Triplet-sensitized photorearrangement of 2-phenylallyl phosphites

Article abstract of DOI:10.1021/ja960428r

The triplet-sensitized photorearrangements of 2-phenylallyl phosphites 1, 3, and 4 to the corresponding 2-phenylallylphosphonates 1a, 3a, and 4a are shown to proceed with complete regioselectivity (5 → 6 and Scheme 1). A mechanism is proposed in which the

Full text of DOI:10.1021/ja960428r

6192
J. Am. Chem. Soc. 1996, 118, 6192-6201
Triplet-Sensitized Photorearrangements of 2-Phenylallyl
Phosphites
Wesley G. Bentrude,* Kevin P. Dockery, Srinivasan Ganapathy, Sueg-Geun Lee,
Michael Tabet, Yuh-Wern Wu, R. Thomas Cambron, and Joel M. Harris*
Contribution from the Department of Chemistry, UniVersity of Utah, Salt Lake City, Utah 84112
ReceiVed February 9, 1996X
Abstract: The triplet-sensitized photorearrangements of 2-phenylallyl phosphites 1, 3, and 4 to the corresponding
2-phenylallylphosphonates 1a, 3a, and 4a are shown to proceed with complete regioselectivity (5 f 6 and Scheme
1). A mechanism is proposed in which the 1,2-biradical-like styryl triplet, 9, adds oxidatively to three-coordinate
phosphorus to generate a cyclic, triplet, phosphoranyl 1,3-biradical, 10, that undergoes rapid subsequent â scission
to generate product 2-phenylallylphosphonate, 11. Phosphonate 1a is formed near-quantitatively from dimethyl
2-phenylallyl phosphite, 1. Phosphites 2-4, with phosphorus contained in a five-, six-, or seven-membered ring,
respectively, photorearrange much more slowly, although yields of phosphonates 3a and 4a of 50-70% are generated.
Quantum yields for the formation of phosphonates 1a, 3a, and 4a, on sensitization by triplet triphenylene, were
determined to be 0.25, 0.003, and 0.005, respectively, in benzene. Similar values were found in acetonitrile and
benzene using benzophenone as triplet sensitizer. Rate constants, kq, for efficient quenching of the triplets of
benzophenone (phosphorescence quenching) and triphenylene (photothermal lense measurements) were all in the
range 4.0-5.4 × 109 M-1 s-1. The low quantum yields for phosphonate formation from 3 and 4 (and presumably
2), therefore, result from the relative inefficiencies of reactions of their 2-phenylallyl triplets toward phosphorus.
This is ascribed to the large reduction in the rate of isomerization of the kinetically formed initial 1,3-biradical, 23,
to the thermodynamically more stable species, 24, an effect found in spiro phosphoranyl monoradicals. A reduction
in the rate of â scission of biradical 23, caused by its spiro structure, also may play a role. The failure of the
3-phenylallyl phosphite 7 to undergo the same cyclic photorearrangement as 1 is ascribed to the inability of the
benzyl radical-like terminus of triplet 7 to react with phosphorus. Phosphite homolog 8 also is inert toward cyclic
photorearrangement via a six-membered ring 1,3-biradical (28) analogous to 10. The reactivity patterns of phosphites
1-4, 7, and 8 can be rationalized in terms of the 1,2-biradical nature of the triplet styryl moiety and factors known
to govern the formation and permutational properties of phosphoranyl monoradicals.
Introduction
present research are hydrogen abstractions by triplet arylalkenes
that occur in intramolecular fashion.3
Aryl-substituted alkenes in the first triplet excited state have
been intensely studied of late.1 The Caldwell group has shown
them to be accurately characterized as orthogonal 1,2-
biradicals.1e,k,l As such they behave like alkyl mono free radicals
and can undergo ring-opening reactions1e and, along with other
triplet alkenes, add to alkenes2 and abstract hydrogen from
carbon-hydrogen bonds.2,3 Of particular relevance to the
The known reversible addition of primary alkyl radicals to
trialkyl phosphites to form phosphoranyl radicals, e.g. 16 (eq
5),4 suggested to us the possible intramolecular trapping of a
triplet alkene moiety by three-coordinate phosphorus to give a
phosphoranyl 1,3-biradical (e.g. 10, Scheme 1). Indeed we
reported preliminary investigations of 2-phenylallyl phosphites
whose triplet-sensitized photorearrangements are consistent with
this model.5a,b The present paper records mechanistic studies
directed at understanding the surprising inefficiency of these
X Abstract published in AdVance ACS Abstracts, June 15, 1996.
(1) See e.g.; (a) Lewis, F. D.; Bassani, D. M.; Caldwell, R. A.; Unett,
D. J. J. Am. Chem. Soc. 1994, 116, 10477. (b) Caldwell, R. A.; Diaz, J. F.;
Hrncir, D. C.; Unett, D. J. J. Am. Chem. Soc. 1994, 116, 8138. (c) Lewis,
F. D.; Bassani, D. M. J. Am. Chem. Soc. 1993, 115, 7523. (d) Tsubakiyama,
K.; Miyagawa, K.; Kaizaki, K.; Yamamoto, M.; Nishijima, Y. Bull. Chem.
Soc. Jpn. 1992, 65, 837. (e) Caldwell, R. A.; Zhou, L. J. Am. Chem. Soc.
1994, 116, 2271. (f) Caldwell, R. A.; Jacobs, L. D.; Furlani, T. R.; Nalley,
E. A.; Laboy, J. J. Am. Chem. Soc. 1992, 114, 1623. (g) Brennan, C. M.;
Caldwell, R. A. Photochem. Photobiol. 1991, 53, 165. (h) Caldwell, R. A.;
Riley, S. J.; Gorman, A. A.; McNeeney, S. P.; Unett, D. J. J. Am. Chem.
Soc. 1992, 114, 4424. (i) Waldeck, D. H. Chem. ReV. 1991, 91, 415. (j)
Saltiel, J.; Sun, Y.-P. In Photochromism, Molecules and Systems; Du¨rr, H.,
Bouais-Laurent, H., Eds.; Elsevier: Amsterdam, 1990; p 64. (k) Ni, T.;
Caldwell, R. A.; Melton, L. A. J. Am. Chem. Soc. 1989, 111, 457. (l) Aria,
T.; Sakuragi, H.; Tokumaru, K. Bull. Chem. Soc. Jpn. 1982, 55, 2204. (m)
Caldwell, R. A.; Cao, C. V. J. Am. Chem. Soc. 1982, 104, 6174. (n)
Bonneau, R. J. Am. Chem. Soc. 1982, 104, 2921. (o) Bonneau, R.; Herran,
B. Laser Chem. 1984, 4, 151.
(3) (a) Hornback, J. M.; Proehl, G. S. J. Am. Chem. Soc. 1979, 101,
7367. (b) Kropp, P. J.; Tise, F. P. J. Am. Chem. Soc. 1981, 103, 7293. (c)
Scully, F.; Morrison, H. J. Chem. Soc., Chem. Commun. 1973, 529. (d)
Kropp, P. J. J. Am. Chem. Soc. 1969, 91, 5783. (e) Padwa, A. Acc. Chem.
Res. 1979, 12, 310. (f) Padwa, A.; Chou, C. S.; Rosenthal, R. J.; Terry, L.
W. J. Org. Chem. 1988, 53, 4193 and references therein. (g) Padwa, A.
Org. Photochem. 1979, 4, 261. (h) Aoyama, H.; Omote, Y. J. Chem. Soc.,
Chem. Commun. 1985, 1381. Planar triplet arylcyclopropenes are especially
reactive in intramolecular hydrogen abstractions.3e-g
(4) (a) Baban, J. A.; Roberts, B. P. J. Chem. Soc., Perkin Trans. 2 1980,
876. (b) Dockery, K. P.; Bentrude, W. G. J. Am. Chem. Soc. 1994, 116,
10332.
(5) (a) Bentrude, W. G.; Lee, S.-G.; Akutagawa, K.; Ye, W.; Charbonnel,
Y. J. Am. Chem. Soc. 1987, 109, 1577. (b) Ganapathy, S.; Cambron, R. T.;
Dockery, K. P.; Wu, Y. W.; Harris, J. M.; Bentrude, W. G. Tetrahedron
Lett. 1993, 34, 5987. (c) For reviews of phosphoranyl radical chemistry
see: Bentrude, W. G. In The Chemistry of Organophosphorus Compounds;
Hartley, F. R., Ed.; John Wiley and Sons: Chichester, 1990; Vol. 1, Chapter
14. Bentrude, W. G. In ReactiVe Intermediates; Abramovitch, R. A., Ed.;
Plenum Press: New York, 1983; Vol. 3, Chapter 4. Bentrude, W. G. Acc.
Chem. Res. 1982, 15, 117. Roberts, B. P. In AdVances in Free Radical
Chemistry; Williams, G. H., Ed.; Heyden: London, 1979; Vol. 6, pp 225-
284.
(2) For a recent review of the photochemistry and photophysics of triplet
alkenes, see: Unett, D. J.; Caldwell, R. A. Res. Chem. Intermed. 1995, 21,
665. For a review of photoinduced isomerization of singlet and triplet aryl
olefins, see: Arai, T.; Tokumaru, K. Chem. ReV. 1993, 93, 23. For earlier
reviews of the photochemistry of both excited singlet and triplet alkenes
see: Kropp, P. J. J. Org. Photochem. 1979, 4, 1. Kropp, P. J. Mol.
Photochem. 1978/79, 9, 39.
S0002-7863(96)00428-3 CCC: $12.00 © 1996 American Chemical Society

Photorearrangements of 2-Phenylallyl Phosphites
J. Am. Chem. Soc., Vol. 118, No. 26, 1996 6193
Table 1. Yields of Triplet-Sensitized Photorearrangements of
0.1-0.5 M Deoxygenated Solutions of 1, 3, and 4 in Pyrex at 350
nm
Table 2. Quantum Yields φP for Triplet-Sensitized Formation of
Phosphonates 1, 3, and 4a
b
φP
% yielda
phosphite
solvent
Bc
Tc
phosphite
sensb
% conv
NMR/GC
isolated
solvent
1
CH3CN
C6H6
0.32
0.22
1
1
3
3
4
4
4
Tc
Bd
Tc
Pe
Tc
Bf
Pf
100
100
72
34
83
39
42
97g
95h
57g
33h
68g
71h
69h
86
i
54
i
59
i
i
C6H6
C6H6
C6H6
C6H6
C6H6
CH3CN
CH3CN
0.25
3
4
CH3CN
C6H6
CH3CN
C6H6
0.004
0.003
0.005
0.004
0.003
0.005
a Degassed solutions, 0.05 M sensitizer, 0.01-0.07 M phosphite,
335 nm light. b One sigma error of 2-6% from 3-5 determinations at
several conversions. c B, benzophenone; T, triphenylene.
a Based on phosphite consumed. b B, benzophenone; P, p-methoxy-
acetophenone; T, triphenylene. c 0.5 M in phosphite and sensitizer.d 0.1
M phosphite, 0.1 M sensitizer. e 0.1 M phosphite, 0.2 M sensitizer,
300 nm lamps. f 0.1 M phosphite, 0.3 M sensitizer. g 31P NMR. h GC.
i Not determined.
ment slowly, phosphites 3 and 4 with phosphorus in a ring gave
reduced yields of phosphonate compared to 1. A careful look
at the sluggish, triplet-sensitized photorearrangement of 3 shows
it to be accompanied by the formation of the apparent dimer of
the 2-phenylallyl radical, (PhCdCH2CH2)2 (ca. 0.2%), and
R-methylstyrene (ca. 1%). These side products, which may
come from P-O homolysis, are not found in the very efficient
triplet-sensitized photorearrangement of 1.
