszynski,
P.; Michl, J. J. Org. Chem. 1988, 53, 4593. (p) Zefirov, N. S.; Sadovaya,
N. K.; Surmina, L. S.; Godunov, I. A.; Koz’min, A. S.; Potekhin, K. A.;
Maleev, A. V.; Struchkov, Yu.T. IzV. Akad. Nauk SSSR, Ser. Khim. 1988,
2648. (q) Wiberg, K. B.; Waddell, S. T.; Laidig, K. Tetrahedron Lett. 1986,
27, 1553.
Results
Free Radical Substitution-Propellane Insertion Reactions.
Tests of the abilities of 3-substituted bicyclo[1.1.1]pent-1-yl
radicals (5), generated on addition of an alkyl radical (R•) to 1
(eq 3), to displace benzyl and n-pentyl radicals from phospho-
(3) The strength of the central bond of [1.1.1]propellane (1), based on
theoretical estimates and experimental measurements, is 60-70 kcal/mol4-6
which is close to that of cyclopropane. The value for the heat of addition
of H2 to the central C-C bond of [1.1.1]propellane (1) is -37.9 kcal/mol
(Kozina, M. P.; Pimenova, S. M.; Lukyanova, V. A.; Surmina, L. S. Dokl.
Akad. Nauk SSSR 1985, 283, 661). The bond dissociation energy for the
bridge head C-H bond of bicyclo[1.1.1]pentane is calculated to be 104.4
kcal.6
nites 5a and 5b (eq 4) were carried out under photoinitiation
(4) Wiberg, K. B.; Walker, F. H. J. Am Chem. Soc. 1982, 104, 5239.
(5) Feller, D.; Davidson, E. R. J. Am. Chem. Soc. 1987, 109, 4133.
(6) Wiberg, K. B.; Hadad, C. M.; Sieber, S.; Schleyer, P. v. R. J. Am.
Chem. Soc. 1992, 114, 5820.
(7) A series of radicals, tertiary butoxy, phenylthiyl, triethylsilyl, benzene-
complexed chlorine atom, and p-methoxybenzoyl add to [1.1.1]propellane
(1) about 2-3 times more rapidly than they add to styrene. (Scaiano, J. C.,
private communication. Data given in ref 1a).
(8) Bicyclopentyl radical (2, X ) H) is estimated from ESR measure-
ments to have about 20% 2s character in the SOMO (Aiso 13C ) 223 G at
-196 °C.9
(9) Rhodes, C. J.; Walton, J. C.; Della, E. W. J. Chem. Soc., Perkin
Trans. 2 1993, 2125.
(10) For reviews of phosphoranyl radicals, see: (a) Bentrude, W. G. In
The Chemistry of Organophosphorus Compounds; Hartley, F. R., Ed.;
Wiley: Sussex, 1990; Vol. 1, p 531; (b) Bentrude, W. G. In ReactiVe
Intermediates, Abramovitch, R. A., Ed.; Plenum: London, 1983; Vol. 3, p
199. (c) Bentrude, W. G. Acc. Chem. Res. 1982, 15, 117. (d) Roberts, B. P.
In AdVances in Free Radical Chemistry; Williams, G. H., Ed.; Heyden and
Sons: London, 1980; Vol. 6, p 225. (e) Schipper, P.; Janzen, E. H. J. M.;
Buck, H. M. Top. Phosphorus Chem. 1977, 9, 407. (f) Bentrude, W. G. In
Free Radicals; Kochi, J. K., Ed.; Wiley-Interscience: New York, 1973; p
595.
conditions in 100 µL of deoxygenated benzene solutions of GC-
purified 1 (≈0.5 mmol) and excess 6 (0.7-1.4 mmol). The
reaction with 5b was initiated with di-tert-butyl peroxide (0.02
mmol) while that with 6a was more efficient with the benzyl
radical source, bis(phenylmethyl)diazene BPMDA (0.04 mm),
as initiator. The yields of these reactions, based on consumed
phosphonite, were determined by 31P NMR spectroscopy
following oxidation of phosphonites 7a and 7b to the phospho-
nates (7a-O and 7b-O) by t-BuOOH and are recorded in Table
1 along with the isolated yield (12%) of pure 7b-O. (7a-O is
the same as 12 formed in eq 5.) The yields are considered to
(11) Bentrude, W. G.; Alley, W. D.; Johnson, N. A.; Murakami, M.;
Nishikida, K.; Tan, H.-W. J. Am. Chem. Soc. 1977, 99, 4383.
(12) (a) Fu, J.-J. L.; Bentrude, W. G. J. Am. Chem. Soc. 1972, 94, 7710.
(b) Fu, J.-J. L.; Bentrude, W. G.; Griffin, C. E. J. Am. Chem. Soc. 1972,
94, 7717. (c) Kryger, R. G.; Lorand, J. P.; Stevens, N. R.; Herron, N. R. J.
Am. Chem. Soc. 1977, 99, 7589. (d) Reference 11 and 12 are concerned
with reactions of dimethylamino and phenyl radicals with trialkyl phosphites.
(e) With phenylphosphonites (PhP(OR)2) and diphenylphosphinites (Ph2-
POR), a greater range of radicals promote free radical Arbuzov reactions,
as summarized in ref 10. This may be related to the ligand-π structure10a
of phenyl-substituted phosphoranyl radicals which may be of increased
stability compared to their TBP σ radical counterparts.
(13) Tables of first-order rate constants for reactions of radicals with
three-coordinate phosphorus molecules are found in refs 10a-10d.
(14) This sort of understanding of the reactions of radicals with three-
coordinate phosphorus molecules, in terms of three reaction classes, was
first set forth in refs 15 and 16.
be conservative because of losses in transfers of the phosphonite
solution to the reaction tube (see Experimental Section) and
possible less-than-quantitative oxidation of products. Nonethe-
less, reasonable amounts of 7a and 7b as their oxides were
obtained. A few percent of 7a-O was isolated in pure form by
HPLC. Major peaks in the 31P NMR spectra of the crude
(15) Bentrude, W. G.; Hansen, E. R.; Khan, W. A.; Min, T. B.; Rogers,
P. E. J. Am. Chem. Soc. 1973, 95, 2286. Bentrude, W. G.; Rogers, P. E. J.
Am. Chem. Soc. 1976, 98, 1674.
(16) Bentrude, W. G.; Fu, J.-J. L.; Rogers, P. E. J. Am. Chem. Soc. 1973,
95, 3625.
(17) Davies, A. G.; Griller, D.; Roberts, B. P. J. Chem. Soc., Perkin
Trans. 2 1972, 2224.

1390 J. Am. Chem. Soc., Vol. 119, No. 6, 1997
Dockery and Bentrude
Table 1. Free Radical Chain Reactions of Three-Coordinate
Phosphorus with [1.1.1]Propellane at 0 °C
Table 2. Mean Chain Lengths for the Reactions of 6a, 7, and 11
with [1.1.1]Propellane
%
%
%
%
% D
chain
reactanta mmol initiator mmol conversionb product yieldc
compd reactanta conversionb yieldb incorprtnc product lengthd
6a
1.33 BPMDA 0.02
31
49
7a
9
46
40
6a
6a
50
41
49
28
77
36
37
40
61
45
1.6
1.8
2.0e
1.8
2.6
7a
7a
9
12
12
46
41
37
41
28
trans- and 1.23 BPMDA 0.02
cis-8
trans- and cis-8
11
11
11
1.38 BPMDA 0.03
28
69
15e
f
12
7b
7c
61 (45)
69 (12)
f (7)
6b
0.67 DTBPd
0.77 TMT
0.02
0.5
0.66
6ch
a 6a and its products were analyzed as the oxide, trans- and cis-8
and their products as the sulfides. b By quantitative 31P NMR. c Deter-
mined by GC-EIMS (70 eV) from the product [M + 1]+/[M]+ ratio
with labeled initiator BPMDA-d2 vs the product [M + 1]+/[M]+ ratio
with unlabeled BPMDA. d Determined from the ratio of unlabeled to
labeled product and multiplying by 0.75 (75% D incorporation in
BPMDA-d2). e From the ratio of fragment m/z 92 to fragment m/z 91.
(EtO)3Ph 11.9 Et2N2
13
15 (6)g
a 6a-6c and 8 were converted to the oxide or sulfide, the products
7a and 7b to the oxides, and products 7c and 9a to the sulfides.
b Determined by quantitative 31P NMR vs 0.317 M (MeO)3PO in C6D6
as external standard. c By quantitative 31P NMR based on disappearance
of starting compound. Numbers in parentheses indicate isolated yields.
d Di-tert-butyl peroxide. e Estimated from crude 31P NMR. Conversion
f
too low to determine. g Yield based on total starting [1.1.1]propellane.
stereospecific retentive19 conversion of the diastereomers of 9
with S8 to the sulfides, cis- and trans-9-S. The latter were found
to be in trans:cis ratio of 38:62.19 The sulfides (8-S) of
unreacted starting 8 (trans:cis ratio, 67:33) were purified as a
mixture, trans:cis ratio, 67:33. The predominant stereochemistry
of this displacement reaction, therefore, is inVersion.
The small increase in percentage of trans product, compared
to that predicted from the 71:29 trans:cis ratio of starting
phosphonite 8, could come from a slightly greater reactivity of
the cis isomer or from a partial equilibration of the diastereomers
of 8. Both possibilities are consistent with the reduction in trans:
cis ratio of the reactants from 71:29 to 67:33 over the course of
the reaction. However, the trans:cis ratio of products 9-S, 38:
62, is far removed from the thermodynamic equilibrium trans:
cis ratio known for 9 of about 65:35.19
h At 24 °C.
reactions mixtures could be reasonably assigned to 7a (δ 172.6)
and 7b (δ 172.0). For 6a assignments of 1H NMR resonances
for the nonaromatic protons in the crude product mixture also
could be made by comparisons to resonances for purified
analogs. Purified phosphonate 7b-O was fully characterized
by H and 13C NMR spectroscopy and by both high and low
1
resolution mass spectrometry (see later paragraph and Experi-
mental Section). Isolated phosphonate 7a-O was identified by
comparison of its spectral properties and GC retention time to
those of a sample of 7a-O (12) that was obtained in greater
quantities from the reaction of (MeO)2POCH2Ph with [1.1.1]-
propellane (1) (see below, eq 5).
