Triplet-sensitized photorearrangements of six-membered-ring 2- phenylallyl phosphites. Reaction efficiency and stereochemistry at phosphorus

Full text of DOI:10.1021/jo9918302

2786
J . Org. Chem. 2000, 65, 2786-2791
Notes
Sch em e 1
Tr ip let-Sen sitized P h otor ea r r a n gem en ts of
Six-Mem ber ed -Rin g 2-P h en yla llyl
P h osp h ites. Rea ction Efficien cy a n d
Ster eoch em istr y a t P h osp h or u s
David C. Hager and Wesley G. Bentrude*
Department of Chemistry, University of Utah,
Salt Lake City, Utah 84112
Received November 30, 1999
In tr od u ction
2-Phenylallyl phosphites undergo triplet benzophe-
none- or triphenylene-sensitized photorearrangement to
the corresponding phosphonates with the regiochemistry
illustrated (deuterium label) by the acyclic example 1 f
2 (Scheme 1).1 The overall process is energetically very
favorable. Chemical yields with 1 are near-quantitative
and accompanied by relatively high quantum yields for
phosphonate formation, φP, of 0.2-0.3.1b-d Mechanisti-
cally, it has been proposed1 that the relaxed (90°) π-π*
triplet of 1 behaves as a 1,2-biradical, 3, whose terminus
mimics a primary monoradical and oxidatively adds to
phosphorus to give phosphoranyl 1,3-biradical 4 (Scheme
1).1d â Scission of 4 provides the two π bonds in 2.
Strikingly, the quantum yields for 5 are reduced to
0.002-0.003.1c,d However, the chemical yields of phos-
phonate from 5 and its analogue 7 remain reasonably
high. These photorearrangements have potential syn-
thetic advantage over the classical Arbuzov reaction
approach to phosphonates in that it avoids the use of allyl
halides, such as CH2dCPhCH2Br, which may be ac-
companied by allylation at multiple nucleophilic sites.
Allylphosphonates are readily converted to synthetically
useful vinylphosphonates.2
enhanced3b when the double bond is in a five-membered
ring (6, n ) 1) which likely imparts higher energy and a
longer lifetime to the planar alkene π-π* triplet.4
We report here studies of the stereochemistry at
phosphorus of the triphenylene triplet-sensitized photo-
rearrangements of diastereomeric phosphites cis- and
trans-7. Cis and trans refer to the relationship of the tert-
butyl and 2-phenylallyloxy substituents on the 1,3,2-
dioxaphosphorinane rings of 7. The product phosphonate
8 is designated cis when the tert-butyl and phosphoryl
oxygen (PdO) are attached cis to one another on the ring.
The present assignment for 8 differs from previous
practice5 but follows the usual priority rules (O > C).
Indeed, we show that the photorearrangements of 7
proceed in good yields with nearly exclusive retention of
configuration at phosphorus, i.e., cis-7 f cis-8 and trans-7
f trans-8 (eq 1). This new finding is in accord with the
Allyl phosphites whose double bonds are not phenyl-
substituted also undergo regiochemically analogous,
triplet xylene-sensitized photorearrangements to allyl-
phosphonates.3 Relative quantum yields, φP, are strongly
(1) (a) Bentrude, W. G.; Lee, S.-G.; Akutagawa, K.; Ye, W.-Z.;
Charbonnel, Y J . Am. Chem. Soc. 1987, 109, 1577. (b) Bentrude, W.
G.; Wu, Y. W.; Ganapathy, W.; Baik. W.; Lee, S.-G.; Cambron, R. T.;
Harris, J . M. Phosphorus, Sulfur, Silicon 1993, 75, 312. (c) Ganapathy,
S.; Cambron, R. T.; Dockery K. P.; Wu, Y.-W.; Harris, J . M.; Bentrude,
W. G. Tetrahedron Lett. 1993, 34, 5987. (d) Bentrude, W. G.; Dockery
K. P.; Ganapathy, S.; Lee, S.-G.; Tabet, M.; Wu, Y.-W.; Cambron, R.
T.; Harris, J . M. J . Am. Chem. Soc. 1996, 118, 6192.
operation of a mode 4 permutation 15 f 16 (Scheme 2),
(2) Minami, T.; Motoyoshiya, J . Synthesis 1992, 333.