Regiochemistry. Phosphites 1, 3, and 4, deuterium labeled
at C1 as shown in 5a, 5c, and 5d, underwent photorearrangement
under triplet sensitized conditions in the regiospecific manner
represented by the conversion 5 f 6 (eq 1). The regiochemistry
photorearrangements when phosphorus is included in a 5-, 6-,
or 7-membered ring (2-4). A rationale for the reported results
is offered that is based on knowledge of reactions involving
phosphoranyl monoradicals. The reluctance of phosphite
analogs 7 and 8 to undergo photoisomerization to phosphonates
is likewise explained in terms of principles derived from earlier
work on phosphoranyl monoradicals.5c
Results
Product Studies. The photorearrangements of 0.01-0.1 M
deoxygenated solutions of phosphites 1-4 were carried out in
cyclohexane, benzene, and acetonitrile. Sensitizers were chosen
with triplet energies that should be above that for the 2-phe-
nylallyl functionality of 1-4, assumed to be close to the value
for R-methylstyrene (ET ) 61.7,1k 62.66 kcal/mol): Ph2CO (ET
) 69.2 kcal/mol7 ), triphenylene (ET ) 67.0 kcal/mol7), and
p-methoxyacetophenone (PMAP) (ET ) 71.7 kcal/mol7). Light
of appropriate wavelengths was transmitted through Pyrex to
ensure that energy was absorbed only by the sensitizer. A
benzene solution, 0.1 M in both phosphite 1 and benzophenone
was followed by 2H NMR which showed the product phospho-
nates (6) to have deuterium attached only to the carbon
designated C1 in eq 1. E.g., phosphonate 6a from phosphite
5a displayed equal-intensity deuterium resonances only at δ 5.28
and 5.43 (C6H6)) assigned to the diastereotopic deuteria at C1
(eq 1). This regiochemistry was confirmed for phosphites 5a
and 5c with PMAP as sensitizer and for 5a, 5c, and 5d with
the sensitizers benzophenone and triphenylene. The limits of
2
detection by H NMR give a lower limit of regioselectivity of
about 95%. In a parallel experiment, deuterium-labeled 1 (5a)
was directly irradiated through quartz with light from a 450-W
medium-pressure lamp. Unlike the sensitized reactions, deu-
terium was nearly randomly scrambled to give diastereomers
6a and 15a in approximately equal amounts. Deuterium at C1
of 15a was observed at δ 2.65 (5a f 6a + 15a).
Quantum Yields. To place the relative reactivities of
phosphites 1, 3, and 4 on a quantitative basis, the quantum yields
(φP) for formation of phosphonates 1a, 3a, and 4a were
determined with the sensitizers (0.05 M) triphenylene and
benzophenone which were irradiated with 335-nm light in a
Photon Technology Quantacount electronic actinometer. The
observed invariance of φP over a range of phosphite concentra-
tions (0.01-0.07 M) and the measured rate constants of triplet
quenching (see below), along with the known triplet lifetimes
of the sensitizers,8 ensured that the triplet energy of the
senzitizers was totally transferred to the phosphites. Results
are recorded in Table 2. The wide spread in quantum efficien-
cies confirms the more-qualitative reactivity differences noted
for the phosphites.
sensitizer, irradiated through Pyrex, showed 1 to be totally
converted to its phosphonate, 1a, in 4 h in near-quantitative
yield (Table 1). By contrast an initial indication of the
remarkable effect of placing phosphorus in a ring was seen when
a parallel reaction of phosphite 2 gave <1% of the correspond-
ing phosphonate in 48 h. GLC and/or isolated yields of
phosphonate (product accountabilities based on consumed
phosphite) from photorearrangement of 1, 3, and 4 are recorded
in Table 1. Notably, in addition to undergoing photorearrange-
Triplet Quenching Kinetics. The relative rate constants for
triplet quenching, kq, by phosphites 1, 3, 4, and (EtO)3P were
determined for the sensitizers triphenylene and benzophenone.
(6) Crosby, P. M.; Dyke, J. M.; Metcalfe, J.; Rest, A. J.; Salisbury, K.;
Sodeu, J. R. J. Chem. Soc., Perkin Trans. 2 1977, 182.
(7) Murov, S. L.; Carmichael, I.; Hug, G. L. Handbook of Photochem-
istry; Marcel Dekker: New York, 1993.
(8) τT for benzophenone 50 µs7 and for triphenylene 55 µs.7
(9) Kuhlmann, R.; Schnabel, W. Polymer 1976, 17, 419.

6194 J. Am. Chem. Soc., Vol. 118, No. 26, 1996
Bentrude et al.
Table 3. Quenching of Triplet States of Triphenylene and
Benzophenone
Scheme 1
10-9kq,a M-1 s-1
quencher
solvent
benzophenoneb
triphenylenec
phosphite 1
CH3CN
C6H6
CH3CN
C6H6
CH3CN
C6H6
CH3CN
C6H6
4.0 ( 0.07
4.0 ( 0.2
4.4 ( 0.3
phosphite 3
phosphite 4
(C2H5O)3P
4.5 ( 0.08
4.1 ( 0.01
1.3 ( 0.04
3.7 ( 0.1
5.4 ( 0.4
<0.001
3.9
R-methylstyrene
CH3CN
C6H6
4.6 ( 0.3
2.7 ( 0.3d
(OMe)2) has been measured to be exothermic by 24 kcal/mol.1k
The benzylic nature of the carbon radical center may stabilize
10 and facilitate the photorearrangement.
a Errors at 1 σ confidence level. b From decay of benzophenone
phosphorescence. c From photothermal lense spectroscopy. d Reference
9.
Triplet ketones are known to add oxidatively to three-
coordinate phosphorus,12 and the phosphorescence of benzophe-
none was quenched by (EtO)3P (Table 3). The quenching of
benzophenone phosphorescence by phosphites 1, 3, and 4,
therefore, might be chemical in nature. Readily envisaged is a
process for the observed photoisomerizations whereby adduct
12, formed by oxidative addition of triplet benzophenone to
phosphorus, cycloadds to the styryl moiety to form 13. (The
trapping of phosphoranyl monoradicals by alkenes is a known
reaction.13 ) Elimination of benzophenone from 13 generates
1,3-biradical 10. The possibility that the observed photorear-
rangements are initiated by single electron transfer from the
styryl phosphites to triplet benzophenone is ruled out by our
recent study of the rearrangements of 1, 3, and 4 induced by
Quenching of the phosphorescence of benzophenone, induced
by 325-nm laser excitation, was monitored in acetonitrile and
benzene. The quenching kinetics in benzene, however, were
not straightforwardly interpretable. Thus, kq was obtained in
acetonitrile only. Since triphenylene is not phosphorescent in
solution, kq was determined by photothermal lense spectros-
copy.10 The kq values (Table 3) in both types of experiments
were determined from observed pseudo-first-order rate constants
via the Stern-Volmer approach (see Experimental Section).
Thermal Rearrangement. Numerous control reactions
demonstrated the thermal stabilities of phosphites 1, 3, and 4
at ambient temperature. However, at 165-170 °C a 1.6 M
p-cymene solution of 5a was at least 60% consumed in 24 h to
give phosphonate exclusively deuterium labeled at C3, 6a, along
with lesser amounts of a second phosphonate, MeP(O)(OMe)2
(identified by GC), in 3/2 to 5/4 ratios. An approximately 18%
yield (based on phosphite consumed) of 6a were isolated by
flash chromatography. Neither reaction nor isolation procedures
for 6a was optimized.
Isomers and Homologs of 1. Placement of the phenyl group
at C3 of the allyl functionality (phosphite 7) rendered that
molecule inert to photorearrangement under benzophenone
electron transfer to the first singlet excited state of 9,10-
dicyanoanthracene (DCA) in acetonitrile.14 Though the same
phosphite f phosphonate rearrangement is induced by singlet
DCA, those processes are characterized by low and nearly
identical quantum yields for phosphonate formation from 1 (φP
) 0.03) and 3 (φP ) 0.02)14 and are clearly different mechan-
istically from the reactions discussed in the present paper.
Triphenylene is an ideal triplet sensitizer for the photorear-
rangements of 1, 3, and 4 as it does not possess a carbonyl
functionality; and the energy of its first excited singlet state
(83.4 kcal/mol7) is much too low for it to be a singlet sensitizer
(ES for R-methylstyrene is 99.6 kcal/mol6). Moreover, the free
energy change for electron transfer to triplet triphenylene in
benzene is unfavorable by 44 kcal/mol.15 Hence triphenylene
can only transfer triplet energy to phosphites 1, 3, and 4 (triplet
energy, ET, for triphenylene ) 66.5 kcal/mol7; ET for R-meth-
sensitization. Cis-trans isomerization about the double bond,
however, was observed. Homologation of the allyl (2-propenyl)
chain gave a molecule (phosphite 8) that yielded a variety of
unidentified products on benzophenone-sensitized reaction.
Discussion
2-Phenylallyl Phosphites. The observed efficient triplet
quenchings and specific regiochemistries of the reactions of
phosphites 1, 3, and 4 are consistent with the formation of the
triplet styryl unit (9) with the unpaired electrons in adjacent
orthogonal orbitals. This moiety adds oxidatively to phosphorus
to generate a phosphoranyl 1,3-biradical (10, Scheme 1).
Subsequent intersystem crossing and â scission (10 f 11) to
form PdO and CdC π bonds yields the 2-phenylallylphos-
phonate (11). Chemically, this transformation amounts to an
intramolecular Arbuzov rearrangement. ∆H° for transformation
of trimethyl phosphite ((MeO)3P) to its phosphonate (MeP(O)-
(11) Lewis E. S.; Colle, K. S. J. Org. Chem. 1981, 46, 4369. An earlier
estimate of -44 kcal/mol was based on measured and calculated heats of
formation from the literature (Fu, J.-J.; Bentrude, W. G.; Griffin, C. E. J.