The stereochemistry of the displacement of the benzyl group
by the 3-benzylbicyclo[1.1.1]pent-1-yl radical (5a, R ) PhCH2)
was determined from the reaction of a 71:29 (trans:cis) ratio of
starting diastereomers of phosphonite 819 to give phosphonites
cis- and trans-9.20 (A Kimax filter was employed to avoid
photoisomerization of 8.21) These reactions are illustrated below
using cis-8. Major peaks assignable to these phosphonites were
seen in the 31P NMR spectra of the crude product mixture at δ
183.3 (cis) and δ 177.7 (trans). An overall yield of 40% for
The structure of 9-S was determined spectroscopically. 1H
and 31P NMR spectra were obtained from very small amounts
of pure individual diastereomers, while a pure cis/trans mixture
of isomers yielded 13C NMR parameters and HRMS data (see
later paragraph and Experimental Section).
Since these insertion reactions are obviously radical chain
processes, the mean kinetic chain lengths for the benzyl-radical
induced reactions of [1.1.1]propellane (1) with phosphonite 6a
and cyclic phosphonite 8 were determined by use of deuterium-
labeled BPMDA from which PhCHD• was generated photolyti-
cally. Reaction conditions and workup duplicated those for the
reactions with unlabeled BPMDA. Data from these studies are
found in Table 2. The percentage of product with deuterium
incorporation was determined from the increase in the [M + 1]
ion intensity in the mass spectrum relative to product from
reaction using unlabeled BPMDA. (GC/MS on 7a-O, trans-
and cis-9-S). Equations 3 and 4 (R ) PhCH2) were assumed
to be operative. The chain length was then calculated from the
ratio, in the oxidized product mixture, of the unlabeled
3-benzylbicyclo[1.1.1]pent-1-ylphosphonate (7a-O or 9-S) con-
taining PhCH2 to the labeled 3-benzylbicyclo[1.1.1]pent-1-
ylphosphonate (7a-O or 9-S) containing PhCHD. No deuterium
incorporation into recovered starting material could be detected.
Kinetic chain lengths of the order 30-50, recorded in Table 2,
show these reactions to be reasonably efficient chain processes.
(MS data on which Table 2 is based are found in the Supporting
Information.)
this process was determined (Table 1) by 31P NMR following
(18) Preliminary studies of the reactions of alkyl radicals with trialkyl
phosphites and alkylphosphonites were carried out in this laboratory by E.
R. Hansen and P. E. Rogers. For a preliminary report of a portion of the
results reported in the present paper, see: Dockery, K. P.; Bentrude, W. G.
J. Am. Chem. Soc. 1994, 116, 10332.
(19) The preparation and assignment of cis and trans geometries to 8
and its sulfides were reported previously from this group. Bentrude, W.
G.; Tan, H.-W. J. Am. Chem. Soc. 1976, 98, 1850. Cis and trans designations
refer to the relation of the ring methyl and alkyl substituent on phosphorus.
The stereospecific conversion of three-coordinate phosphorus to the sulfide
with retention of configuration is well established. (Young, D. P.; McEwen,
W. E.; Velez, D. C.; Johnson, J. W.; Van-derWerf, C. A. Tetrahedron Lett.
1964, 359. Horner, L.; Winkler, H. Tetrahedron Lett. 1964, 175.)
(20) The assignment of cis and trans stereochemistries were made from
the relative 31P chemical shifts, a previously established criterion.15
(21) Unpublished results from this laboratory.
Dialkylamino groups have been shown to be readily displaced
from phosphorus by tert-butoxy radicals15 which undoubtedly
form strong phosphorus-oxygen bonds. Reaction of phos-
phoramidite 6c with [1.1.1]propellane (1), under conditions of
photoinitiation by tetramethyltetrazine (Me2NNdNNMe2) at O
°C, gave only trace amounts of insertion product 7c, presumably
generated by attack of 5c (R ) Me2N) on phosphoramidite 6c
(eqs 3 and 4). However, at higher temperatures (24 °C), use

Free Radical Chain Reactions of [1.1.1]Propellane
J. Am. Chem. Soc., Vol. 119, No. 6, 1997 1391
of near-stoichiometric amounts of initiator (0.77 mmol 6c, 0.60
mmol tetramethyltetrazine) and higher dilution in C6D6 gave
insertion product 7c, isolated as the sulfide (7c-S) (85% purity).
The initial product 7c was evident in the crude reaction mixture
from its 31P NMR resonance at δ 171.9. The isolated yield of
7c in Table 5 (7%) takes into account its 85% purity and is a
rough estimate at best in view of the very low conversion of
phosphoramidite 6c. 1H NMR parameters from the 85% pure
7c-S sample were readily assignable by comparisons to analo-
gous compounds, 7b-O, 9-S, and 12. GC-MS fragmentation
patterns and HRMS confirmed the assignment (see later
paragraph and Experimental Section).
Free-Radical Arbuzov-Propellane Insertion Reactions.
The potential reaction of 3-substituted bicyclo[1.1.1]pent-1-yl
intermediate 5a (R ) PhCH2), formed via eq 3, with benzyl
dimethyl phosphite (11) to yield adduct phosphonate 12 (eq 5)
was explored. The optimal conditions determined for this
reaction involved photoinitiation by BPMDA, rather than di-
tert-butyl peroxide, and use of excess 11 (Table 1). Both 31P
NMR (61%) and isolated (45%) yields of 12 are quite good.
excess phosphite (phosphite/1 approximately 12) and a lower
phosphite/azobisethane ratio. Isolable amounts of phosphonate
13 were obtained. The 31P NMR yield, about 15% (6%
isolated), was based on total starting [1.1.1]propellane (1) instead
of (EtO)3P, because the very low conversion of the phosphite
could not be accurately measured.
Finally, a reaction parallel to eq 6 was carried out at 0 °C
with (MeO)3P (40 mmol) in 30 mL of an ether solution of crude
[1.1.1]propellane (1) using BPMDA (0.85 mmol) as the radical
source. Insertion product, phosphonate 12 (identified by
comparison of spectral data to those of 12 from reaction of 6a
with 1) was generated in low, undetermined yield along with
some bis adduct 10-O (R ) PhCH2). A large amount of
bibenzyl also was formed. This is clearly a nonchain process.
Furthermore, the analog of 12 with methyl in place of benzyl
apparently also was formed (GC-MS evidence only).
Evidence for Product Structure from NMR and MS Data.
Detailed listings of 31P, 13C, and 1H NMR data for the insertion
products 7-O, 9-S, 10-O, 12, and 13 are found in the
Experimental Section and fully support the structures written.
Indicative of the incorporation of the bicyclo[1.1.1]pentyl ring
1
system were the H chemical resonances for the methylene
groups of that ring at δ 1.72-1.94 that featured three-bond
couplings to phosphorus of 0.9-1.3 Hz. Molecules containing
a methylene group attached to the bridgehead carbon, i.e., 7a-
O, 9-S, 12, and 13 (substituent at C3 ) CH2CH3, CH2C6H5),
resulting from addition of ethyl or benzyl radicals to [1.1.1]-
propellane (1) (eq 3), displayed 0.7-1.5 Hz five-bond proton-
phosphorus couplings. The ring methylene carbon resonances
of the insertion products all appeared at δ about 50 ppm with
2JCP = 2 Hz. Bridgehead carbons (C1, C3) were distinguished
Insertion product 12, a solid, was thoroughly characterized by
1
31P, H, and 13C NMR spectroscopy, GC-MS fragmentation
patterns, and HRMS (see later paragraph and Experimental
Section). When the reaction was run with an excess of [1.1.1]-
propellane (1) (0.57 mmol of 11, approx. 1 mmol of 1), the
crystalline bis adduct (10-O) was isolated in 8% yield (11%
NMR yield). By contrast the substitution-insertion reactions,
discussed earlier, of 6a and 6b with [1.1.1]propellane (1) to
yield 7a and 7b (eq 4) gave bis adducts, e.g., phosphonite 10
(reactions of 1 with 6a), in only barely detectable amounts (GC/
MS evidence only). (See Experimental Section and later
paragraph for characterization of 10-O.)
3
by large couplings to phosphorus (1JCP = 150 Hz, JCP = 50
Hz). The products 7b-O, 9-S, 10-O, and 12 displayed long-
4
range JP-C couplings, transmitted through the bicyclo[1.1.1]-
pentyl framework, of 25.8-26.9 Hz.
The EI/MS fragmentation patterns for 7a-O, 7b-O, 7c-S, 9-S,
12, and 13 strongly supported their structures. Peaks assignable
to the 3-substituted bicyclo[1.1.1]pent-1-yl ion, the (MeO)2P-
(O) or (EtO)2P(O) ion, and the 3-substitutents were seen in each
case and often had intensities 10-25% or more of the base peak.
The spectral parameters, NMR and MS, compare well to those
found for many other molecules containing the bicyclo[1.1.1]-
pentyl structural unit.2
Nonchain Reactions of Alkyl Radicals with Trialkyl
Phosphites and Phosphonites 6a and 6b. In order to better
evaluate the propensities of 3-substiuted bicyclo[1.1.1]pent-1-
yl radicals, formed via reaction 3, to participate in free radical
substitution (eq 4) and Arbuzov (eq 5) reactions with those of
other radicals, parallel comparison reactions of primary and
secondary alkyl radicals and benzyl radicals were carried out.
Equation 7 depicts the potential nonchain free radical ArbuzoV
reaction of alkyl radicals with dimethyl benzyl phosphite (11)
to yield dimethyl alkylphosphonates 14a-c, presumably via
phosphoranyl radical 4. When deoxygenated benzene solutions,
Table 2 records the kinetic chain length for formation of 12,
again determined by use of deuterated BPMDA. The chain
mechanism assumed is the repeating steps of eqs 3 (R ) PhCH2)
and 5 and again is reasonably efficient. (The involvement of a
phosphoranyl radical in reaction 5 will be discussed below.)
The lower number (28) determined at higher conversions of
phosphite 11 probably reflects the decrease in rate of chain-
carrying step 5 as the concentration of 11 decreases, and other
processes compete more effectively for 3-substituted bicyclo-
[1.1.1]pent-1-yl radical 5a.
A less efficient radical-Arbuzov trapping of 5 by a phosphite
((EtO)3P) is seen in the overall reaction 6 (presumed intermedi-
ate 5d, R ) Et). Under conditions that were successful for
reaction 5 (Table 1), no 13 could be detected. The use of
(EtO)3P as solvent at 0 °C likewise did not give insertion product
13. Large quantities of a waxy material were formed, along
with what may be polymeric phosphonates. Use of azobisethane
instead of di-tert-butyl peroxide, however, was successful when
the reaction was carried out at 24 °C instead of 0 °C.