(4) (a) Caldwell, R. A.; Zhou, L. J . Am. Chem. Soc. 1994, 116, 2271.
(b) Unett, D. J .; Caldwell, R. A. Res. Chem. Intermed. 1995, 21, 665.
(c) Unett, D. J .; Caldwell, R. A.; Hrncir, D. C. J . Am. Chem. Soc. 1996,
118, 1682.
(3) (a) Bentrude, W. G. Photorearrangements of Allyl Phosphites.
In Phosphorus Chemistry, Developments in American Science, ACS
Symposium Series 486, Walsh, E. N., Griffith, E. J ., Parry, R. W., Quin,
L. D., Eds.; American Chemical Society: Washington D. C., 1992; Ch.
11. (b) Huang, Y.; Bentrude, W. G. Tetrahedron Lett. 1997, 38, 6989.
(5) For
a recent example, see: Bhanthumnavin, W.; Arif, A.;
Bentrude, W. G. J . Org. Chem. 1998, 63, 7753. See also ref 6.
10.1021/jo9918302 CCC: $19.00 © 2000 American Chemical Society
Published on Web 04/04/2000

Notes
J . Org. Chem., Vol. 65, No. 9, 2000 2787
Sch em e 2
Sch em e 3
known from ESR studies6 for cyclic phosphoranyl mono-
radicals. The retentive stereochemistry provides conclu-
sive evidence against a mode 1 permutation (16 f 18,
Scheme 2) that predominates for truly pentacovalent
phosphorus species. Furthermore, Scheme 2 is totally
consistent with arguments1c,d that explain the low quan-
tum yield, φP, for phosphonate formation from 5. Unfor-
tunately, the synthetically interesting triplet triphenylene-
sensitized photorearrangement of the nucleoside-based
2-phenylallyl phosphite 9 is found to produce only low
yields of phosphonate 10.
Cyclic phosphite 11 does not give phosphonates 12 or
13 on m-xylene triplet sensitization nor do the corre-
sponding phosphonates arise from allyl and 2-methylallyl
phosphites 14 and 15. The findings for 11, 14, and 15,
along with those for 1, 5, 7, and 9, are discussed in terms
of the energy diagram of Scheme 3 which unifies for the
first time our understanding of the effects of allyl phos-
phite structure on the quantum efficiencies of these
photorearrangements in terms of a 1,2-diradical π-π*
model.4
Resu lts a n d Discu ssion
Tr ip let -Sen sit ized P h ot or ea r r a n gem en t of cis-
a n d tr a n s-7. Phosphite 7 was prepared under conditions
of kinetic control (trans/cis ) 92/8) and then partially
equilibrated thermally to give a second nonequilibrium
ratio of diastereomers (trans/cis ) 57/43). Thoroughly
deoxygenated dilute solutions of phosphite 7, containing
triphenylene as triplet sensitizer, were irradiated in
Pyrex tubes with light from the 350 nm lamps of a
Rayonet reactor (Table 1).
(6) Five-coordinate phosphorus modes of permutation have been
defined (Musher, J . I. J . Chem. Educ. 1974, 51, 94). (a) Griller, D.;
Roberts, B. P. J . Chem. Soc, Perkin Trans. 2 1973, 1416. (b) Dennis,
R. W.; Roberts, B. P. J . Chem. Soc, Perkin Trans. 2 1975, 140. (c)
Davies, A. G.; Parrott, M. J .; Roberts, B. P.; Skowronska, A. J . Chem.
Soc., Perkin Trans. 2 1976, 1154. (d) Cooper, J . W.; Parrot, M. J .;
Roberts, B. P. J . Chem. Soc., Perkin Trans. 2 1977, 730. (e) Roberts,
B. P.; Singh, K. J . Chem. Soc., Chem. Commun. 1979, 980.
The reaction was followed by GC (internal standard
method) to assay the total consumption of the diastereo-
mers of phosphite 7 (cis plus trans) and to quantitate the
amount of each individual diastereomer of 8 formed.