Am. Chem. Soc. 1972, 94, 7717).
(12) (a) Fox, M. A. J. Am. Chem. Soc. 1979, 101, 5339 and references
therein. (b) Chow, Y. L.; Marciniak, B. J. Org. Chem. 1983, 48, 2910. (c)
Okazaki, R.; Tamara, K.; Hirabayashi, Y.; Inamoto, N. J. Chem. Soc., Perkin
Trans 1 1976, 1924. (d) Alberti, A.; Griller, D.; Nazran, A. S.; Pedulli, G.
F. J. Org. Chem. 1986, 51, 3959. (e) Griffin, C. E.; Bentrude, W. G.;
Johnson, G. M. Tetrahedron Lett. 1969, 969. (f) McGimpsey, G.; Depew,
M. C.; Wan, J. K. S. Phosphorus Sulfur 1984, 21, 135.
(13) (a) Griller, D.; Roberts, B. P. J. Chem. Soc., Perkin Trans. 2 1973,
1416. (b) Davies, A. G.; Parrot, M. J.; Roberts, B. P. J. Chem. Soc., Perkin
Trans. 2 1976, 1066.
(14) Ganapathy, S.; Dockery, K. P.; Sopchik, A. E.; Bentrude, W. G. J.
Am. Chem. Soc. 1993, 115, 8863.
(10) Braslavsky, S. E.; Hiebel, G. E. Chem. ReV. 1992, 92, 1381.

Photorearrangements of 2-Phenylallyl Phosphites
J. Am. Chem. Soc., Vol. 118, No. 26, 1996 6195
ylstyrene ) 62 kcal/mol1k,6). Moreover, PMAP, whose π-π*
triplet is generally less chemically reactive16 than the nfπ*
triplet of benzophenone, effectively sensitizes these photorear-
rangements (Tables 1 and 2). The similarities of the kq and φP
values for benzophenone and triphenylene sensitizations suggest
strongly that both act as triplet sensitizers. Evidently triplet
energy transfer from benzophenone to the styryl moiety of these
phosphites is faster than any chemical quenching process.
Lacking a triplet acceptor, (EtO)3P is unable to quench triplet
triphenylene. Its 3-fold reduction in kq with benzophenone,
compared to the values for phosphites 1, 2, 3, and 4, is consistent
with its quenching of benzophenone phosphorescence by a
chemical mechanism.
or phosphonate stage (Experimental Section) but is intrinsic to
the photorearrangement process.19 Two possible mechanisms
are the following: (1) competing, concerted 1,2- and 2,3-
sigmatropic rearrangements; and (2) scission to a caged phos-
phonyl/2-phenylallyl radical pair that bonds randomly to
phosphorus at C1 and C3. These studies will be reported
separately.19 They are relevant to the present study in that they
show that the styryl-like singlet excited states of 1, 3, and 4 do
not undergo rapid intersystem crossing to the triplet (9) prior
to reaction, which would have led to the regiospecific process
observed on triplet sensitization.
Effects of Ring Size on ΦP. The greatly reduced quantum
efficiencies encountered when phosphorus is included in a 6-
or 7-membered ring (phosphites 3 and 4) could arise from
inefficient triplet energy transfer or because the styryl-like triplet
excited state is relatively unreactive. The essentially equal, near
diffusion-controlled kq values measured for phosphites 1, 3, and
4 (Table 3) show triplet quenching to be very rapid in all cases
and comparable to that measured for R-methylstyrene itself
(Table 3). Indeed at the concentrations of phosphites used,
triplet energy transfer, in competition with triplet-singlet decay,
should be 100% efficient.20 Evidently, based on the φP values
of Table 2, triplet 1 undergoes cyclization and formation of
product phosphonate relatively efficiently, while for triplet 3
and 4 these steps are inefficient.
The triplet-sensitized photorearrangements of the above
phosphites find some intermolecular analogies,12,17 the most
direct being the xanthone-sensitized reaction of 1,1-diphe-
nylethene with tetraphenyldiphosphine (eqs 2-4).17 Attack at
A reasonable understanding of the inefficiencies of the
photorearrangements of 3 and 4 can be obtained by consideration
of known features of phosphoranyl monoradicals: the revers-
ibility of the addition of alkyl radicals to trialkyl phosphites
and the retardation of the rates of apical-equatorial exchange
of phosphorus substituents when phosphorus is part of a spiro
ring system. The reversibility of the addition of methyl radical
to a series of trialkyl phosphites (eqs 5 and 6) has been
phosphorus by the terminal carbon of the excited alkene was
proposed to lead to displacement (eq 3), perhaps via a
phosphoranyl 1,3-biradical. Alkoxy-radical-like reactions of
benzophenone triplets12d with tetraphenyldiphosphine12c and
(EtO)2POP(OEt)212d have been reported. Photoreactions of
other aromatic ketones,12d 4,4′-dimethoxybenzothioketone,12f and
quinones12d with (EtO)2POP(OEt)2 and/or tetraethyldiphosphine
likewise illustrate the 1,2-biradical-like reactivity of triplet
functionality toward three-coordinate phosphorus. Moreover,
aroylmethyl and acylmethyl dimethyl phosphites, RCOCH2OP-
(OMe)2, undergo photorearrangement to CH2dCRP(O)(OMe)2,
perhaps via a triplet 1,3-biradical intermediate similar to 10.12e
Singlet Processes. The observed thermal rearrangement of
1, in contrast to its room-temperature triplet-sensitized photo-
rearrangement, occurs only at relatively high temperatures. Its
regiochemistry and the temperature required are consistent with
earlier results for other allyl phosphites.18 Side product MeP-
(O)(OMe)2 is the obvious result of an intermolecular thermal
Arbuzov process involving two molecules of 1. Whether this
formally 2,3-sigmatropic process (based on its regiochemistry,
5a f 6a) occurs via a concerted or stepwise mechanism is
unknown. Conceivably, triplet phosphoranyl biradical 10 may
be converted by intersystem crossing to the same ground state
singlet cyclic biradical formed in a stepwise, thermal rearrange-
ment that is then rapidly and irreversibly converted to 11.
By contrast, direct photoirradiation of 5a (Results Section),
5c, and 5d yields both phosphonate regioisomers (e.g. 6a and
15a from 5a) in nearly equal amounts.19 Control experiments
show that this scramble of label does not occur at the phosphite
demonstrated by ESR techniques that showed the signal intensity
for 16, monitored under conditions of steady-state methyl radical
flux, to be strongly temperature dependent.4a Furthermore, alkyl
radical from â scission could not be observed, suggesting that
this cleavage is too slow to compete with R scission to
regenerate P(OR)3. This was confirmed by chemical studies4b
that demonstrated that no stable products result from the room
temperature reaction of ethyl radical with (MeO)3P or (EtO)3P.
By contrast, ethyl radicals react with PhCH2OP(OMe)2 to give
C2H5P(O)(OMe)2 and bibenzyl, presumably by way of 17.4b The
formation of the relatively stable benzyl radical by â scission
is a comparatively rapid process that effectively traps 17 in
competition with reformation of ethyl radical (R scission).
(15) (a) Rehm, D.; Weller, A. Isr. J. Chem. 1970, 8, 259. (b) Calculations
made using for triphenylene ES ) 3.62 eV,7 Ered ) -2.22 eV vs SCE,
CH3CN (Julliard, M.; Chanon, M. Chem. Br. 1082, 18, 558); for (EtO)3P
Eox ) 1.57 eV vs SCE, CH3CN (Ohmori, H.; Nakai, S.; Masui, M. J. Chem.
Soc., Perkin Trans 1 1979, 7023); and R-methylstyrene Eox ) 1.76 eV vs
SCE, CH3CN (Katz, M.; Riemenschneider, P.; Wendt, H. Electrochem. Acta
1972, 17, 1595).
(16) Turro, N. J. Modern Molecular Photochemistry; Benjamin/Cum-
mings Pub.: Menlo Park, CA, 1978; Chapter 10.
(17) Okazaki, R.; Hirabyashi, Y.; Tamura, K.; Inamoto, N. J. Chem. Soc.,
Perkin Trans 1 1976, 1034.
(19) Lee, S.-G.; Baik W.; Wu, Y.-W.; Tabet, M. Unpublished results
from this laboratory.
(20) This can be seen by consideration of the measured values of kq of
Table 3 of 109-1010 M-1 s-1, the triplet lifetimes for benzophenone (50
µs8) and triphenylene (55 µs8), and the concentration range of phosphites
employed (0.01-0.07 M).
(18) Herriott, A. W.; Mislow, K. Tetrahedron Lett. 1968, 3013.

6196 J. Am. Chem. Soc., Vol. 118, No. 26, 1996
Bentrude et al.
Kinetic ESR work6 has established that the most rapid
permutation available to the near-trigonal bipyramidal phos-
phoranyl radical represented by 18 is an exchange involving
the two exocylic substituents (X and Y) and the position
occupied by the odd electron (18 h 19). This exchange is of
mode 4 (M4).21 Because it involves ligands not included in the
ring, it is termed M4(exo). It occurs rapidly even at 200 K with
Scheme 2
substituent-dependent rate constants (kexo) of 107-109 s-1 22
. At
radicals,4 then kR (23f24) must be of the order 107-108 s-1
.
higher temperatures an M4(ring) permutation participates (18
f 20) that includes the ring atoms. At 273 K kring is about
106-108 s-1. Strikingly, the spiro radical 21 could not be
induced to undergo rapid exchange even at 393 K.23
This value is not unreasonable considering the range noted above
for kring for species such as 18 (18 f 20). With the spiro system
depicted in Scheme 2, rate-determining kexo(ring) should be
much decreased. Therefore, 23 will largely revert to reactant
phosphite rather than undergo relatively slow isomerization to
24. This leads to an inefficient conversion of phosphite to
phosphonate.
The role of the spiro ring system in suppressing the rate of
permutation of substituents is not well understood but almost
certainly is related to geometric constraints imposed by the ring
on the transition state for exchange. Proposed geometries for
the transition state or intermediates for permutational exchange
of ligands attached to phosphorus in phosphoranyl monoradicals
include the following: local C3V σ*; trigonal bipyramidal with
odd-electron apical; and square pyramidal with odd electron
basal.26 Less sterically constrained rings might allow for rapid
permutation to occur. The effect on φP of a seven- rather than
six-membered ring (phosphites 3 vs 4, Table 2), however, is
marginal at best.