Additionally, more dilute conditions were employed along with
0.23 M in both azobisethane and 11, were irradiated at 24 °C

1392 J. Am. Chem. Soc., Vol. 119, No. 6, 1997
Dockery and Bentrude
through Pyrex, a 90% yield of 14a (R ) Et) resulted, based on
consumed phosphite 11 (18%), along with generous amounts
of bibenzyl. This finding is clearly consistent with the operation
of the two-step process depicted by eq 7 which is analogous to
process 2. Significantly, the benzyl radicals generated failed
to take part in a potential chain reaction analogous to 7 but
with R ) PhCH2. Thus, no dimethyl benzylphosphonate (14c)
was noted (31P NMR, GLC). By contrast to the case with
azobisethane, the parallel photogeneration of isopropyl radicals
from azobisisopropane (0.28 M) in a 0.35 M benzene solution
of phosphite 11 did not yield dimethyl isopropylphosphonate
(14b) or any other new phosphorus-containing molecule (31P
NMR).18 Furthermore, the photolysis of azobisethane (0.27 M)
in benzene solution 0.28 M in (MeO)3P or (EtO)3P led to no
new phosphorus-containing products (e.g., EtP(O)(OMe)2 or
EtP(O)(OEt)2)) or evidence for consumption of the phosphites
(31P NMR) by any other reaction.
However, isopropyl radicals, photogenerated in deoxygenated
benzene, were reactiVe with phosphonite 6a in a substitution
process, eq 8 (R′ ) isoPr), to generate dimethyl isopropylphos-
phonite (15a) with bibenzyl as biproduct. The yield of dimethyl
isopropylphosphonite 15a, based on conversion of 6a (69%),
was 36% (31P NMR of the unoxidized product mixture, tri-n-
propyl phosphate internal standard). Ethyl radicals also reacted
cleanly (31P NMR) with dimethyl benzylphosphonite (6a)
according to eq 8 to form dimethyl ethylphosphonate 15b. (The
The failure of ethyl radical to react in radical-Arbuzov fashion
with (EtO)3P, when its reaction with 11 (eq 7) occurs so readily,
can be easily understood in terms of the relatively slow â
scission of the phosphoranyl radical of sequence 9 (analogous
to reaction 7) which cannot compete with R scission to reform
ethyl radical and the phoshpite. These results give chemical
confirmation to previously reported ESR studies of the reVersible
addition of methyl radical to PhCH2OP(OMe)2, (MeO)3P and
(EtO)3P. Only PhCH2OP(OMe)2 gave evidence of â scission,
i.e., the rapid formation of benzyl radical (eq 7, R ) Me•),
though reversible phosphoranyl radical formation was noted in
other cases.17 For the process (EtO)3P f EtP(O)(OEt)2 ∆H°
has been measured (-25 kcal/mol).22 However, kinetic factors
preclude its catalysis by ethyl radicals.
Both ethyl and isopropyl radicals replace benzyl radical in
PhCH2P(OMe)2 (6a) in a radical substitution process, eq 8. This
is a reasonable result in terms of the overall thermodynamics
of process 8 with these combinations of R and R′ radicals. It
also likely means that isopropyl radical is able to bond to 11
and 6a to form both 4 and 3, respectively. However, only in
reaction 8 is the scission step (R-scission, eq 8, 3 f 15a) rapid
enough to give product formation in competition with reversal
of the formation of 3 and 4. Indeed, R scission is often more
rapid than competing â scission options for phosphoranyl
radicals.3 The failure of isopropyl radical to replace n-pentyl
radical from phosphonite 6b to afford 15a is totally consistent
with the unfavorable thermodynamics of the overall process and
the faster R scission reforming reactants (reverse of reaction
8a).
The kinetics and thermodynamics of formation of 3 and 4,
of course, are closely intertwined in that for a reversible
endothermic process, the activation free energy is at least as
great as the free energy change for the reaction. Thus it was
noted that the thermoneutral exchange of benzyl groups was
not observed when deuterium-labeled benzyl radical was
generated in the presence of PhCH2P(OMe)2 (6a). Likely, the
formation of 3 (R ) R′ ) PhCH2) is thermodymanically and
kinetically unfavorable.
Earlier work from this laboratory showed that a facile radical
Arbuzov reaction ensues (reaction 7) with PhCH2OP(OMe)2
when R ) Me2N but not with (MeO)3P and (EtO)3P.11 Clearly,
the dimethylamino radical (like the ethyl radical) bonds suf-
ficiently well to add reversibly to phosphorus; but product
formation requires rapid subsequent â scission. Phenyl radical,
however, forms a strong bond with phosphorus. Phosphites of
all kinds undergo radical Arbuzov reactions, as exemplified by
sequence 2, when R ) Ph and R′ ) Me. We have shown the
formation of 4 to be highly irreVersible for R ) Ph.16
There are then three general classes of reactions of free
radicals with three-coordinate phosphorus.14 When the radical
bonds too weakly to phosphorus (e.g., benzyl), no reaction
ensues. When the bonding is reVersible (e.g., ethyl and
dimethylamino), a sufficiently rapid R or â scission step will
trap the intermediate phosphoranyl radical to generate stable
product. IrreVersible formation of the phosphoranyl radical
intermediate by radicals such as phenyl that form very strong
bonds to phosphorus gives product even in the absence of
particularly rapid R or â scission step. Irreversible oxidative
addition occurs with alkoxy, thiyl, and phenyl radicals and is
yield of 15b was not determined.) However, the isopropyl
radical failed to replace the presumably more strongly bound,
less stable n-pentyl radical from the n-pentylphosphonite 6b to
form isopropylphosphonite 15a.
Discussion
Nonchain Radical Substitution and Arbuzov Reactions of
Alkylphosphonites and Phosphites. These processes are
expressed in terms of the assumed phosphoranyl radicals, 3 and
4, by eqs 7 (Arbuzov) and 8 (substitution). For these processes
to proceed, the intermediates 3 and 4 must be accessible
kinetically, and the overall reactions must be thermodynamically
favorable. The results reported above established an ability to
effect overall Arbuzov reaction or substitution, reactions for
radicals R (eq 7) and R′ (eq 8) of Et > i-Pr > PhCH2.
Thus, in the radical ArbuzoV reactions (eq 7), ethyl radical
reacts readily with PhOCH2OP(OMe)2 (11) to generate dimethyl
ethylphosphonate (14a). However, the benzyl radical, generated
on â scission of 4, fails to react analogously with PhCH2OP-
(OMe)2 (11) to yield PhCH2P(O)(OMe)2 (14c). Similarly,
isopropyl radicals, generated independently in the presence of
phosphite 11, do not give i-PrP(O)(OMe)2 (14b). Clearly â
scission of intermediate 4, when R ) Et, is rapid enough to
compete with the reverse of step 7a (R scission) to regenerate
11 and ethyl radical. When R is an i-Pr or benzyl radical,
â-scission of 4 to yield phosphonate 14b or 14c does not
compete with a more rapid R scission which now gives more
stable radicals, isopropyl and benzyl. The fact that exchange
of deuterium labeled benzyl radical with 6a (eq 8) is not
observed makes questionable the ability of the very stable benzyl
radical to bond to three-coordinate phosphorus to form either 3
or 4.
(22) Lewis, E. S.; Colle, K. S. J. Org. Chem. Soc. 1981, 46, 4369. An
earlier estimate of -44 kcal/mol was based on measured and calculated
heats of formation from the literature.12b

Free Radical Chain Reactions of [1.1.1]Propellane
J. Am. Chem. Soc., Vol. 119, No. 6, 1997 1393
Figure 1. Free-radical Arbuzov and substitution reactions. Energetics of phosphoranyl radical formation and subsequent R and â scissions.
then followed by R scission (substitution) or â scission with
transfer of attacking oxygen or sulfur (oxidation). These ideas
are depicted energetically in Figure 1, which includes both
substitution and Arbuzov processes using a single reactant A,
and will be discussed more thoroughly below.14,23
Free-Radical Chain Substitution-Insertion and Arbuzov-
Insertion Reactions. The above studies of free-radical reactions
of dimethyl benzylphosphonite (6a) and n-pentylphosphonite
(6b) and the free-radical Arbuzov reactions of trialkyl phosphites
provide a background against which to assess the propensity of
3-substituted bicyclo[1.1.1]-pent-1-yl radicals (5) to undergo
analogous reactions. The observed radical insertion-substitution
processes that incorporate the bicyclo[1.1.1]pent-1-yl moiety into
7 are reasonably depicted by a chain sequence, reactions 3 and
10. The results in Table 1 show that radical 5 is able to attack
phosphorus (eq 10a) to displace not only benzyl radical to give
7a but also the primary n-pentyl radical to yield 7b. This means
that the new phosphorus-carbon bond formed in 7b is not only
stronger than the phosphorus-carbon bond to the benzyl
substituent broken in 6a but is even stronger than the n-pentyl-
phosphorus bond in 6b. For 7a a mean chain length of about
then reacts readily with [1.1.1]propellane (1) to give 5a in chain
carrying reaction 3.25 Remarkably, as in the substitution reaction
(eq 8), the tertiary 3-benzylbicyclo[1.1.1]pent-1-yl radical (5a)
evidently forms a stronger bond to phosphorus in 17 than does
the unreactive secondary isopropyl radical and, by this com-
parison, a bond at least comparable in strength to that formed
by ethyl radical. (Recall that the isopropyl radical fails to react
with PhCH2OP(OMe)2 (eq 7, R ) i-Pr), while ethyl radical gives
the phosphonate EtP(O)(OMe)2; and radical 5d with (EtO)3P
yields phosphonate 13.) In the chain reactions of both 6a (via
eq 3 and 10) and 11 (via eq 3, 11, and 12), a relatively stable
benzyl radical that is unreactive toward three-coordinate phos-
phorus is trapped by addition to [1.1.1]propellane (1) to give a
reactive radical (5a) able to propagate a free-radical chain
process by oxidative addition to phosphorus.10 These constitute
noVel processes to giVe modest yields of structurally unusual
products, 7 and 12. In fact bis adduct 10-O, which incorporates
two molecules of [1.1.1]propellane (1), results on reaction of
PhCH2OP(OMe)2 11 but not PhCH2P(OMe)2 6a. This indicates
that attack by 3-benzylbicyclo[1.1.1]pent-1-yl radical 5a on the
phosphorus of 6a and subsequent R scission, are more efficient
than attack on PhCH2OP(OMe)2 11 followed by â scission. This
result also is consistent with the trapping of the isopropyl radical
by PhCH2P(OMe)2 6a but not by PhCH2OP(OMe)2 11.
45 was measured (Table 2), along with a similar number, close
to 30, for the analogous reactions of cis- and trans-8. Reactions
3 and 10 are evidently both reasonably efficient.