Consumption of the individual cis and trans diastereo-
mers of 7, interconverted on GC, had to be monitored

2788 J . Org. Chem., Vol. 65, No. 9, 2000
Notes
Ta ble 1. Ster eoch em istr y of P h otor ea r r a n gem en t of 7a
was nearly unchanged at 24 and 72 h. (Interestingly, the
diastereomers of 7 are consumed at nearly equal rates.)
Second, the accountabilities of converted phosphite 7, in
terms of phosphonate 8 formed, are not quantitative.
Therefore, for the conclusion regarding stereochemistry
to be valid, the same fraction of side products must be
formed from each diastereomer of 7. The consistency of
results and at several conversions, for two initial dia-
stereomeric ratios of 7, supports this conclusion.
time, cis/trans % conv
% 8
cis/trans 7e cis/trans 8f
h
7
of 7d
formedd
consump
formed
0b,f
24
72
0c,f
24
48
72
8/92
8/92
9/91
43/57
40/60
43/57
44/58
0
32
61
0
39
58
85
0
90
80
0
66
72
69
8/92
9/93
0
8/92
10/90
0
47/53
44/56
43/57
43/57
42/58
44/56
a
b
By 31P NMR vs internal standard (n-PrO)3PO. 0.86 mol of
Scheme 2 depicts for cis-7 a mechanism that accounts
for the observed stereochemistry at phosphorus (16 f
17 f cis-8). Featured is the mode 4 (M4) that was
postulated previously for the triplet-sensitized photore-
arrangement of 5 to trap the otherwise reversibly formed
16.1c,d Equation 2 shows the apical introduction of the
1,2-biradical-like4 styryl moiety to form 1,3-biradical 16.
7, 0.009 M 7. c 1.2 mol of 7 (0.012 M). Total cis plus trans. e Based
on moles of cis- and trans-7 individually consumed. f Calculated
from moles of individual cis and trans isomers of 8 formed. f Both
reactions contained an approximate 0.6 equiv of triphenylene.
d
quantitatively, along with those for the diastereomers of
8, by integration of their 31P NMR signals against
internal standard, (n-PrO)3PO. Alternatively, essentially
identical results were obtained if the product mixture was
concentrated under argon and then oxidized by tert-
BuOOH to convert remaining phosphite to phosphate
which was quantitatively assayed by GC. This oxidation
occurs near-quantitatively with retention of configuration
at phosphorus.7 The results in Table 1 were obtained by
the 31P NMR method. Photolysis in the absence of tri-
phenylene gave no conversion of 7 to 8.
Phosphonates 8 were isolated by chromatography and
identified by comparison to authentic material prepared
independently. The cis and trans identities of the indi-
vidual diastereomers of 7 and 8 were readily assigned.
(The 31P shift of the diastereomer with the RO or R group
positioned axially on phosphorus in its predominant
conformer is invariably more upfield8).
This stereochemistry is analogous to that for the revers-
ible apical addition of alkyl radicals to trialkyl phos-
phites.9 The trigonal bipyramidal geometry of 16 is based
on the known structures of similar phosphoranyl mono-
radicals.10 It is significant that a mode 1 rearrangement
of 16 to give 18 would lead to inversion of configuration
at phosphorus and generation of trans-8 (Scheme 2). The
driving force for 16 f 17 (and 16 f 18) is the known
thermodynamic preference with phosphoranyl mono-
radicals for the oxygen and carbon substituents to be
apical and equatorial, respectively. Presuming that rear-
rangement of 16 precedes â-scission to form cis-8, the
stereochemical results of this study provide further
evidence for the contrasting permutational properties of
phosphoranyl radicals and truly pentacovalent phospho-
rus molecules. The latter exchange ligands via mode 1
processes. However, the possibility that 16 gives cis-8
directly cannot be excluded since we have previously
postulated that 16 f 17 is relatively slow.1c,d
Phosphonate yields in Table 1 are not quantitative but
reasonably good. Workup of a 151-mg scale photoreaction
gave a 54% total isolated yield of a mixture of the two
diastereomers of 8 (99% conversion). The amounts of each
individual diastereomer of 7 consumed are expressed in
Table 1 as percentage ratios of the total amount of 7
consumed (cis/trans 7 consumed). Likewise, the account-
ability yields of each diastereomer of phosphonate 8
formed are given as percentage ratios. The cis/trans ratio
of consumed phosphite is close to the cis/trans ratio of
phosphonate formed in each case. Thus, the photorear-
rangement proceeds with essentially complete retention
of configuration at phosphorus. This is clearly repre-
sented for cis-7 f cis-8 by eq 1. This conclusion, however,
requires that two things be true of these reactions.