In Scheme 2 triplet-singlet intersystem crossing (ISC) of the
initial 1,3-biradical 23 precedes its isomerization, its â scission
to 25, and its R scission to ground state phosphite. (Thus,
R-scission of 23 in Scheme 2 yields ground state phosphite
rather than its triplet (22).) However, whether this should be
true for all processes available to 23 is not certain. Chemical
processes generally compete with ISC of long-lived triplet
biradicals.27 Indeed, 1,3-biradicals in all-carbon 5-membered
rings (the cyclopentane-1,3-diyl system) often have triplet
ground states with lifetimes with respect to ISC of 10-100 ns
or greater.27,28 In phosphoranyl 1,3-biradicals such as 10 or
23, the phosphorus atom may act to enhance the ease of
intersystem crossing and/or perturb the singlet-triplet energy
levels. From the rate constants given above for M4(ring),
lifetimes with respect to exchange of the order 1-100 µs are
predicted for non-spiro phosphoranyl 1,3-biradicals analogous
to 18 (e.g. 10 with X ) MeO), long enough to allow ISC to
occur before permutation of the sort illustrated by 18 f 19.
The estimate made above for kR for 23 (107-108 s-1) would
give it a lifetime towards R scission of 10-100 ns, during which
ISC may occur leading to R scission to ground state phosphite
rather than 22.
Scheme 2 depicts the intramolecular reaction of the styryl
triplet of 22 with phosphorus. The new carbon phosphorus bond
in the trigonal bipyramidal 1,3-biradical is formed apically,
consistent with the implications of kinetic studies of the reverse
process, R-scission of carbon-phosphorus bonds of phospho-
ranyl monoradicals.24 (The near-trigonal bipyramidal nature of
closely related phosphoranyl monoradicals possessing alkyl and
oxy substituents is well-established.25 ) If the lifetime of the
triplet styryl moiety of these phosphites is similar to that of
styrene itself (22-25 ns1k,m,n,o) and â-methylstyrene (26,1k 27,1o
461m ns), then to compete with decay of the triplet, 22 f 23
must occur with a rate constant (ka) of 107-108 s-1. It is
proposed that when a spiro system is not involved (i.e. a radical
analogous to 18 is present), kex(ring) of Scheme 2 is rapid and
traps kinetically formed 23 by its conversion to 24, a lower-
energy form. (In 24 ring P-O and P-C bonds are in apical
and equatorial positions, respectively, where they are known
to be favored thermodynamically.6c,25) Assuming that 22 f
23 is reversible, as is known for the addition of primary alkyl
(21) The modes of permutation of a molecule containing five-coordinate
phosphorus have been defined (Musher, J. I. J. Chem. Ed. 1974, 51, 94.
Musher, J. I. J. Am. Chem. Soc. 1972, 94, 5662). Mode refers to the position
of substituents on the trigonal bipyramidal framework before and after the
permutation that exchanges their positions without regard to pathway
(mechanism). For phosphoranyl radicals the odd electron or vacant position
is considered stereochemically to be a fifth ligand. In the mode 4(exo)
permutation two ligands that are not part of the ring, along with the odd
electron, change positions. Recent calculations have indicated that the
permutational possibilities and pathways for phosphoranyl radicals without
ligands attached to the phosphorus part of a ring system may be more
extensive than for those determined experimentally for 18 and sensitive to
the nature of the ligands.25b
(22) Cooper, J. W.; Parrot, M. J.; Roberts, B. P. J. Chem. Soc., Perkin
Trans 2 1977, 730.
(26) See ref 25b and citations therein.
(23) Griller, D.; Roberts, B. P. J. Chem. Soc., Perkin Trans. 2, 1973,
1416.
(24) Cooper, J. W.; Roberts, B. P. J. Chem. Soc., Perkin Trans. 2, 1976,
808.
(25) (a) See ref 5c for experimental evidences. Also, Schipper, P.; Jansen,
E. H. J. M.; Buck, H. M. Top. Phosphorus Chem. 1977, 9, 407. (b) For a
recent theoretical paper on the relative energies of TBP structures for [MeP-
(OMe)3]• and (MeO)4P•, including those with the odd electron either
equatorial or apical, see: Gustafson, S. M.; Cramer, C. J. J. Phys. Chem.
1995, 99, 2267.
(27) For recent reviews of biradicals, see: (a) Johnston, L. J. Chem. ReV.
1993, 93. 251. (b) Dougherty, D. A. Acc. Chem. Res. 1991, 24, 94. (c)
Adam, W.; Grabowski, S.; Wilson, R. M. Acc. Chem. Res. 1990, 23, 165.
(d) Caldwell, R. A. In Kinetics and Spectroscopy of Carbenes and
Biradicals, Platz, M. S., Ed.; Plenum Publ. Co.: New York, 1990; pp 77-
116. (e) Johnston, L. J.; Scaiano, J. C. Chem. ReV., 1989, 89, 521.
(28) For recent considerations of the effects of structure on the lifetimes
of 1,3-cyclopentanediyl triplet biradicals, see: Kita, F.; Nau, W. M.; Adam,
W J. Am. Chem. Soc. 1995, 117, 8670. Engel, P. S.; Lowe, K. L.
Tetrahedron Lett. 1994, 35, 2267.

Photorearrangements of 2-Phenylallyl Phosphites
J. Am. Chem. Soc., Vol. 118, No. 26, 1996 6197
Scheme 2 might mistakenly be taken to suggest that â scission
is necessarily more rapid for 24 than for 23. Indeed, whether
there is an apical or equatorial preference for this process in
phosphoranyl monoradicals is as yet unclear.29 In fact in
Scheme 2, â scission of 23 may simply not be able to compete
effectively with its R scission to phosphite 22. In fact rapid
irreversible trapping of 23 by the rate-determining permutation
23 f 24 would allow â scission of 24 to occur with kâ for 24
less than kex(ring) for 23 and perhaps even less than kâ for 23.29
In our view of Scheme 2 outlined above, for the overall
photoprocess to be efficient, kex(ring) for 23 f 24 must be of
the same order or greater than that for 23 f phosphite. This
would place kâ for 23 f 25 at less than 107-108 s-1. However,
if 23 f 24 is very slow, as in the spiro systems, the product
forming â scission for the slow, inefficient production of 25
might even be 23 f 25.
It is difficult to estimate a priori what kâ for either 23 or 24
should be. These scissions might be extremely rapid since the
formation of two π bonds renders the reaction very favorable
energetically. Carbon-centered 5-membered-ring triplet 1,3-
biradicals, however, do not typically readily undergo such ring
openings.27 Nevertheless, the phosphorus center and the avail-
ability of d orbitals set this case apart.
A second potential factor that also should be considered is
the known reduction in kâ for phosphoranyl monoradicals on
placement of phosphorus in a 5-membered ring. Indeed, the
octamethyl ring-substituted analog of 21 gives no evidence for
â scission even at 45 °C13a and must be generated well above
100 °C for the chemical products of ring-opening to be
observed.30 As noted above, the ESR spectrum for 19 persists
at 120 °C.23 By contrast, (t-BuO)4P• and [t-BuOP(OEt)3]• are
subject to â scission at -70 °C.31 In all of the phosphoranyl
1,3-biradicals proposed as intermediates for the photorearrange-
ments of 1-4, it is the same 5-membered ring that must undergo
â scission. Indeed, part of the effect of the OCH2CH2O attached
to phosphorus in 21, and its tetramethyl-substituted analog, may
be to slow the â scission within the 5-membered ring. What
the influence of a 6- or 7-membered ring in a phosphoranyl
monoradical is on the â scission of the 5-membered ring of a
species such as 26, and by inference on 23 or 24, is unknown
and intrinsically difficult to study. Thus in radicals analogous
to 20, but with phosphorus in a 6- rather than 5-membered ring,
only ring-opening â scission occurs to the exclusion of â scission
of ethoxy groups also attached to phosphorus (X ) Y )
ethoxy).31 For 20 itself for X ) Y ) ethoxy both ring and
ethoxy C-O â scission occur.31 Radical 26, therefore, is likely
to undergo exclusive scission of the 6-membered ring.
Structural Variations in the Styryl Chromophore. The
failure of the attempted triplet-sensitized photo-Arbuzov reaction
of 7 is consistent with the potential reactive carbon center being
benzylic, a reflection of the 1,2-biradical character of such triplet
moieties. Thus, it has been found that benzyl radicals from
intermediate 17 fail to react with PhCH2OP(OMe)2 or at least
are not trapped by formation of PhCH2P(O)(OMe)2.4b This is
in accord with the relatively stable nature of the benzyl radical
and the weakness of the phosphorus-carbon bond to be formed.
By analogy, formation of phosphoranyl 1,3-biradical from triplet
7 (7 f 27) likely is too unfavorable thermodynamically. The
carbon radical center of 27 lacks the stabilizing phenyl sub-
stituent found in 10. Moreover, â scission of any 27 formed,
to generate the 1-phenylallylphosphonate, probably is unable
to compete with the R scission reforming phosphite 7.
Disappointingly, the potential 1,3-biradical 28 from 8, the
homolog of 1, failed to yield a clean formation of phosphonate
product of any kind. Evidently, even if 28 is generated, the â
scission shown is not rapid enough to generate 1,3-biradical 29
that might close to cyclopropyl product 30. Ethoxy phospho-
ranyl monoradicals (EtO)4P• undergo â scission at 25 °C with
kâ of only 3 × 103 s-1 32
.
Clearly, this step would be too slow
to trap an intermediate (28) undergoing R scission to reform
phosphite 8 with kR of the order 107-108 M-1 s-1 (see above).
Moreover, the formation of 28 involving a 6-membered ring
may be kinetically and/or thermodynamically disfavored com-
pared to the formation of 5-membered ring 1,3-biradicals like
10.
Conclusions
The photorearrangements of 2-phenylallyl phosphites, inti-
tiated by triplet sensitization, proceed with essentially complete
regioselectivity. The results are consistent with the mechanism
of Scheme 1 involving a novel phosphoranyl 1,3-biradical, 10.
The efficient reaction of phosphite 1 and the failure of phosphite
7 to undergo an analogous photorearrangement are understood
in terms of the 1,2-biradical model of triplet arylalkenes set
forth by Caldwell.1e,k,m Thus, reaction of 1 involves attack at
phosphorus by what is essentially a primary alkyl radical, a
species known to add oxidatively to phosphites. The terminus
of triplet 7, by contrast, resembles a relatively stable benzyl
radical, an intermediate characteristically unreactive toward
phosphites. The low quantum yields encountered for the
photorearrangements of phosphites 3, 4, and presumably 2 result
not from inefficient transfer of triplet energy but rather because
the triplet styryl moiety has a reduced reactivity toward
phosphorus. A reasonable understanding of the latter is found
by consideration of the known greatly reduced rates of ligand
exchange for phosphoranyl monoradicals of spiro structure.23
It is proposed that the rate-determining trapping of reversibly
formed intermediate 23 by conversion to 24, Scheme 2, is
greatly reduced when phosphorus is placed in a 5-, 6-, or
7-membered ring. This leads to predominant reformation of
phosphite rather than a rapid 23 f 24 step, followed by â
scission to phosphonate 25. The known reduction in rate of â
scission for phosphoranyl monoradicals, when phosphorus is
contained in a small-ring spiro system,13a,30 also may be
important. The reactivity trends seen for phosphites 1-4, 7,
and 8 are generally interpretable in terms of previously
discovered principles5 governing the reactions of monoradicals
with three-coordinate phosphorus, the permutational properties
of phosphoranyl monoradicals, and the assumption that the
(29) ESR results interpreted as evidence for a preference for â scission
from the equatorial position have been put forth. See: Roberts, B. P.; Singh,
K J. Chem. Soc., Perkin Trans. 2 1980, 1549.