Equations 3, 11, and 12 account nicely for the reaction of 1
with PhCH2OP(OMe)2 in the radical Arbuzov-insertion process
that provides 12 (reaction 5). Like the reaction of 6a via process
8, reactions 11 and 12 are quite efficient, as indicated by the
measured mean chain length of 30-40. Evidently, both R
scission (step 10b) and â scission (step 12) are rapid when the
relatively stable benzyl radical is formed. The benzyl radical
Additional evidence that the 3-substituted bicyclo[1.1.1]pent-
1-yl radical forms an especially strong bond to phosphorus is
(24) Tables of average bond strengths for these substituents are recorded
in ref 10 a-c,f, and 16.
(25) The addition of PhCH2Br to [1.1.1]propellane (1), presumably
involving attack benzyl radical on [1.1.1]propellane (1), is well documented
although yields were not reported.2a,n
(23) For earlier presentations of figures similar to Figure 1 that represent
the potential reactions of free radicals with three-coordinate phosphorus
molecules, including substitution, oxidation, and free-radical Arbuzov, see
refs 10a,b,c,f, 15, and 16.

1394 J. Am. Chem. Soc., Vol. 119, No. 6, 1997
Dockery and Bentrude
the observed reaction (eq 10, R ) Me2N) at room temperature
of 3-(dimethylamino)bicyclo[1.1.1]pent-1-yl radical 6c (R )
Me2N) with (MeO)2PNMe2 (6c) to displace dimethylamino
radical. This provides substitution-insertion product 7c (R )
Me2N) in a process that requires the phosphorus-carbon bond
of 7c to be stronger than the phosphorus-nitrogen bond broken
in 6c.
process, the â scission of (EtO)4P• at 0 °C has a rate constant17
of only 3 × 102 s-1.) Moreover, 3-benzylbicyclo[1.1.1]pent-
1-yl radical 5a (R ) PhCH2), formed on photolysis of
stoichiometric amounts of BPMDA in the presence of [1.1.1]-
propellane (1), reacted with (MeO)3P, eVen at 0 °C, to yield
small amounts of insertion product 12. This requires â scission
of the even stronger O-Me bond of phosphoranyl radical 22b.
Though this â scission is relatively slow, the methyl radicals
from 22b initiated formation of minor amounts of phosphonate
23 via phosphoranyl radical 22c.
For comparison, the experimentally determined average bond
strength of the phosphorus-carbon bond in Et3P is 62 kcal/
mol24 and that of the phosphorus-nitrogen bond in (Me2N)3P
is 69 kcal/mol.24,26 This difference in bond strengths also is
reflected in the relative rates of R-scission in the phosphoranyl
radical derived from addition (eq 13) of benzyloxy radical to
(EtO)2PEt (18a) and (EtO)2PNBu2-n (18b). For 18a the ratio
of 20 (â scission) to 21 (R scission) is 0.19. The same ratio
(20/21) for 18b is 1.7. On the premise that the weaker bond
would undergo R scission more rapidly, it is apparent that the
P-Et bond is weaker than the P-NBu2 bond. The difference
in the efficiencies of the reactions of n-C5H11P(OMe)2 (6b) and
Me2NP(OMe)2 (6c) with [1.1.1]propellane (1) is likely due to
the difference in bond strengths (i.e., n-C5H11-P < Me2N-P).
Thus, n-C5H11P(OMe)2 (6b) undergoes a much more efficient
reaction with [1.1.1]propellane (1) than does Me2NP(OMe)2 6c.
A tert-butoxy radical from photolysis of di-tert-butyl peroxide
Free radical Arbuzov reactions of (MeO)3P and (EtO)3P with
ethyl radicals, as noted above, do not occur. Normally, only
phenyl radicals carry out such a reaction (eq 2, R ) Ph, R′ )
Me12). Phenyl radicals, which are σ radicals with the odd
electron in an sp2 hybridized orbital, no doubt form relatively
strong bonds to phosphorus in the intermediate phosphoranyl
radical.12d,e Moreover, the latter, based on ESR evidence, is
electronically best represented by structure 24 with the odd
electron in a π* orbital.27 Such a ligand-π10d phosphoranyl
radical may have intrinsic stability greater than that of the more
usual near-trigonal bipyramidal species like 3 and 4 of the
present study. The oxidative addition of 3-ethylbicyclo[1.1.1]-
pent-1-yl radicals (5d) to triethyl phosphite to give phosphoranyl
radical 22a followed by its â scission to generate phosphonate
13, finds precedent in the recently reported reaction of
3-(methoxycarbonyl)bicylo[1.1.1]pent-1-yl radicals with (EtO)3P
at 30 °C to give a 28% yield of 25.2d
displaces n-pentyl radical from 6b to initiate the process. In
the chain reaction represented in steps 3 and 10, the n-pentyl
radical is again readily generated on R scission of phosphoranyl
radical 16b, eVen at 0 °C. By contrast, the reaction of Me2-
NP(OMe)2 (6c) with [1.1.1]propellane (1) requires a higher
temperature (24 °C) and a high flux of dimethylamino radicals
from near-stoichiometric amounts of Me2NNdNNMe2 to give
reasonable quantities of the 3-(dimethylamino)bicyclo[1.1.1]-
pent-1-ylphosphonate (7c) in a process with a short kinetic chain
length at best. The higher temperature, along with dilute
reaction conditions, ultimately allows the R scission of the P-N
bond of phosphoranyl radical 16c to compete with the dominant
formation of polymer observed under normal conditions (0 °C
and higher concentrations of [1.1.1]propellane (1)). Neverthe-
less, the tertiary bicyclo[1.1.1]pent-1-yl radical 5c was able to
displace the dimethylamino radical (eq 10, R ) Me2N). Clearly,
bicyclo[1.1.1]pent-1-yl substituent 5 forms a stronger bond to
three-coordinate phosphorus than do the n-pentyl and dimeth-
ylamino groups.
The above results show the propensity of a 3-substituted
bicyclo[1.1.1]pent-1-yl radical to add oxidatively to three-
coordinate phosphorus to be similar to that of a phenyl radical.
This behavior is a reflection of the high degree of s character
in the bridgehead orbital containing the odd electron which
results from the pyramidal nature of this strained radical.
Indeed, ESR measurements of 13C hyperfine splittings at the
bridgehead carbon of the bicyclo[1.1.1]pent-1-yl radical allow
assignment of 20% 2s character to the orbital bearing the odd
electron and lead to an estimation of 100° for the bridgehead
C-C-C bond angle for this highly pyramidal species.8,9
Bicyclo[1.1.1]pent-1-yl radicals, therefore, differ markedly from
tert-butyl radicals which are estimated by theoretical calcula-
tions28 and ESR measurements29 to be bent only 11-24° from
planarity.27,28 Furthermore, theoretical and experimental evi-
dences have been presented for the increased strength of bonds
formed to the bridgehead carbon of various bicyclo[1.1.1]-
Most dramatic, however, is the ability of the 3-ethylbicyclo-
[1.1.1]pent-1-yl radical (5d, R ) Et), generated under dilute
conditions from near stoichiometric quantities of azobisethane,
to react with (EtO)3P at 24 °C to form 13. This requires the
3-ethylbicyclo[1.1.1]pent-l-yl radical (5d) to bind tightly, and
presumably irreversibly, to phosphorus in phosphoranyl radical
22a so that generation of ethyl radicals via â scission of the
O-Et bond can take place at 24 °C. The higher temperature
(24 °C) facilitates the â scission of 22a. (As a model for this
(26) A somewhat higher average bond strength (and in our view an
unreasonably high one) is calculated using a more recent heat of formation
of the dimethylamino radical. Golden, D.; Solly, R. K.; Gac, N. A.; Benson,
S. W. J. Am. Chem. Soc. 1972, 94, 363.
(27) (a) Davies, A. G.; Maxwell, J. P.; Roberts, B. P. J. Chem. Soc.,
Chem. Commun. 1974, 973. (b) Boekestein, G.; Jansen, E. H. J. M.; Buck,
H. M. J. Chem. Soc., Chem. Commun. 1974, 118.

Free Radical Chain Reactions of [1.1.1]Propellane
J. Am. Chem. Soc., Vol. 119, No. 6, 1997 1395
pentanes.1d,2d,6,30 This is undoubtedly true also of bonds formed
to phosphorus in phosphoranyl radical intermediates and in their
products of R and â scission. These factors, combined with
potentially reduced steric interactions with the phosphorus
center, give the bicyclo[1.1.1]pent-1-yl radical a thermodynamic
and presumably also kinetic propensity toward reaction with
three-coordinate phosphorus that is greater than that of other
alkyl radicals (primary, secondary, or tertiary) and com-
mensurate with that of the phenyl radical.
Scheme 1
Energetics and Kinetics of Reactions of Free Radicals with
Three-Coordinate Phosphorus. The approximate relative
energetics of addition of free radicals to three-coordinate
phosphorus and the alternative R and â scission routes available
to the phosphoranyl radical formed are shown in Figure 1.14,23
A reactant, A, is employed that combines processes 1 and 2,
substitution and free-radical Arbuzov reactions, through a single
phosphoranyl radical intermediate, B. The aVerage P-X bond
strengths for several PX3 are known experimentally:24 X )
MeO, 84 kcal/mol; X ) Ph, 77 kcal/mol; X ) NMe2, 69 kcal/
mol; X ) Et, 62 kcal/mol. This allows the radicals to be ordered
along the energy axis such that the approximate thermodynamic
ease of generation of a given radical (Z•) by its replacement by
a particular attacking radical (X•) is indicated. For the carbon
radicals methyl through benzyl, the ordering comes from relative
radical stabilities and presumed consequent energies of the bonds
formed to phosphorus in A and C, even though average bond
strength values are available only for Me3P (67 kcal/mol) and
Et3P (62 kcal/mol).24 No attempt is made to precisely space
the energies of the various Z• or X•. However, for addition of
methyl radical to (EtO)3P, ∆H° has been determined to be -7
kcal/mol,17 a value used to guide the placement of the methyl
radical (and hence the others relative to it) along the energy
axis with respect to the phosphoranyl radical B. We previously
estimated the highly irreversible oxidative addition of a phenyl
radical to a phosphite to be exothermic by 10-13 kcal/mol.16
The assumption is made that variations in X and Z attached to
B will not greatly affect its energy. Figure 1 then allows one
to compare the thermodynamics of formation of phosphoranyl
radical B by oxidative addition of various X• to A followed by
loss of Z• (R scission) to give C (radical substitution product)
or â scission to give D (radical Arbuzov product). The difficulty
of forming B by oxidative addition of the more stable radicals
to A is readily apparent.