(30) Akhmetkhanova, F. M.; Rol'nik, L. Z.; Pastushenko, E. V.;
Proskurnina, M. V.; Zlot-skii, S. S.; Rakhmankulov, D. L. J. Gen. Chem.
USSR 1985, 55, 1810.
(31) (a) Davies, A. G.; Griller, D.; Roberts, B. P. J. Chem. Soc., Perkin
Trans. 2 1972, 2224. (b) Davies, A. G.; Griller, D.; Roberts, B. P. J. Chem.
Soc., Perkin Trans. 2 1972, 993.
(32) Calculated from the activation parameters given in ref 31a.

6198 J. Am. Chem. Soc., Vol. 118, No. 26, 1996
Bentrude et al.
triplet 2-phenylallyl moiety behaves as a 1,2-biradical,1e,l,n the
methylene terminus of which behaves as a primary monoradical.
The formation of phosphoranyl 1,3-biradical 10(23, 24) is most
likely further favored by the benzylic nature of the radical
centered on carbon.
NMR). Only the resonance corresponding to the 2-phenylallylphos-
phonate was observed in the 31P NMR of the reaction mixtures. 2H
NMR analysis of the concentrated, crude reaction mixtures for both
solvents showed that the rearrangement occurred with complete
regioselectivity. Flash chromatography gave only the phosphonate
regioisomer, 5a. 31P NMR (121.4 MHz, C6D6) δ 29.10; 2H NMR (46.0
MHz, C6H6) δ 5.43 (s, br, 0.5 2H, CdCH2H), 5.28 (s, br, 0.5 2H,
CdCH2H) (50:50 mixture of E and Z isomers); 1H NMR (299.9 MHz,
C6D6) δ 2.85 (d, 2 H, PCH2, 2JHP ) 22.2 Hz), 3.29 (d, 6 H, OCH3, 3JHP
) 10.8 Hz), 5.25 (dt, 0.5 H, CdCH2H, 4JHP ) 5.5 Hz, 4JHH ) 0.9 Hz),
5.38 (d, 0.5 H, CdCH2H, 4JHP ) 5.4 Hz), 7.01-7.16 (m, 3 H), 7.36-
Experimental Section
Materials. Ether (Et2O) and tetrahydrofuran (THF) were purified
by distillation under nitrogen from sodium/benzophenone. Benzene
(Baker Photrex) was used as received. Acetonitrile was distilled over
calcium hydride under nitrogen. Trimethyl phosphite was distilled
prior to use. Benzophenone (99% pure) and p-methoxyacetophenone
(PMAP), both from Aldrich, were used as received. Triphenylene was
recrystallized from ethanol. 3-Bromo-2-phenylpropene33 was prepared
in 73% yield in routine manner from phenylallyl alcohol on reaction
with phosphorus tribromide. Cyclic methyl phosphites,34 used in the
Arbuzov reactions, were prepared by the exchange reaction of the
appropriate phosphoramidite (i.e., N,N-diethyl-1,3,2-dioxaphosphorinan-
2-amine or N,N-diethyl-1,3,2-dioxaphosphepan-2-amine) with excess
methanol. 2,5-Diphenyl-1,5-hexadiene was prepared by the literature
method.35 Unless otherwise noted, distillations were performed using
a short-path apparatus without distillation column.
Physical Methods. Melting points were obtained on a Thomas
Hoover Capillary Melting Point apparatus, and are uncorrected. Proton
(1H), deuterium, (2H), carbon (13C), and phosphorus (31P) NMR spectra
were taken on Varian XL-300 and Varian Unity 300 NMR spectrom-
eters. Occasional 31P NMR (32.2 MHz) spectra were obtained on a
Varian FT-80 NMR. Chemical shifts (δ) are recorded in parts per
million from tetramethylsilane (0.00) with the solvent as internal
standard: C6HD5 (1H) 7.15, (13C) 128.0; CHCl3 (1H) 7.24, (13C) 77.0;
C6DH5 (2H) 7.15. 31P NMR spectra were referenced to external 85%
phosphoric acid.
7.44 (m, 2 H). 13C NMR (75.4 MHz, C6D6) δ 32.57 (d, PCH2, 1JCP
)
138.7 Hz), 52.17 (d, OCH3, 2JCP ) 6.6 Hz), 116.71 (td, C)CH2H, 1JCD
3
) 24.2 Hz, JCP ) 10.9 Hz), 126.63, 127.86, 128.50, 139.19 (d, ipso-
3
2
Ph, JCP ) 10.1 Hz), 141.17 (d, C)CH2H, JCP ) 4.2 Hz); GC-EIMS
(EI -70 eV) m/z (relative intensity) 227 [M]+ (71), 226 [M - 1]+
(12), 212 [M - CH3]+ (6), 132 (82), 131 (89), 130 (42), 119 (40), 118
(30), 117 (30), 116 (100), 115 (35), 109 [(MeO)2PO]+ (55), 95 (27),
94 (25), 92 (41), 91 (23), 79 (45), 77 [Ph]+ (15). C11H142HO3P [M]+:
HRMS [M]+ (calcd) 227.0822, (obsd) 227.0814.
Triphenylene-Sensitized Photorearrangement of Dimethyl [1-2H]-
2-Phenyl-2-propenyl Phosphite (5a). A solution of triphenylene (47.2
mg, 0.207 mmol, 0.0414 M) and dimethyl [1-2H]-2-phenyl-2-propenyl
phosphite (5a) (332.0 mg, 1.46 mmol, 0.292 M) in C6H6 (Photrex) was
apportioned and then irradiated for 6 h, after which analysis by 31P
NMR showed that about 30% of the starting phosphite had been
converted to phosphonate. Flash chromatography gave exclusively
dimethyl (E)-, (Z)-[3-2H]-2-phenyl-2-propenylphosphonate, 6a. (Spec-
tra were identical with those of phosphonate from PMAP-sensitized
photorearrangement of 6a.)
Triphenylene-Sensitized Photorearrangement of 2-([1-2H]-2-
Phenyl-2-propenoxy)-1,3,2-dioxaphosphorinane (5c). A solution of
triphenylene (52.6 mg, 0.231 mmol, 0.0462 M) and 2-([1-2H]-2-phenyl-
2-propenoxy)-1,3,2-dioxaphosphorinane (5c) (141.0 mg, 0.590 mmol,
0.118 M) in C6H6 (Photrex) was apportioned and irradiated for 7 days.
2H NMR showed about 40% conversion of 5c to phosphonate 6c with
complete regioselectivity. Flash chromatography gave exclusively (E)-,
(Z)-2-oxo-2-([3-2H]-2-phenyl-2-propenyl)-1,3,2-dioxaphosphorinane, 6c.
(Spectra of phosphonate 6c were obtained from 80% 2H-labeled
phosphite 5c.) 31P NMR (121.4 MHz, C6D6) δ 21.70; 2H NMR (46.0
MHz, C6F6) δ 5.23 (s, br, 0.5 2H, CdCH2H), 5.43 (s, br, 0.5 2H,
CdCH2H) (50:50 mixture of E and Z isomers); 1H NMR (299.9 MHz,
C6D6) δ 0.72-0.84 (m, 1 H, CH2CH2CH2), 1.00-1.12 (m, 1 H,
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 fused silica capillary column.
Other low-resolution EIMS (70 eV) along with HRMS (EI, 70 eV)
data were obtained on a Finnegan MAT 95 mass spectrometer.
Intensities (parentheses) are percentages of the base peak. The
molecular ion is indicated by [M]+. Elemental analyses were performed
by Atlantic Microlab, Inc., Norcross, GA, and Galbraith Laboratories,
Inc., Knoxville, TN.
Photorearrangement Regiochemistry. Reaction solutions were
typically 5.00 mL in volume, prepared under nitrogen atmosphere in a
glovebag, and then apportioned between a Pyrex 5 mm NMR tube and
one or two 12 mm Pyrex tubes, all fitted with 10/30 joints. After 3 to
4 freeze-pump-thaw degassings (0.01 mmHg), the tubes were flame-
sealed. Rearrangement on irradiation through Pyrex in a Rayonet
2
CH2CH2CH2), 2.93 (d, 2 H, PCH2, JHP ) 22.1 Hz), 3.26-3.40 (m, 2
H, CH2CH2O), 3.78-3.92 (m, 2 H, CH2CH2O), 5.26-5.32 (m, 0.6 H,
4
CdCH2H), 5.39 (d, 0.6 H, CdCH2H, JHP ) 5.3 Hz), 7.02-7.20 (m,
3 H), 7.36-7.42 (m, 2 H); GC-EIMS (EI -70 eV) m/z (relative
intensity) 239 [M]+ (64), 238 [M - 1]+ (27), 157 (18), 156 (18), 141
(23), 129 (22), 119 (51), 118 (51), 117 (37), 116 (100), 115 (49), 105
(21), 92 (32), 91 (38), 80 (23), 78 (23), 77 [Ph]+ (23), 65 (32), 51
(28), 41 (80). C12H142HO3P [M]+: HRMS [M]+ (calcd) 239.0822,
(obsd) 239.0818. In a parallel experiment, a 0.1 M benzene solution
of 2-([1-2H]-2-phenyl-2-propenoxy)-1,3,2-dioxaphosphorinane (5c)
with added p-methoxyacetophenone as sentizer was irradiated through
Pyrex at 300 nm to give chromatographically isolated phosphonate with
2
apparatus equipped with 350 nm lamps was monitored by 31P and H
NMR. Control experiments showed rearrangement did not occur in
the absence of UV light and/or appropriate sensitizers. Flash chro-
matography on silica gel (ethyl acetate) yielded the pure phosphonate
product which was characterized by 31P, 2H, 13C, and HRMS and
comparison with authentic unlabeled material. Deuterium was attached
to C1 of product phosphonate with random stereochemistry resulting
in an approximately 50/50 ratio of E and Z 2H regioisomers, as observed
2
a H NMR spectrum identical to that from the above triphenylene-
2
by H NMR. The regioselectivity, i.e. formation of 1-2H (6) vs 3-2H
sensitized study.
2
Triphenylene-Sensitized Photorearrangement of 2-([1-2H]-2-
Phenyl-2-propenoxy)-1,3,2-dioxaphosphepane (5d). A solution of
triphenylene (50.6 mg, 0.222 mmol, 0.0444 M) and 2-([1-2H]-2-
phenyl-2-propenoxy)-1,3,2-dioxaphosphorinane (159.0 mg, 0.628
mmol, 0.125 M) in C6H6 (5.00 mL, Photrex) was apportioned and
irradiated for 5 days (∼40% conversion, 31P NMR). Purification by
flash chromatography gave exclusively (E)-, (Z)-2-oxo-2-([3-2H]-2-
phenyl-2-propenyl)-1,3,2-dioxaphosphepane, 6d. 31P NMR (121.4
MHz, C6D6) δ 25.11; 2H NMR (46.0 MHz, C6F6) δ 5.22 (s, br, 0.5 2H,
(15) phosphonate regioisomer, was also determined by H NMR.
PMAP-Sensitized Irradiation of Dimethyl [1-2H]-2-Phenyl-2-
propenyl Phosphite (5a). Two solutions were prepared and divided
as described above: (1) a 5.00 mL solution of phosphite 5a (155 mg,
0.689 mmol, 0.138 M) and p-methoxyacetophenone (PMAP) (74 mg,
0.50 mmol, 0.10 M) in benzene (Photrex); (2) a 5.00 mL solution of
phosphite 1-1d (127 mg, 0.564 mmol, 0.113 M) and p-methoxyac-
etophenone (PMAP) (82 mg, 0.55 mmol, 0.11 M) in acetonitrile. After
18 h of irradiation, all of the starting phosphite was consumed (31P
2
CdCH2H), 5.42 (s, br, 0.5 H, CdCH2H) (50:50 mixture of E and Z
1
(33) Hatch, L. F.; Patton, T. L. J. Am. Chem. Soc. 1954, 76, 2705.
(34) For the preparation of 2-methoxy-1,3,2-dioxaphosphorinane and
other cyclic methyl phosphites as well as their physical properties, see:
White, D. W.; Bertrand, R. D.; McEwen, G. K.; Verkade, J. G. J. Am.
Chem. Soc. 1970, 92, 7125.
isomers); H NMR (299.9 MHz, C6D6) δ 0.91-1.06 (m, 2 H, CH2-
2
CH2O), 1.08-1.24 (m, 2 H, CH2CH2O), 2.94 (d, 2 H, PCH2, JHP
)
22.3 Hz), 3.26-3.40 (m, 2 H, CH2CH2O), 3.72-3.87 (m, 2 H,
CH2CH2O), 5.32 (dt, 0.5 H, CdCH2H, 4JHP ) 5.4 Hz, 4JHH ) 1.0 Hz),
4
(35) Kolobielski, M.; Pines, H. J. Am. Chem. Soc. 1975, 79, 5820
5.40 (d, 0.5 H CdCH2H, JHP ) 5.4 Hz), 7.02-7.20 (m, 3 H), 7.38-

Photorearrangements of 2-Phenylallyl Phosphites
J. Am. Chem. Soc., Vol. 118, No. 26, 1996 6199
7.44 (m, 2 H); GC-EIMS (EI -70 Ev) m/z (relative intensity) 253 [M]+
(97), 252 [M - 1]+ (6), 199 (52), 198 (30), 143 (21), 119 (80), 118
(50), 117 (39), 116 (100), 115 (36), 97 (20), 92 (42), 78 (19), 77 [Ph]+
(17), 55 (65), 54 (45), 51 (19), 41 (26). C13H162H03P [M]+: HRMS
[M]+ (calcd) 253.0978, (obsd) 253.0977.
Benzophenone-Sensitized Photorearrangement of Dimethyl [1-2H]-
2-Phenyl-2-propenyl Phosphite (5a). A solution of benzophenone
(99.5 mg, 0.546 mmol, 0.109 M) and dimethyl [1-2H]-2-phenyl-2-
propenyl phosphite, 5a (134.9 mg, 0.594 mmol, 0.119 M), in acetonitrile
(5.00 mL) was apportioned and irradiated 14 h. Flash chromatography
gave exclusively dimethyl (E)-, (Z)-[3-2H]-2-phenyl-2-propenylphos-
phonate, 6a. Spectra were identical to those of phosphonate 6a from
the PMAP-sensitized process.
Benzophenone-Sensitized Photorearrangement of 2-([1-2H]-2-
Phenyl-2-propenoxy)-1,3,2-dioxaphosphorinane (5c). A solution of
benzophenone (96.0 mg, 0.527 mmol, 0.105 M) and 2-([1-2H]-2-phenyl-
2-propenoxy)-1,3,2-dioxaphosphorinane, 5c (160.2 mg, 0.670 mmol,
0.134 M), in C6H6 (5.00 mL) was apportioned and irradiated for 40 h
to ∼25% conversion (31P NMR). Flash chromatography gave exclu-
sively (E)-, (Z)-2-oxo-2-([3-2H]-2-phenyl-2-propenyl)-1,3,2-dioxaphos-
phorinane, 6c. Spectra were identical with those of phosphonate 6c
from the triphenylene-sensitized process.
PMAP-Sensitized Photorearrangement of 2-([1-2H]-2-Phenyl-2-
propenoxy)-1,3,2-dioxaphosphorinane (5c). Deoxygenated 0.1 M
benzene solutions of phosphite in Pyrex tubes also containing 0.2 M
and 1.0 M PMAP were irradiated at 300 and 350 nm, respectively.
Flash chromatography followed by HPLC (ethyl acetate/ethanol ) 9/1)
gave 6c whose 2H NMR spectrum was identical to that of phosphonate
6c from triphenylene-sensitized photorearrangement.
were identical to those of phosphonate 6a from PMAP-sensitized
photorearrangement.
General Procedure for Determination of Yields of Triphenylene-
Sensitized Photorearrangements of 2-Phenylallyl Phosphites. A 2.00
mL solution containing approximately 0.1 mmol each of triphenylene,
phosphite, and internal standard (tri-n-propyl phosphate) in benzene
(Photrex) was apportioned in 0.75-, 0.75-, and 0.50-mL amounts into
three Pyrex NMR tubes. All operations between weighings were
conducted in an argon-flushed glovebag. After 3 freeze-pump-thaw
degassing cycles (0.01 mmHg), the tubes were flame sealed. The two
tubes containing 0.75 mL of solution were irradiated with 350-nm lamps
in a Rayonet reactor while the third was a control maintained in the
dark at 0 °C. The progress of the reactions was followed quantitatively
by 31P NMR (decoupler off, 60 s repetition rate). Flash chromatography
of the contents of one of the tubes (0.75 mL) on silica gel (100% ethyl
acetate) isolated the pure product phosphonate. 31P NMR showed none
of the controls to have undergone rearrangement. Phosphonate products
were compared spectroscopically to authentic samples of 1a, 3a, and
4a for which independent preparations and spectroscopic and analytical
data are given below.
Yields of Phosphonate with Sensitizers Other Than Triphenylene.
Appropriately degassed 0.1 M solutions of phosphite with an appropriate
amount of benzophenone and internal standard were irradiated under
conditions indicated in Table 1. The photorearrangements were
monitored by GLC/FID techniques. GLC response factors of phosphite
and product phosphonate relative to the internal standard were carefully
determined.
(R,S)-[1-2H]-2-Phenyl-2-propen-1-ol. To a vigorously stirred
mixture of activated MnO2 (90%, Fluka, used as received, 220 g, 2.3
mol) and 250 mL of methylene chloride at 0 °C was added, under
nitrogen over a 15-min period, a solution of >98% pure 2-phenylallyl
alcohol (10.5 g, 0.078 mol) in 30 mL of methylene chloride. The
reaction was allowed to warm to room temperature and then stirred
for 3 h. (Longer reaction times were found to be undesirable.) The
solution was filtered through a sintered glass frit that was then rinsed
with methylene chloride. The filtrate was dried over MgSO4, and the
solvent was removed by rotary evaporation. Crude 2-phenyl-2-propenal
(∼8 g) was taken up in 40 mL of dry diethyl ether and used in the
subsequent reduction.
Direct Irradiation of Dimethyl [1-2H]-2-Phenyl-2-propenyl Phos-
phite (5a). A 25.0 mL solution of dimethyl [1-2H]-2-phenyl-2-propenyl
phosphite (305.6 mg, 1.35 mmol, 0.0540 M) in C6H6 (Photrex) was
divided equally between three 10 mm quartz tubes and one Pyrex NMR
tube. The quartz tubes were stoppered with rubber septa and purged
with argon for 10 min. The NMR tube was flame sealed. The solutions
were irradiated for 3 h with light from a medium pressure 450-W Hg
lamp in a quartz thimble to 100% conversion of 5a (31P NMR). The
products in one of the tubes were isolated by flash chromatography to
2
give, as determined by H NMR, a 1:1 mixture of dimethyl (E)-, (Z)-
[3-2H]-2-phenyl-2-propenylphosphonate (6c) and dimethyl (R,S)-[1-2H]-
2-phenyl-2-propenylphosphonate (15c): 31P NMR (121.4 MHz, C6D6)
To 100 mL of dry distilled ether, magnetically stirred under nitrogen,
was added 98% 2H-labeled powdered lithium aluminum deuteride
(Aldrich, 2.5 g, 0.060 mol). The stirred mixture was cooled to 0 °C,
and to it was added, dropwise, the solution of crude 2-phenyl-2-propenal
over a 15 min period. The reaction mixture was diluted to 500 mL
with ether, quenched with water at 0 °C, and filtered. The filtrate was
dried over MgSO4. Solvent removal by rotary evaporation and
distillation of the crude reaction mixture gave 3.5 g (0.026 mol, 33%
yield) of 97-98% pure 2H-labeled 2-phenylallyl alcohol (normally
>98% 2H incorporation by 1H, 13C NMR): 2H NMR (46.0 MHz, C6H6)
2
2
δ 28.97; H NMR (46.0 MHz, C6F6) δ 2.76 (s, br, 0.5 H, CH2HP),
5.26 (s, br, 0.25 H, CdCH2H), 5.41 (s, br, 0.25 H, CdCH2H); H
2
2
1
NMR (299.9 MHz, C6D6) δ 2.84 (m, PCH2HC, 2JHP ) 22.3 Hz), 2.85
2
3
(d, PCH2C, JHP ) 22.2 Hz), 3.30 (d, 6 H, CH3O, JHP ) 10.8 Hz),
5.27-5.33 (m), 5.36-5.43 (m), 7.02-7.15 (m, 3 H), 7.40-7.46 (m, 2
H).