Depicted by Figure 1 are the three classes of reactiVity
mentioned above.14,23 Relatively stable radicals (isopropyl, tert-
butyl, and benzyl) bond weakly or not at all to phosphorus in
B. The failure, shown in this study, of isopropyl to replace
n-pentyl (thermodynamically unfavorable) and benzyl to replace
benzyl (energetically inaccessible B) are consistent with Figure
1. The second class of radicals, dimethylamino, methyl, and
ethyl, form stronger bonds to phosphorus, but bonding is
reVersible and only leads to free radical Arbuzov product when
R′ is benzyl whereupon rapid â scission ensues. Substitution
potentially can occur if the radical to be displaced is lower on
the energy scale than the attacking radical (e.g., dimethylamino
radical should readily replace an alkyl radical.) The replace-
ments of benzyl by isopropyl and ethyl radicals and failure of
isopropyl radical to displace n-pentyl, reported in the present
research, are in accord with Figure 1. Those radicals that form
very strong bonds to phosphorus (alkoxy, phenyl, and 3-sub-
stituted bicyclo[1.1.1]pent-1-yl) constitute a third class, those
that react irreVersibly and give radical-Arbuzov product even
when the alkyl radical formed on â scission is less stable than
benzyl (e.g., ethyl or methyl) and is formed more slowly. The
3-substituted bicyclo[1.1.1]pent-1-yl radical (5) is grouped with
the phenyl radical in Figure 1. These radicals, of course, readily
replace those below them on the energy scale in substitution
processes, as exemplified in this study by the reactions of 5
with 6a-c.
Stereochemistry of Free Radical Displacement-Insertion.
The predominant inversion stereochemistry of the reactions of
cis- and trans-2-benzyl-5-methyl-1,3,2-dioxaphospholanes (cis-
and trans-8) to form the displacement-insertion products cis-
and trans-9 is completely consistent with the known inversion
stereochemistry of other radical substitutions at phosphorus that
were reported earlier by the Bentrude group.31 A rationale for
this outcome is depicted in Scheme 1. Radical 5a is introduced
into trigonal bipyramidal 26 axially in keeping with conclusions
from ESR studies of R scission reactions (the microscopic
reverse).32 The benzyl group is then moved to the axial position
(27) from which it leaves (R scission) to yield product (cis-9).
The overall result is substitution with inversion at phosphorus.
The isomerization of 26 to 27 is of mode 4, as defined by
variable temperature ESR studies3 of the kinetics of such
processes when phosphorus is included in a 1,3-dioxaphos-
pholane ring. Theoretical calculations also support mode 4 low-
barrier processes for at least certain phosphoranyl radicals.33
A
straightforward inline process via 28 also would proceed with
inversion at phosphorus. However, the stereochemical conse-
quences of initial formation of intermediate 26 must be
considered because of the known thermodynamic preference of
such rings to be attached in apical-equatorial fashion3 which
also may be true kinetically.
(31) Bentrude, W. G.; Kahn, W. A.; Murakami, M.; Tan, H.-W. J. Am.
Chem. Soc. 1974, 96, 5566. Nakanishi, A.; Bentrude, W. G. J. Am. Chem.
Soc. 1978, 100, 6271.
(32) Cooper, J. W.; Roberts, B. P. J. Chem Soc., Perkin Trans. 2 1976,
808. Davies, A. G.; Dennis, R. W.; Roberts, B. P. J. Chem. Soc., Perkin
Trans. 2 1974, 1101.
(28) For selected papers on the geometries of unconstrained alkyl radicals,
see: (a) Pacansky, J.; Koch, W.; Miller, M. D. J. Am. Chem. Soc. 1991,
113, 317 and references therein. (b) Paddon-Row, M. N.; Houk, K. N. J.
Am. Chem. Soc. 1981, 103, 5046.
(29) Griller, D.; Ingold, K. U.; Krusic, P. J.; Fischer, H. J. Am. Chem.
Soc. 1978, 100, 6750.
(30) Maillard, B.; Walton, J. C. J. Chem. Soc., Chem. Commun. 1983,
900.
(33) See Gustafson, S. M.; Cramer, C. J. J. Phys. Chem. 1995, 99, 2267.
A double pseudorotation as well as two successive pseudorotations (mode
1) via an electron apical transition state or intermediate are equivalent
stereochemically to a mode 4 permutation of the type shown for 26-27.

1396 J. Am. Chem. Soc., Vol. 119, No. 6, 1997
Dockery and Bentrude
Waters Differential Refractometer detector using Waters µ Porasil (25
× 0.46 cm) and Rainin Microsorb Dyanmax 60A silica columns (25
× 2.14 cm).
Conclusions
The above results demonstrate the great ease of reaction of
3-substituted bicyclo[1.1.1]pentyl radicals (5) with three coor-
dinate phosphorus. E.g., free radical Arbuzov reactions with
(EtO)3P, that only occur with the very reactive phenyl radical,
are found for 5. The replacement of a primary radical, the
n-pentyl, by 5 further shows the propensity of these highly
pyramidal tertiary radicals (5) to bond to three-coordinate
phosphorus. Moreover, [1.1.1] propellane (1) is able to trap
free radicals (eq 3) to give 3-substituted bicyclo[1.1.1]pentyl
radicals (5) that participate in chain carrying reactions 4 and 5
as part of reasonably efficient radical chain processes (chain
lengths 30-50). Reactions 4 and 5 involve the key R scission
step of radical substitution (eq 10b) and â scission step of the
free radical Arbuzov process (eq 12) and yield novel products,
e.g., 7 and 12, featuring the insertion of the bicyclo[1.1.1]-
pentane structure. These chain processes are enabled by the
ability of 5 to form strong bonds to three-coordinate phosphorus
molecules (eqs 10a and 11a). The propensities of various
radicals, including 5, to bond to three-coordinate phosphorus
and promote radical substitution or radical-Arbuzov reactions
are summarized in Figure 1 in which three reactivity classes of
radicals are depicted.
[1.1.1]Propellane (1) was prepared in ether according to the
procedure of Szeimies et al.38 from the reaction of 1,1-dibromo-2,2-
bis(chloromethyl)cyclopropane (ca. 20 mmol scale) with methyllithium
(ca. 40 mmol). Volatile products, including [1.1.1]propellane (1), were
distilled at 30-40 mmHg (water aspirator) into a flask cooled to -78
°C in which crude [1.1.1]propellane (1) was stored prior to purification.
Neat [1.1.1]propellane (1), employed in preparing most of the reaction
solutions, was isolated by preparative GLC on 15% Apiezon L or 15%
SE-30 on Chromsorb W utilizing a 25 ft × 1/4 in. column operated at
ambient temperature. A 3-5 mL portion of the ether solution of crude
[1.1.1]propellane, warmed to warm temperature, was injected directly
on column via a 5.00 mL Hamilton gastight Teflon luerlock syringe.
Neat [1.1.1]propellane (10-30 mg/injection) was collected at -78 °C.
General Procedure for the Reactions of Three-Coordinate
Phosphorus Molecules with [1.1.1]Propellane. All operations with
the exception of weighings were done in a glove bag under a N2
atmosphere. Liquid transfers were made by syringe techniques. Thus,
1-2 mmol by weight of the three-coordinate phosphorus compound
were transferred to a vial to which was added a solution of 0.05 mmol
of initiator in 100 µL of degassed C6D6. By way of a Hamilton 250
µL Teflon plunger luerlock syringe, 30-50 µL of neat [1.1.1]propellane
(30-50 µL), isolated by preparative GLC, was added to a nitrogen-
flushed quartz test tube (12 mm × 40 mm), fitted with a rubber septum,
and maintained at -78 °C under a N2 sweep. For reactions for which
yields (based on [1.1.1]propellane) were determined, the quartz test
tube was immediately reweighed to determine the amount of [1.1.1]-
propellane added. Quick transfers and weighings helped to minimize
introduction of water by condensation as well as thermal oligomerization
of the [1.1.1]propellane. The solution of the three-coordinate phos-
phorus compound and initiator was transferred from the vial to the
quartz test tube maintained at -78 °C. The solution was warmed to
melting, agitated to insure thorough mixing, and irradiated for 1 h at 0
°C (450 W medium pressure Hg lamp, filtered through a Kimax sleeve).
Experimental Section
Materials. Ether (Et2O) and tetrahydrofuran (THF) were distilled
from sodium/benzophenone under nitrogen. Benzene (Baker Photrex)
was used as received. Acetonitrile was distilled over calcium hydride
under nitrogen. Benzene-d6 (C6D6) was degassed prior to use by
purging with argon. Triethyl phosphite and trimethyl phosphite were
distilled from sodium prior to use. Cis- and trans-7, bis(phenylmethyl)-
diazene,34 azobisethane (Et2N2),35 azobisisopropane (i-Pr2N2),36 and
tetramethyltetrazene (TMT)36 were prepared by literature procedures.
Et2NPCl2, i-Pr2NPCl2, (MeO)2PNEt2, and i-Pr2NP(OMe)2 were prepared
by standard methods.37 In the Supporting Information Available may
be found details of the nearly routine preparations of [2,2′-2H]bis-
(phenylmethyl)diazene, 6a, 6b, 6c, 11, and 15b. Unless otherwise
noted, all distillations were performed using a short-path apparatus.
Physical Methods. Melting points were obtained on a Thomas
Hoover Capillary Melting Point apparatus and are uncorrected. 1H,
2H, 13C, and 31P NMR spectra were obtained on Varian XL-300 and
Unity 300 NMR spectrometers. Some 31P NMR spectra were taken
on a Varian FT-80 NMR. The 125 MHz 13C NMR spectra were
measured on a Varian VXR-500 NMR spectrometer. Chemical shifts
(δ) are recorded in parts per million with respect to tetramethylsilane
General Procedure for Oxidation of Reaction Products. Fol-
lowing irradiation the reaction mixture was transferred in a N2-swept
glove bag to a round bottom flask, diluted to 10 mL with benzene,
fitted with a stir bar, a rubber septum, and nitrogen inlet, and cooled
to 0 °C. Via syringe, 1.1 equiv of tert-butyl hydroperoxide (3.0 M in
2,2,4-trimethylpentane) was added. The solution was stirred for 1 h
at 0 °C, and the solvent was removed in vacuo.
General Procedure for Sulfur-Oxidation of Reaction Products.
To a round bottom flask were added 15 mL of degassed benzene, 3
equiv of S8, and 100 µL of NEt3. The flask was cooled to 0 °C under
a sweep of nitrogen with stirring. The crude photolysis mixture, diluted
with benzene to approximately 1 mL, was added to the above stirred
solution via a cannula. After 1 h, the reaction mixture was filtered,
and the solvent was removed in vacuo.
(0.00) with the solvent as an internal standard for the H, H, and 13C
NMR spectra: (C6HD5 (1H) 7.15, (13C) 128.0; CHCl3 (1H) 7.24, (13C)
77.0; C6DH5 (2H) 7.15). A capillary of external 85% H3PO4 in C6D6
or CDCl3 served as reference for the 31P NMR spectra. GC-EIMS (70
eV) analyses were done on a Hewlett-Packard 5971A Mass Selective
Detector utilizing a 5890 Series II Gas Chromatograph equipped with
a 30 m X 0.25 mm HP-5 on fused silica capillary column. Other low
resolution EIMS (70 eV) along with HRMS (EI, 70 eV) data were
obtained on a Finnegan MAT 95 gas chromatograph/mass spectrometer.
Intensities (parentheses) are reported as a percent of the base peak
intensity. The molecular ion is designated as [M]+. Because of the
small scale of the reactions, HRMS rather than elemental analyses were
obtained for new compounds.
1
2
General Procedure for Determination of Yields by 31P NMR
(Table 5). The residue was diluted to 2.00 mL in a volumetric flask
with C6D6. Unreacted and product three-coordinate phosphorus
compounds were converted to the oxides as above. Yields, based on
starting three coordinate phosphorus-containing material consumed,
were determined by quantitative 31P NMR vs a 0.317 M solution of
trimethyl phosphate in C6D6 inserted into the NMR tube as an external
standard (sealed in a capillary tube). Relative peak areas for proton-
coupled spectra were measured under conditions of 60 s delay times.
A series of phosphorus compounds, over a range of known concentra-
tions, were employed to obtain a sensitivity factor of 0.42 ( 0.03 mmol/
mL which was multiplied by the area ratio and the number of milliliters
of solution to obtain mmol of each phosphorus compound. The yields
are considered to be conservative, as some of the starting phosphorus
compound is inevitably lost in transfer, resulting in lowered recovery
of unreacted starting material. The oxidations may be less than
quantitative as well. Avoidance of an internal standard allowed products
to be isolated more readily from the same solution used for quantitation
HPLC separations and product isolations (isocratic) were done with
a Waters 590 solvent delivery system and either an ISCO V4 UV or
(34) Bandlish, B. K.; Garner, A. W.; Hodges, M. L.; Timberlake, J. W.
J. Am. Chem. Soc. 1975, 97, 5856.
(35) Leitch, L. C.; Renaud, R. Can. J. Chem. 1954, 32, 545.
(36) McBride, W. R.; Kruse, H. W. J. Am. Chem. Soc. 1957, 79, 572.
(37) Nakazawa, H.; Kadoi, Y.; Miyoshi, K. Organometallics 1989, 8,
2851. Phosphoramidous acid esters: Mag, M.; Engels, J. W. Nucleosides
and Nucleotides 1988, 7, 725. Arbuzov, B. A.; Yarmukhametova, D. Kh.
Dokl. Akad. Nauk SSSR 1955, 101, 675. 31P NMR chemical shifts: Quin,
L. D.; Gordon, M. D.; Lee, S. O. Org. Magn. Reson. 1974, 6, 503 and
references therein.
(38) Belzner, J.; Bunz, U.; Semmler, K.; Szeimies, G.; Opitz, K.; Schlu¨ter,
A.-D. Chem. Ber. 1989, 122, 397.

Free Radical Chain Reactions of [1.1.1]Propellane
J. Am. Chem. Soc., Vol. 119, No. 6, 1997 1397
by 31P NMR. Isolated yields, given below and in Table 5, are also
based on converted three-coordinate reactant as determined by 31P
NMR.
to two HPLC (on silica) purifications (EtOAc, UV detection) to give
6 mg of the insertion product 1 as a colorless oil with NMR and MS
spectra identical to those for 12 reported above. From the crude reaction
mixture, prior to oxidation, the resonances for dimethyl 1-(3-(phenyl-
methyl)-bicyclo[1.1.1]pentyl)phosphonite (7a) could be readily as-
signed: 31P NMR (121.4 MHz, C6D6) δ 172.49; 1H NMR (299.9 MHz,
For example, a solution of bis(phenylmethyl)diazene (BPMDA) (4.8
mg, 0.023 mmol), benzyl dimethyl phosphite, 11 (113 mg, 0.57 mmol),
excess [1.1.1]propellane (60-70 µL, approx. 1 mmol), and 100 µL of
C6D6 was irradiated for 1 h at 0 °C.39 The photolysis mixture was
diluted to 2.00 mL with C6D6. The 31P NMR spectrum showed the
presence of two product peaks, identified as the mono- and bis-bicylo-
[1.1.1]pentylphosphonate adducts 12 and 10-O, in addition to starting
phosphite and a small amount of benzyl phosphate, resulting from
oxidation. Yields, determined by quantitative 31P NMR, were 45%
(0.20 mmol) for the mono adduct 12 and 11% (0.048 mmol) for the
bisadduct 10-O, based on disappearance of starting phosphite. Product
isolation by flash chromatography (100% ethyl acetate), followed by
HPLC (100% ethyl acetate), gave 29.4 mg (0.11 mmol, 25% isolated
yield) of 12 and 11.1 mg (0.033 mmol, 8% isolated yield) of 10-O.
Product dimethyl 1-(3-(phenylmethyl)bicyclo[1.1.1]pentyl)phosphonate
(12): mp 45-46 °C; 31P NMR (121.4 MHz, CDCl3) δ 22.51; 31P NMR
(121.4 MHz, C6D6) δ 22.71; 1H NMR (299.9 MHz, CDCl3) δ 1.91 (d,
3
C6D6) δ 1.72 (d, 6 H, PCCH2C, JHP ) 0.9 Hz), 2.57 (s, 2 H, CCH2-
3
Ph), 3.40 (d, 6 H, OCH3, JHP ) 11.0 Hz), aromatic protons obscured
by overlap.
Reaction of trans- and cis-8 with [1.1.1]Propellane. The reaction
mixture (e.g., from trans- and cis-8 (71:29 trans:cis starting ratio by
31P NMR) (240.5 mg, 1.23 mmol), BPMDA (5.6 mg, 0.03 mmol),
[1.1.1]propellane (∼0.5 mmol), and 110 µL of C6D6) was treated with
sulfur (see Table 5 for 31P yield data) to give the products cis- and
trans-9-S in a ratio of 62:38 cis:trans (by 31P NMR). Starting material
was recovered as the sulfides trans- and cis-8-S (67:33 trans:cis by
GLC). The sulfides were then passed through silica gel (methylene
chloride), followed by two HPLC purifications (hexanes:ethyl acetate,
90:10). A difficult separation of the isomers gave approximately 2
mg of pure 9-S in 90/10 cis/trans ratio (31P NMR and GLC) and
approximately 1 mg of the pure trans isomer, trans-9-S. These samples
yielded good 31P and 1H NMR data, while 13C NMR data were obtained
from fractions containing 9 mg of the cis and trans isomers in a 3:1
ratio (13C NMR). cis-9 31P (121.4 MHz, C6D6) δ 183.26 (crude reaction
mixture). trans-9 31P (121.4 MHz, C6D6) δ 177.65 (crude reaction
3
5
6 H, CCH2C, JHP ) 1.1 Hz), 2.74 (d, 2 H, CCH2Ph, JHP ) 1.4 Hz),
3
1
3.69 (d, 6 H, OCH3, JHP ) 10.7 Hz), 7.03-7.34 (m, 5 H); H NMR
3
(299.9 MHz, C6D6) δ 1.81 (d, 6 H, CCH2C, JHP ) 1.0 Hz), 2.42 (d,
5
3
2 H, CCH2Ph, JHP ) 1.2 Hz), 3.40 (d, 6 H, OCH3, JHP ) 10.5 Hz),
6.82-7.20 (m, 5 H). 13C NMR (75.4 MHz, C6D6) δ 32.80 (d, PCCH2C,
1JCP ) 161.0 Hz), 39.50 (d, CCH2Ph, JCP ) 26.2 Hz), 44.60 (δ,
4
1
mixture). cis-9-S 31P NMR (121.4 MHz, C6D6) δ 106.86; H NMR
PCCH2C, 3JCP ) 35.5 Hz), 51.24 (d, CCH2C, 2JCP ) 2.4 Hz), 51.58 (d,
OCH3, 2JCP ) 6.1 Hz), 126.33, 128.58, 129.08, 138.57 (ipso-Ph); GC-
EIMS (70 eV) m/z (rel intensity) 266 [M]+ (36), 265 [M - 1]+ (15),
251 [M - CH3]+ (32), 189 [M - Ph]+ (60), 175 [M - PhCH2]+ (6),
156 (39), 155 [M - 111]+ (100), 141 (51), 115 (29), 109 [(CH3O)2-
PO]+ (21), 91 (49), 79 (31), 77 (9), 65 (24). C14H19O3P: HRMS (calcd)
266.1072, (obsd) 266.1073. Product dimethyl 3-[3′-(phenylmethyl)-
1,1′-bicyclo[1.1.1]pentyl]phosphonate (10-O): mp 106-108 °C; 31P
3
(299.9 MHz, C6D6) δ 0.58 (d, 3 H, CHCH3, JHH ) 6.2 Hz), 1.72 (d,
3
5
6 H, CCH2C, JHP ) 1.2 Hz), 2.37 (d, 2 H, CCH2Ph, JHP ) 1.5 Hz),
2.88 (m, 1 H), 3.56 (m, 1 H), 3.97-4.06 (m, 1 H), 6.78-7.44 (m, 5 H,
C6H5); 13C NMR (125.7 MHz, C6D6) δ 17.95 (d, CHCH3, JCP ) 7.9
3
4
1
Hz), 39.25 (d, CCH2Ph, JCP ) 26.9 Hz), 39.85 (d, PCCH2C, JCP
)
3
96.1 Hz), 43.55 (d, PCCH2C, JCP ) 39.3 Hz), 51.33 (d, PCCH2C,
2JCP ) 2.5 Hz), 72.71 (d, OCH2, JCP ) 2.7 Hz), 74.50 (d, OCHCH3,
2
2JCP ) 1.5 Hz), 138.51 (ipso), remaining phenyl resonances masked
by C6D6 triplet (13C NMR spectrum of 3:1 cis:trans mixture); GC-
EIMS (70 eV) m/z (rel intensity) 294 [M]+ (1), 261 [M - SH]+ (11),
203 [M - PhCH2]+ (33), 157 [3-phenylmethyl bicyclo[1.1.1]pentyl]+
(69), 156 [M - 138]+ (100), 155 (44), 145 (19), 141 (75), 105 (24),
91 (75), 77 [Ph]+ (13), 65 (49), 41 (48); C15H19O2PS: HRMS (90/10
cis/trans mixture) (calcd) 294.0844, (obsd) 294.0836. trans-9-S: 31P
1
NMR (121.4 MHz, C6D6): δ 22.46; H NMR (299.9 MHz, C6D6) δ
3
1.27 (s, 6 H, CCCH2C), 1.84 (d, 6 H, PCCH2C, JHP ) 1.2 Hz), 2.55
3
(s, 2 H, CCH2Ph), 3.40 (d, 6 H, OCH3, JHP ) 10.5 Hz), 6.86-7.44
(m, 5 H); 13C NMR (125.7 MHz, C6D6): δ 31.31 (d, PCCH2C, 1JCP
)
160.0 Hz), 39.13 (s, CCH2CCH2Ph), 39.45 (s, CCH2Ph), 39.68 (d,
CCH2CC, 4JCP ) 29.9 Hz), 44.13 (d, CCH2CC, 3JCP ) 32.6 Hz), 48.83,
2
2
1
50.35 (d, PCCH2C, JCP ) 1.9 Hz), 51.87 (d, OCH3, JCP ) 6.5 Hz),
126.16, 128.47, 129.20, 139.68 (s, ipso-Ph); GC-EIMS (70 eV) m/z
(rel intensity) 331 [M - 1]+ (1), 317 [M - CH3]+ (2), 255 [M - Ph]+
(2), 241 [M - PhCH2]+ (4), 177 (12), 149 (23), 131 (37), 115 (17), 91
[PhCH2]+ (100), 77 [Ph]+ (9); C19H24O3P [M - 1]+: HRMS [M -
1]+ (calcd) 331.1463, (obsd) 331.1460.