Control Reaction on Scrambling of Deuterium in Dimethyl
[3-2H]-2-Phenyl-3-(2H)-2-propenylphosphonate (6a). Deoxygenated
solutions in benzene and in cyclohexane, 0.1 M in both 6a and
undeuterated dimethyl 2-phenyl-2-propenyl phosphite, were irradiated
in a quartz tube with a 450 W Hanovia medium-pressure lamp. After
the phosphite had been nearly totally consumed (24 h), the samples
2
1
δ 4.18 (d, CH2HOH, JDH ) 1.9 Hz); H NMR (299.9 MHz, C6D6) δ
1.86 (d, 1 H, OH, JHH ) 5.6 Hz), 4.22 (s, br, 1 H, CH2HOH), 5.26
3
(m, 1 H, CdCH2), 5.34 (m, 1 H, CdCH2), 7.02-7.16 (m, 3 H), 7.25-
7.31 (m, 2 H); 13C NMR (75.4 MHz, C6D6) δ 64.33 (t, CH2HOH, 1JCD
) 21.8 Hz), 111.97 (CdCH2), 126.34, 127.91, 128.61, 139.23 (ipso-
Ph), 147.82 (CdCH2).
2
were analyzed by H NMR which showed no deuterium attached to
C1 in the deuterium-labeled phosphonate.
Thermally-Induced Rearrangement of Dimethyl [1-2H]-2-Phenyl-
2-propenyl Phosphite (5a). A 5.00-mL solution of phosphite 5a (191.4
mg, 0.843 mmol, 0.169 M) in freshly distilled p-cymene was divided
between two thick-walled Pyrex NMR tubes (∼0.5 mL each) and one
12-mm Pyrex tube (∼4 mL), freeze-thaw degassed, flame-sealed, and
then heated to 165-170 °C. After 24 h, approximately 35% of the
starting phosphite remained (31P NMR based on total peak areas). Two
major phosphonate products were observed in the 31P NMR in a ratio
of approximately 5:4 in one tube and 3:2 in another. The major
phosphonate (δ31P, 28.95) was the deuterium-labeled 2-phenylal-
lylphosphonate 6a, and the minor phosphonate (δ31P, 32.72) was
dimethyl methylphosphonate. 2H NMR (CdCH2H, δ 5.3, 5.4) showed
completely regioselective phosphonate formation (6a) along with some
side products. Solvent and volatile impurities were removed from the
contents of the NMR tube in vacuo (40 °C, 0.05 mmHg). Flash
chromatography gave 18 mg (0.08 mmol) of exclusively dimethyl (E)-,
(Z)-[3-2H]-2-phenyl-2-propenylphosphonate (6a) (2H NMR). Spectra
Dimethyl 2-Phenylallyl Phosphite (1). In a glovebag, under a
nitrogen atmosphere, dimethyl diethylphosphoramidite (18.6 g, 0.113
mol), 400 mg of 1H-tetrazole (0.006 mol), 99% pure (GLC) 2-pheny-
lallyl alcohol (8.4 g, 0.063 mol), and 50 mL of freshly distilled, dry
acetonitrile were combined in a 100-mL round bottom flask equipped
with a stir bar. The flask was capped with a rubber septum which was
replaced outside the glovebag with a nitrogen inlet. The reaction
mixture was stirred at room temperature under nitrogen for 10 h and
then heated to 30 °C to remove the solvent and remaining phosphora-
midite through a vacuum adaptor (0.05 mmHg). The crude product
was transferred in a glovebag to a short-path distillation apparatus, and
the product mixture was distilled in vacuo (67 °C, 0.02 mmHg) to give
three fractions of >99% purity (GLC) (11 g, 0.049 mol, 58% yield) of
1. When flame-sealed in ampules under nitrogen, phosphites 1, 3, and
4 could be stored indefinitely at -10 °C. High-purity 2-phenylallyl
alcohol and rigorous exclusion of oxygen are essential to the preparation
of pure phosphite, as separation from phosphate contamination cannot

6200 J. Am. Chem. Soc., Vol. 118, No. 26, 1996
Bentrude et al.
be achieved by short-path distillation. Distillation on a long-path
column at higher temperatures removed phosphate but led to a small
amount of thermal isomerization of the phosphite to the phosphonate:
C6D6) δ 0.67 (m, 1 H, CH2CH2CH2), 1.99 (m, 1 H, CH2CH2CH2), 3.30-
3.42 (m, 2 H, OCH2CH2), 4.08-4.20 (m, 2 H, OCH2CH2), 4.58 (m, 2
H, OCH2C), 5.38-5.46 (m, 2 H, CdCH2), 7.04-7.19 (m, 3 H), 7.28-
7.36 (m, 2 H); 13C NMR (75.4 MHz, C6D6) δ 28.64 (d, CH2CH2CH2,
1
31P NMR (121.4 MHz, C6D6) δ 140.90; H NMR (299.9 MHz, C6D6)
δ 3.29 (d, 6 H, OCH3, 3JHP ) 10.5 Hz), 4.65 (ddd, 2 H, OCH2, 3JHP
)
2
3JCP ) 5.4 Hz), 59.35 (s, CH2CH2O, 64.60 (d, OCH2C, JCP ) 20.5
13.6 Hz, 4JHH ) 1.2 Hz, 4JHH < 1.0 Hz), 5.40-5.42 (m, 2 H, CdCH2),
Hz), 113.44 (s, C)CH2), 126.41, 127.85, 128.60, 138.80 (ipso-Ph),
145.43 (d, CdCH2, 3JCP ) 5.6 Hz); GC-EIMS (EI -70 eV) m/z (relative
intensity) 238 [M]+ (8), 237 [M - 1]+ (7), 151 -bromo (25), 141
(18), 128 (22), 118 (56), 117 (42), 116 (28), 115 (100), 105 (30), 103
(19), 91 (44), 79 (17), 77 [Ph]+ (32), 65 (30), 51 (22), 41 (63).
C12H15O3P [M]+: HRMS [M]+ (calcd) 238.0758, (obsd) 238.0771.
Anal. Calcd for C12H15O3P: C, 60.50; H, 6.35; P, 13.00. Found: C,
60.25; H, 6.18; P, 12.96.
7.04-7.18 (m, 3 H), 7.28-7.36 (m, 2 H); 13C NMR (75.4 MHz, C6D6)
2
2
δ 48.85 (d, OCH3, JCP ) 9.9 Hz), 63.83 (d, OCH2, JCP ) 12.5 Hz),
113.48 (s, CdCH2), 126.37, 128.00, 128.56, 138.77 (ipso-Ph), 145.26
3
(d, CdCH2, JCP ) 4.7 Hz); GC-EIMS (EI -70 eV) m/z (relative
intensity) 226 [M]+ (4), 225 [M - 1]+ (7), 210 [M-CH3]+ (5), 131
(63), 130 (52), 129 (20), 118 (37), 117 [M - (MeO)2PO]+ (47), 116
(26), 115 (100), 109 [(MeO)2PO]+ (39), 94 (26), 93 (34), 91 (46), 79
(22), 77 [Ph]+ (19), 63 (22), 51 (22); UV λmax(ꢀ) 240 (10 295), 280
(291), 290 (112). C11H15O3P [M]+: HRMS [M]+ (calcd) 226.0759,
(obsd) 226.0745. Anal. Calcd for C11H15O3P: C, 58.38; H, 6.69; P,
13.70. Found: C, 58.28; H, 6.67; P, 13.60.
Preparation and Photolysis of 2-(2-Phenyl-2-propenyl)dioxaphos-
pholane (2). Phosphite 2 was prepared in analogous fashion to the
above phosphite (7) in 46% yield, bp 99-100 °C (0.1 mmHg): 31P
(CDCl3) δ 135.2; 1H NMR (CDCl3) δ 7.33 (m, 5 H, C6H5), 5.48, 5.36
(2 H, m, CdCH2), 4.65 (2 H, m, OCH2), 3.91-4.22 (CH2CH2). A
deoxygenated solution 0.1 M in both phosphite and benzophenone was
irradiated in a 3-mm Pyrex tube with a 450-W Hanovia light source
for 48 h to give less that 1% of the product phosphonate. Under these
conditions a parallel solution of dimethyl 2-phenyl-2-propenyl phosphite
was totally converted to its phosphonate in 4 h.
Dimethyl 2-Phenylallylphosphonate (1a). A solution of 3-bromo-
2-phenylethene (1.0 g, 0.0051 mol) and trimethyl phosphite (7.0 g, 0.056
mol) was heated to 105 °C. Reaction was accompanied by the evolution
of bromomethane. After 0.5 h excess trimethyl phosphite and meth-
ylphosphonate side product were removed in vacuo at room temperature.
Flash chromatography of the residue on silica gel (98:2 CHCl3-CH3-
OH) gave 450 mg (0.0020 mol, 39% yield) of pure phosphonate 1a
2-(2-Phenyl-2-propenoxy)-1,3,2-dioxaphosphepane (4). The above
procedure for 1 was followed using 2-phenylallyl alcohol (8.9 g, 0.066
mol), N,N-diethyl-1,3,2-dioxaphosphepan-2-amine (13.9 g, 0.0727 mol),
1H-tetrazole (0.470 g, 0.007 mol), and 50 mL of freshly distilled
acetonitrile. Removal of the excess 1,3,2-dioxaphosphepan-2-amine
was found to be difficult, even after heating the crude product mixture
to 90 °C in vacuo (0.05 mmHg) for several hours. The product mixture
was therefore distilled collecting the first fractions containing 4 in a
cooled flask (-78 °C). Further distillation gave, as the purest fraction
2.1 g (0.0083 mol, 13% yield) of 4 (104 °C, 0.05 mmHg, 98% pure by
GLC) as a colorless oil. Fractions were stored in several flame-sealed
ampules under nitrogen. The relatively high boiling point of 4 led to
reduced yields because of thermal isomerization of the phosphite: 31P
1
(GLC) as a colorless oil: 31P NMR (121.4 MHz, C6D6) δ 29.10; H
NMR (299.9 MHz, C6D6) δ 2.86 (dd, 2 H, PCH2, 2JHP ) 22.2 Hz, 4JHH
3
) 1.0 Hz), 3.30 (d, 6 H, OCH3, JHP ) 10.8 Hz), 5.30 (ddt, 1 H,
CdCH2, 4JHP ) 5.5 Hz, 2JHH ) 1.0 Hz, 4JHH ) 1.0 Hz), 5.40 (dd, 1 H,
4
2
CdCH2, JHP ) 5.5 Hz, JHH ) 1.0 Hz), 7.04-7.16 (m, 3 H), 7.38-
7.44 (m, 2 H). 13C NMR (75.4 MHz, C6D6) δ 32.74 (d, PCH2, 1JCP
139.0 Hz), 52.23 (d, OCH3, 2JCP ) 6.4 Hz), 117.01 (d, CdCH2, 3JCP
)
)
10.7 Hz), 126.63, 127.86, 128.49, 139.28 (d, ipso-Ph, 2JCP ) 10.0 Hz),
141.19 (d, CdCH2, 3JCP ) 4.6 Hz); GC-EIMS (EI -70 eV) m/z (relative
intensity) 226 [M]+ (54), 225 [M - 1]+ (8), 211 [M - CH3]+ (10),
131 (74), 130 (60), 129 (26), 118 (31), 117 (24), 116 (25), 115 (100),
109 [(MeO)2PO]+ (42), 94 (40), 91 (44), 79 (39), 77 [Ph]+ (15).