NMR (121.4 MHz, C6D6) δ 106.43; H NMR (299.9 MHz, C6D6) δ
3
3
0.80 (d, 3 H, CHCH3, JHH ) 6.2 Hz), 1.72 (d, 6 H, PCCH2C, JHP
)
5
1.3 Hz), 2.37 (d, 2 H, CCH2Ph, JHP ) 1.2 Hz), 3.24-3.42 (m, 2 H),
3.56 (m, 1 H), 3.64-3.78 (m, 1 H), 6.80-7.44 (m, 5 H, C6H5); 13C
3
NMR (125.7 MHz, C6D6) δ 18.57 (d, CHCH3, JCP ) 6.1 Hz), 39.28
4
1
(d, CCH2Ph, JCP ) 26.9 Hz), 39.89 (d, PCCH2C, JCP ) 96.3 Hz),
The same procedures were used to determine the 31P yields of the
reactions of 6a, 6b, 6c, 11, and (EtO)3P. Only reactant quantities and
oxidant (tert-BuOOH or S8) are given below. Product isolation was
by flash chromatography on silica gel under nitrogen followed by HPLC
on silica gel with EtOAc as elutant for both chromatographies unless
otherwise specified. UV or RI detection was used in the isolations by
HPLC.
Dimethyl 1-(3-(Phenylmethyl)bicyclo[1.1.1]pentyl)phosphonate
(12) from Reaction of 11 with [1.1.1]Propellane. The reaction
mixture from the reaction of excess 11 (276.2 mg, 1.38 mmol), BPMDA
(5 mg, 0.02 mmol), [1.1.1]propellane (∼0.5 mmol), and 110 µL of
C6D6) gave 45 mg (0.17 mmol, 45% isolated yield based on conversion
of 11, 61% crude yield, 31P NMR) of product 12 as a colorless white
solid. (Little or no 10-O was detected by 31P NMR.) Spectral data
were identical to those for 12 recorded above.
3
2
43.17 (d, PCCH2C, JCP ) 39.1 Hz), 51.33 (d, PCCH2C, JCP ) 2.5
2
2
Hz), 71.82 (d, OCH2, JCP ) 2.1 Hz), 76.16 (d, OCHCH3, JCP ) 2.3
Hz), 138.56 (ipso), remaining phenyl resonances masked by C6D6 triplet.
Reaction of Dimethyl n-Pentylphosphonite (6b) with [1.1.1]-
Propellane. The solution of products from reaction of 6b (109 mg,
0.665 mmol), [1.1.1]propellane (∼0.5 mmol), di-tert-butyl peroxide (3
µL, 2 mg, 0.02 mmol), in 100 µL of C6D6 was oxidized with tert-butyl
hydroperoxide, purified by flash chromatography on silica gel (EtOAc),
and separated by HPLC on silica gel (EtOAc) to give 13 mg (0.053
mmol) of the insertion product 7b-O as a colorless oil (12% yield,
based on 6b consumed (31P NMR). Product dimethyl 1-(3-n-
pentylbicyclo[1.1.1]pentyl)phosphonite 7b: 31P NMR (121.4 MHz,
C6D6) δ 171.98 (crude reaction mixture). Product dimethyl 1-(3-n-
pentylbicyclo[1.1.1]pentyl)phosphonate (7b-O): 31P NMR (121.4 MHz,
1
C6D6) δ 22.20; H NMR (299.9 MHz, C6D6) δ 0.86 (t, 3 H, CH2CH3,
3JHH ) 7.1 Hz), 0.96-1.26 (m, 8 H, (CH2)4), 1.87 (d, 6 H, PCCH2C,
3JHP ) 1.2 Hz), 3.44 (d, 6 H, OCH3, 3JHP ) 10.6 Hz); 13C NMR (75.4
MHz, C6D6) δ 14.21, 22.90, 25.96, 32.17 (s), 32.17 (d, PCCH2C, 1JCP
Reaction of Dimethyl Benzylphosphonite (6a) with [1.1.1]Pro-
pellane. The product mixture (from 6a (244 mg, 1.33 mmol), BPMDA
(9 mg, 0.04 mmol), [1.1.1]propellane (0.5 mmol), and 200 µL of C6D6)
was treated with tert-BuOOH and purified by flash chromatography
on silica gel (EtOAc). A portion of this material was then subjected
4
) 159.8 Hz), 32.26 (d, PCCH2CCH2, JCP ) 25.8 Hz), 44.83 (d,
3
2
PCCH2C, JCP ) 33.1 Hz), 51.41 (d, PCCH2C, JCP ) 2.2 Hz), 51.86
2
(d, OCH3, JCP ) 6.5 Hz); GC-EIMS (70 eV) m/z (rel intensity) 246
[M]+ (0.5), 245 [M - 1]+ (4), 231 [M - CH3]+ (1), 217 [M - Et]+
(3), 203 [M - n-Pr]+ (9), 189 [M - n-Bu]+ (100), 175 [M - n-C5H11]+
(13), 157 (22), 137 [3-n-pentylbicyclo[1.1.1]pentyl]+ (6), 109 [(CH3O)2-
PO]+ (23), 107 (19), 94 (17), 93 (38), 79 (50); C12H22O3P [M - 1]+:
HRMS [M - l]+ (calcd) 245.1306, (obsd) 245.1302.
(39) Use of Kimax filtered light avoids the photo-Arbuzov isomerization
of benzyl phosphites studied in this laboratory. Omelanzcuk, J.; Sopchik,
A. E.; Lee, S.-G.; Akutagawa, K.; Cairns, S. M.; Bentrude, W. G. J. Am.
Chem. Soc. 1988, 110, 6908. Cairns, S. M.; Bentrude, W. G. Tetrahedron
Lett. 1989, 30, 1025; Bentrude, W. G.; Mullah, K. B. J. Org. Chem. 1991,
56, 7218.

1398 J. Am. Chem. Soc., Vol. 119, No. 6, 1997
Dockery and Bentrude
2
Reaction of Dimethyl Dimethylphosphoramidite (6c) with [1.1.1]-
Propellane. Under the above conditions (0.5 mmol of 6c, 1 mmol of
[1.1.1]propellane (1), 0.1 mmol of tetramethyltetrazene at 0 °C), only
trace amounts of product 7c resulted (31P NMR). Alternatively, the
photolysis was done at 24 °C with larger amounts of C6D6: 6c (106
mg, 0.774 mmol), tetramethyltetrazene (70 µL, 0.6 mmol), [1.1.1]-
propellane (∼1 mmol), in 500 µL of C6D6. The reaction products were
treated with S8. Separation of the sulfides by flash chromatography
on silica gel (elution with hexanes:ethyl acetate, 1:1, then ethyl acetate),
monitored by GLC, gave thiophosphonate 7c-S (2 mg, 85% pure by
GLC) as a colorless oil. (Coevaporation of the residual EtOAc with
benzene was necessary to prevent decomposition of 7c-S): Product
dimethyl 1-(3-dimethylaminobicyclo[1.1.1]pentyl)phosphonite (7c): 31P
NMR (121.4 MHz, C6D6) δ 171.88 (crude reaction mixture). Product
dimethyl 1-(3-(dimethylamino)bicyclo[1.1.1]pentyl)phosphonothioate
(7c-S) (sulfide of 7c): 31P NMR (121.4 MHz, C6D6) δ 93.07; 1H NMR
(299.9 MHz, C6D6) δ 1.91 (s, 6 H, N(CH3)2), 1.94 (d, 6 H, PCCH2C,
33.2 Hz), 50.57 (d, PCCH2C, JCP ) 2.1 Hz), 61.26 (d, OCH2CH3,
2JCP ) 6.2 Hz); GC-EIMS (70 eV) m/z (rel intensity) 232 [M]+ (0.4),
231 [M - 1]+ (3), 217 [M - CH3]+ (12), 203 [M - Et]+ (7), 189 [M
- 43]+ (14), 175 [M - 57]+ (24), 161 (66), 137 [(EtO)2PO]+ (2), 111
(21), 95 [3-ethylbicyclo[1.1.1]pentyl]+ (24), 94 (60), 93 (100), 91 (19),
81 (25), 79 (62), 77 (19), 65 (21); C11H20O3P [M - 1]+: HRMS [M -
l]+ (calcd) 231.1150, (obsd) 231.1146.
Reaction of Ethyl Radicals with Benzyl Dimethyl Phosphite. A
5.00 mL solution of benzyl dimethyl phosphite (11) (233 mg, 1.17
mmol), azobisethane (97 mg, 1.13 mmol), and tri-n-propyl phosphate
(73 mg, 0.326 mmol) as internal standard in C6D6 was prepared under
nitrogen, and portions were added to two 12 mm Pyrex NMR tubes
(∼0.7 mL each) that were then degassed at 10-3 mmHg with 4 freeze-
pump-thaw cycles, flame sealed, and irradiated for 14 h (medium-
pressure 450 W Hg lamp, Kimax filter sleeve). The 31P NMR spectrum
showed the presence of dimethyl ethylphosphonate, confirmed by GC
and GC-EIMS, in addition to starting phosphite and phosphate internal
standard. By quantitative 31P NMR (proton-coupled, 60 s repetition
rate), the yield of EtP(O)(OEt)2 was determined to be 90%, based on
18% conversion of starting material. No phosphonate was formed in
a nonirradiated control reaction. The 1H NMR spectrum showed
evidence for the formation of n-butane, ethane, and ethylene.
Reaction of Ethyl radicals with Triethyl Phosphite and Trimethyl
Phosphite. Analogously, aliquots of a stock 5.00 mL solution of 1.37
mmol of the appropriate three-coordinate phosphorus compound
(triethyl phosphite or trimethyl phosphite), 1.14 mmol Et2N2, and 0.283
mmol tri-n-propyl phosphate, in C6D6 were irradiated for 12 h. The
31P NMR spectrum showed the presence of only the starting phosphite
and the phosphate internal standard and in the initial ratio.
Reaction of Isopropyl Radicals with Benzyl Dimethyl Phosphite
(11). Irradiation in completely analogous fashion of aliquots of a 5.00
mL solution of benzyl dimethyl phosphite (11) (352 mg, 1.76 mmol),
azobisisopropane (157 mg, 1.40 mmol), and tri-n-propyl phosphate (129
mg, 0.576 mmol) in C6D6 for 12 h gave no evidence by 31P NMR
spectroscopy for the formation of a phosphonate or consumption of
starting phosphite.
Reaction of Isopropyl Radicals with Dimethyl n-Pentylphospho-
nite (6b). Similarly, aliquots of a 5.00 mL solution of dimethyl
n-pentylphosphonite (6b) (157.5 mg, 0.960 mmol, 0.192 M), azobis-
isopropane (151.4 mg, 1.32 mmol, 0.264 M), and tri-n-propyl phosphate
(84.2 mg, 0.76 mmol, 0.0752 M) in C6H6 were irradiated for 14 h
(medium-pressure 450 W Hg lamp, Kimax filter sleeve). The 31P NMR
spectrum showed only the presence of the starting n-pentylphosphonite
(6b) and the phosphate internal standard in the initial ratio.
3
3JHP ) 1.0 Hz), 3.42 (d, 6 H, OCH3, JHP ) 13.4 Hz); GC-EIMS (70
eV) m/z (rel intensity) 235 [M]+ (0.2), 234 [M - 1]+ (1), 220 [M -
CH3]+ (5), 202 [M - SH]+ (29), 188 (11), 125 [(MeO)2PS]+ (7), 110
[3-(dimethylamino)bicyclo[1.1.1]pentyl]+ (100), 108 (22), 70 (20); C9-
H18NO2PS: HRMS (calcd) 235.0796, (obsd) 235.0779.
Reaction of 3-(Phenylmethyl)bicyclo[1.1.1]pent-1-yl Radicals
from BPMDA and [1.1.1]Propellane with Trimethyl Phosphite.
Formation of 12 and 10-O. A rubber septum fitted Pyrex tube was
charged with a 30 mL of a solution of crude [1.1.1]propellane in ether
(∼1-2% by weight, 4-6 mmol [1.1.1]propellane, not purified by GC
isolation), 178 mg (0.85 mmol) of BPMDA, and 5.0 mL (4.8 g, 38
mmol) of P(OMe)3. The solution, sparged with nitrogen from a syringe
needle, was irradiated at 0 oC for 1.5 h with light from a 450 W medium
pressure Hg lamp. Volatile materials were removed in vacuo. 31P
NMR and GC-EIMS analysis showed the presence of several phos-
phonate products, including adducts 12 and 10-O. (The major product
was bibenzyl (GC and GC-EIMS).) The residue from solvent removal
was purified by flash chromatography followed by HPLC (25 × 0.46
cm column) to give 2 mg (0.008 mmol, 1% yield based on BPMDA)
of phosphonate 12 (85% pure by GC and 1H NMR): identification by
comparison to 31P NMR and 1H NMR spectral data, GC-retention time,
and GC-EIMS data of authentic samples of 12 isolated from reactions
of 6a and 11 with [1.1.1]propellane (see above). Also identified in
the crude reaction mixture (GC-EIMS) were 10b and presumed dimethyl
1-(3-methylbicyclo[1.1.1]pentyl)phosphonate (23): GC-EIMS (70 eV)
m/z (rel intensity) 190 [M]+ (3), 189 [M - 1]+ (35), 175 [M - CH3]+
(25), 157 (12), 143 (13), 110 (23), 81 [3-methylbicyclo[1.1.1]pentyl]+
(15), 80 (41), 79 (100).
Reaction of Isopropyl Radicals with Dimethyl Benzylphosphonite
(6a). As above, aliquots of a 5.00 mL solution of dimethyl ben-
zylphosphonite (5a) (231.8 mg, 1.26 mmol, 0.252 M), azobisisopropane
(129.7 mg, 1.14 mmol, 0.228 M), and tri-n-propyl phosphate (73.8 mg,
0.330 mmol, 0.0659 M) in C6H6 were irradiated in Pyrex NMR tubes
for 14 h (medium-pressure 450 W Hg lamp, Kimax sleeve). A new
resonance at δ 191.79 was observed in the 31P NMR spectrum,
consistent with the formation of an alkylphosphonite. The 31P NMR
yield of the presumed dimethyl isopropylphosphonite (15a) was 0.32
mmol (36% yield based on 69% conversion of starting 6a).
Reaction of Triethyl Phosphite with [1.1.1]Propellane. Under the
above conditions for the reactions of 6a, 6b, and 11 with [1.1.1]-
propellane (irradiation of 1-2 mmol of P(OEt)3, 0.5 mmol of [1.1.1]-
propellane, and 0.1 mmol Et2N2 initiator, in 100 µL of C6D6, 0 °C), a
waxy polymer was formed. A relatively dilute solution in a serum-
capped tube with significantly more initiator (3.0 mL of benzene, neat
[1.1.1]propellane (60-70 µL, ∼1 mmol), P(OEt)3 (1.97 g, 11.9 mmol),
and Et2N2 (70 µL, 57 mg, 0.66 mmol) (total solution volume ) 5.2
mL) was irradiated at 24 °C (450 W medium pressure Hg lamp filtered
through a Pyrex sleeve). The progress of the reaction was followed
by 31P NMR and GC analysis of 300 and 500 µL aliquots of solution
withdrawn after 1.5 and 12 h of irradiation, respectively. Solvent was
removed from the remaining solution in vacuo. The residue was diluted
to 2.00 mL in a volumetric flask. The amount of phosphonate 13 in
the crude mixture was determined by quantitative 31P NMR to be 0.15
mmol (corrected for the 0.800 µL of solution removed from the original
5.2 mL reaction mixture), ca. 15% yield based on [1.1.1]propellane.
Product purification by flash chromatography (EtOAc) followed by
HPLC (EtOAc) yielded 14.5 mg (0.063 mmol, ca. 6% yield based on
[1.1.1]propellane) of diethyl 1-(3-ethylbicyclo[1.1.1]pentyl)phosphonate
(13) as a colorless oil: 31P NMR (121.4 MHz, C6D6) δ 20.11; 1H NMR
(299.9 MHz, C6D6) δ 0.63 (t, 3 H, CCH2CH3, 3JHH ) 7.5 Hz), 1.08 (t,
6 H, OCH2CH3, 3JHH ) 7.1 Hz), 1.17 (dq, 2 H, CCH2CH3, 4JHP ) 0.7
Reaction of Ethyl Radicals with Dimethyl Benzylphosphonite
(6a). A nitrogen flushed solution of dimethyl benzylphosphonite (5a)
(198 mg, 1.08 mmol), azobisethane (38 mg, 0.44 mmol), in 0.6 mL of
benzene in a 5 mm Pyrex NMR, capped with a septum and wrapped
with parafilm, was irradiated at 24 C for 12 h as above. 31P NMR
o
(32.2 MHz) showed only the resonance of starting 5a (δ 180.5) and a
single new resonance (δ 188.1) assigned to dimethyl ethylphosphonite
(15b) (δ 188.2 from a sample of 15b prepared independently from
methanolysis of EtP(NEt2)2,37 lit. 15b,37 δ 188.3.); area ratio, 6a:15b,
87:13. The yield of 15b was not quantitated.
Labeling Studies Using [2,2′-2H]Bis(phenylmethyl)diazene (BP-
MDA-d2) as an Initiator. The general procedures outlined above for
reaction of unlabeled BPMDA with 6a, trans- and cis-8, and 11 were
followed. The amount of 2H incorporation in the products was
determined from the percentage increase in product [M + 1]+ peaks
relative to natural abundance [M + 1]+ reference peaks observed using
unlabeled initiator. The intensities of the [M + 1]+ peaks were
expressed as a percentage of the molecular ion [M]+. Each of the
3
3
Hz, JHH ) 7.4 Hz), 1.85 (d, 6 H, PCCH2C, JHP ) 1.2 Hz), 3.97 (m,
4 H, OCH2CH3); 13C NMR (75.4 MHz, C6D6) δ 10.04 (s, CCH2CH3),
16.72 (d, OCH2CH3, 3JCP ) 5.2 Hz), 25.23 (d, CCH2CH3, 4JCP ) 25.9
Hz), 32.37 (d, PCCH2C, 1JCP ) 160.8 Hz), 45.27 (d, PCCH2C, 3JCP
)

Free Radical Chain Reactions of [1.1.1]Propellane
J. Am. Chem. Soc., Vol. 119, No. 6, 1997 1399
experiments was run twice, i.e., with two different sources of unlabeled
BPMDA. The chain lengths were determined by multiplication of the
peak ratios of unlabeled to labeled product by 0.75 (75% 2H incorpora-
tion in the initiator). GC-EIMS data showed no evidence for 2H
incorporation in recovered starting material analyzed after oxidation
to the oxide or sulfide (6a-O), trans- and cis-8-S)). The GC-EIMS
data used to determine the reaction kinetic chain lengths are contained
in Tables 10-12).
spectrometer were supplied by National Science Foundation
Grant CHE-9002690 and the University of Utah Institutional
Funds Committee.
Supporting Information Available: Tables containing
EIMS data for the deuterium labeling experiments on 6a, 8,
and 11 or the incorporation of deuterium label into bibenzyl
and preparations of 6a, 6b, 6c, 8, 11, and deuterated BPMDA
(6 pages). See any current masthead page for ordering and
Internet access instructions.
Acknowledgment. Support of this research by Grants from
the National Science Foundation and the Public Health Service
of the National Institutes of Health (CA11045 and GM 53309-
26.) Funds for the purchase of the Finnigan MAT 95 mass
JA962287Z