C11H15O3P [M]+: HRMS [M]+ (calcd) 226.0759, (obsd) 226.0765.
Anal. Calcd for C11H15O3P: C, 58.38; H, 6.69; P, 13.70. Found: C,
58.09; H, 6.61; P, 13.89.
1
NMR (121.4 MHz, C6D6) δ 134.27; H NMR (299.9 MHz, C6D6) δ
1.24 (m, 4 H, CH2CH2O), 3.44-3.58 (m, 2 H, CH2CH2O), 3.85-3.99
(m, 2 H CH2CH2O), 4.68 (m, 2 H, OCH2C), 5.41 (m, 1 H, CdCH2),
Dimethyl (R,S)-[1-2H]-2-Phenyl-2-propenyl Phosphite (5a). The
above procedure for the unlabeled phosphite (1) was followed using
[1-2H]-2-phenyl-2-propen-1-ol (3.2 g, 0.024 mol), dimethyl dieth-
ylphosphoramidite (6.0 g, 0.036 mol), 1H-tetrazole (0.300 g, 0.004 mol),
and 25 mL of freshly distilled acetonitrile. The crude product was
distilled in vacuo (74 °C, 0.05 mmHg) to give, as the purest fraction,
1.5 g (0.0066 mol, 28% yield) of 97% pure phosphite 5a (GLC): 31P
5.47 (m, 1 H, CdCH2), 7.00-7.17 (m, 3 H), 7.26-7.32 (m, 2 H); 13
C
NMR (75.4 MHz, C6D6) δ 30.38 (s, CH2CH2O), 63.19 (d, CH2CH2O,
2JCP ) 2.6 Hz), 64.41 (d, OCH2C, 2JCP ) 20.5 Hz), 113.51 (s, CdCH2),
2
126.42, 127.96, 128.56, 138.82 (s, ipso-Ph), 145.39 (d, CdCH2, JCP
) 5.3 Hz); GC-EIMS (EI -70 eV) m/z (relative intensity) 252 [M]+
(7), 251 [M - 1]+ (2), 198 (34), 197 (26), 142 (27), 118 (70), 117
(60), 116 (37), 115 (100), 103 (21), 96 (20), 91 (51), 77 [Ph]+ (33), 55
(80), 54 (27), 51 (19). C13H17O3P [M]+: HRMS [M]+ (calcd) 252.0915,
(obsd) 252.0909. A distillation fraction 100% pure by GLC was used
for elemental analysis. Anal. Calcd for C13H17O3P: C, 61.90; H, 6.78.
Found: C, 61.02; H, 6.88.
2
NMR (121.4 MHz, C6D6) δ 140.92; H NMR (46.0, C6H6) δ 4.60 (s,
br, OCH2H); 1H NMR (299.9 MHz, C6D6) δ 3.29 (d, 6 H, OCH3, 3JCP
) 10.5 Hz), 4.60-4.69 (m, 1 H, OCH2H), 5.39-5.43 (m, 2 H, CdCH2),
7.04-7.18 (m, 3 H), 7.28-7.36 (m, 2 H); 13C NMR (75.4 MHz, C6D6)
δ 48.85 (d, OCH3, 3JCP ) 9.3 Hz), 63.55 (td, OCH2H, 1JCD ) 22.3 Hz,
3JCP ) 9.9 Hz), 113.54 (s, CdCH2), 126.38, 128.00, 128.57, 138.78
(ipso-Ph), 145.22 (d, CdCH2, 4JCP ) 4.7 Hz); GC-EIMS (EI -70 eV)
m/z (relative intensity) 227 [M]+ (5), 226 [M - 1]+ (8), 212 [M -
CH3]+ (4), 132 (67), 131 (70), 130 (33), 119 (46), 118 (57), 117 (30),
116 (100), 115 (31), 109 (50), 95 (23), 93 (45), 92 (47), 91 (23), 79
(27), 77 [Ph]+ (21). C11H142HO3P [M]+: HRMS [M]+ (calcd)
227.0822, (obsd) 227.0830.
Triplet-Quenching Kinetics. Samples were irradiated by 325-nm,
5-ns laser pulses produced by frequency doubling the 650 nm output
of a Quanta Ray Model PDL-2 laser, where the dye (DCM, Exciton)
was pumped by the second harmonic (532 nm) output of a Quanta
Ray Model GCR-2 Nd:YAG laser. The 10 Hz pulse train from the
laser is reduced to 0.1 Hz by a mechanical shutter. Average output
intensities of 300 and 200 µJ/pulse were used for the photothermal
lens and phosphorescence quenching techniques, respectively.
2-(2-Phenyl-2-propenoxy)-1,3,2-dioxaphosphorinane (3). The
procedure for the dimethyl 2-phenylallyl phosphite (1) was followed
with a slight modification, using 2-phenylallyl alcohol (12.2 g, 0.091
mol), N,N-diethyl-1,3,2-dioxaphosphorinan-2-amine (22.8 g, 0.13 mol),
400 mg of 1H-tetrazole (0.006 mol), and 50 mL of acetonitrile freshly
distilled from CaH2. Complete removal of the excess 1,3,2-dioxaphos-
phorinan-2-amine after workup by heating the crude product mixture
to 60 °C in vacuo (0.05 mmHg) for 2 h was essential to avoid
contamination of the distilled phosphite by phosphoramidite. Distil-
lation (103 °C, 0.02 mmHg) gave 14 g (0.058 mol, 65% yield) of 3
(99% pure by GLC). The boiling point of the phosphate is significantly
higher than that of the phosphite allowing their separation by distillation.
When bath temperatures exceed 130 °C, some thermal isomerization
occurs: 31P NMR (121.4 MHz, C6D6) δ 131.35; 1H NMR (299.9 MHz,
Signal detection and data acquisition for both techniques have been
described previously.36 Briefly, phosphorescence was collected at 90°
through an interference notch filter, having a 10 nm pass band centered
at 430 nm, and detected with a photomultiplier. Time-resolved
photothermal lens signals were detected by aligning a 5-mW HeNe
probe laser beam collinearly with the pulsed excitation beam and
directing both beams into the sample. Beyond the sample, the probe
laser beam was separated from the excitation beam by a dispersing
prism, allowed to propagate to the far-field (about 3 m from the sample),
and centered on a mask having a radially symmetric, parabolic
transmittance profile. The intensity of the radiation passed by the mask,
(36) Cambron, R. T.; Harris, J. M. J. Phys. Chem. 1993, 97, 13598.

Photorearrangements of 2-Phenylallyl Phosphites
J. Am. Chem. Soc., Vol. 118, No. 26, 1996 6201
collected with a lens and focussed onto a photomultiplier, is proportional
to the second moment of the probe beam spatial distribution which
reports the strength (inverse focal length) of the thermal lens in the
sample. Signal transients from both techniques were fit to a first-order
kinetic model using nonlinear least-squares to give an observed first
order, k1; this result was employed in the relationship k1 ) ko + kq[Q],
where ko is the rate constant for decay in the absence of quencher, kq
is the quenching rate constant, and [Q] is the quencher concentration.
Quantum yields at 335 nm were determined on a Photon Technol-
ogy International Quanticount electronic actinometer, precalibrated with
ferrioxalate, as previously outlined.38 In these studies, however, the
quartz cuvette (quantum yield cell) was modified such that it could be
flame sealed following degassing of the sample at <5 × 10-6 mmHg
prior to irradiation. Phosphite conversions were less than 6%.
Phosphonite formation and phosphite consumption were monitored by
GC/FID analysis on a DB-1 capillary column with tri-n-phosphate as
sensitivity-calibrated internal standard. Other details are found in Table
2.
Freshly opened Baker Photrex reagent benzene and EM Science
Omnisolve acetonitrile were employed. Benzophenone (50 µM) and
triphenylene (40 µM) solutions with varying concentrations (2-40 µM)
of R-methylstyrene or phosphite as quenchers were prepared. To avoid
oxidation of the phosphite, solutions were prepared in a glovebag with
solvents previously purged with argon and using cells fitted with a
degassing bulb and a high-vacuum Teflon stopcock. The latter was
closed to protect the solutions prior to degassing (6 freeze-thaw-
pump cycles at <5 µTorr) and during the quenching experiments.
Benzophenone phosphorescence quenching in acetonitrile involved
three measurements of each signal averaged over 100 transients for
each experiment. The measured triplet lifetime in the absence of
quencher was 240 ( 10 µs (av of 10 measurements). Triphenylene
triplet quenching in benzene and acetonitrile was measured by
photothermal lens spectroscopy, where three to five replicates of each
signal (the average of 100 transients) were acquired. The unquenched
triplet lifetimes of triphenylene were determined to be 45 ( 6 µs in
benzene solution and 110 ( 16 µs in acetonitrile (average of five
measurements); these results are consistent in magnitude with a
previously reported result of 55 µs for the triplet lifetime of triphenylene
in cyclohexane.37
Acknowledgment. This work was supported by grants from
the National Science Foundation (CHE9319419 and previous
grants to W. G. B.; CHE9006667 and CHE9510312 to J. M.
Harris) and by grants (to W. G. B.) from the Public Health
Service, National Institutes of Health (CA 11045 and GM
53309).
Supporting Information Available: Detailed descriptions
of the triphenylene-sensitized reactions of 1, 3, and 4; prepara-
tions of 2-phenylallyl alcohol, the diethylamino precursors of
3 and 4, the phosphonates 3a and 4a, the deuterated phosphites
5c and 5d, phosphites 7 and 8 (6 pages). Ordering information
is given on any current masthead page.
JA960428R
(37) Porter, G.; Wilkinson, F. Proc. R. Soc. London, Ser. A 1961, 264,
1.
(38) Koptyug, I. V.; Ghatlia, N. D.; Sluggett, G. W.; Turro, N. J.;
Ganapathy, S.; Bentrude, W. G. J. Am. Chem. Soc. 1995, 117, 9486.