First, the diastereomers of 7 must be configurationally
stable under the photoirradiation conditions. Indeed, the
trans/cis ratio of phosphite 7 was unchanged in 96-h at
room-temperature in the absence of UV light. Further-
more, photoreactions on solutions rich in the kinetically
preferred trans-7 (cis/trans ) 8/92) showed no evidence
of interconversion of diastereomers. Thus, the build up
during irradiation of a greater amount of thermodynami-
cally favored cis-7 irradiation than was initially present
was not observed nor was a systematic drift in the ratio
of products with conversion of 7 seen. In fact, the cis/
trans ratios of phosphite diastereomers (8/92 and 43/57)
Tr ip let-Sen sitized P h otor ea r r a n gem en t of Th y-
m id in e-Ba sed Cyclic P h osp h it e 9. The triplet tri-
phenylene-sensitized photolysis of 0.005 M solutions of
phosphite 9 (35/65 cis/trans ratio) with light from the 350
nm lamps of a Rayonet reactor led to 37-41% consump-
tion of the phosphite in 144 h but generated only 7-12%
of phosphonate 10 (quantitative 31P NMR; tri-n-propyl
phosphate internal standard). The individual diastereo-
mers of phosphonate 10, however, can be prepared from
(9) Davies, A. G.; Dennis, R. W.; Roberts, B. P. J . Chem. Soc., Perkin
Trans. 2 1974, 1101. Cooper, J . W.; Roberts, B. P. J . Chem. Soc., Perkin
Trans. 2, 1976, 808.
(10) For the most recent reviews of the chemistry of phosphoranyl
radicals, see: Bentrude, W. G. In The Chemistry of Organophosphorus
Compounds; Hartley, F. R., Ed.; J ohn 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, p 225.
(7) Bentrude, W. G.; Hargis, J . H. J . Am. Chem. Soc. 1970, 92, 7136.
(8) See coverage in: Bentrude, W. G. In Steric and Stereoelectronic
Effects in 1,3,2-Dioxaphosphorinanes; J uaristi, E., Ed.; VCH: New
York, 1995; Ch. 7. Bentrude, W. G. In 31P NMR Spectroscopy; Verkade,
J . G., Quin, L., Eds.; VCH Publishers: New York, 1987; Chapter 11.
Maryanoff, B. A.; Hutchins, R. O.; Maryanoff, C. A. Top. Stereochem.
1979, 11, 187.

Notes
J . Org. Chem., Vol. 65, No. 9, 2000 2789
9 by photoinduced single electron transfer (SET) meth-
ticipated that the same alkene moiety in 11 might lead
to phosphonate formation. Evidently the expected in-
creased lifetime and energy for the π-π* triplet of 114 is
not sufficient to overcome the energetically unfavorable,
highly reversible nature of the cyclization of triplet 11
to form a 1,3-biradical intermediate analogous to 16 and
its slow subsequent mode 1 isomerization.
ods.11 (See Supporting Information.)
Reasonable yields of phosphonate 8 were obtained on
irradiation of 7 over extended periods of time (4-5 days).
The disappointing outcome with cyclic phosphite 9 obvi-
ously results from the highly inefficient nature of its
photorearrangement. To a much greater degree than with
phosphite 7, which has phosphorus in a six-membered
ring, phosphite 9 is converted to unidentified side prod-
ucts that are seen in the 31P NMR spectra obtained
following photolysis. The thymin-1-yl ring of 9 can be
readily activated by triplet sensitizers. However, triplet
triphenylene (ET ) 67 kcal/mol12) is not energetic enough
to transfer energy to 9, based on the reported ET value
(>73 kcal/mol) for thymidine 5′-monophosphate.13 The
reluctance of the π-π* triplet of 9 to rearrange to 10
allows other undefined photoprocesses to take over.
Attem p ted Tr ip let-Sen sitized P h otor ea r r a n ge-
m en ts of Oth er Cyclic P h osp h ites. A deoxygenated
solution in cyclohexane of phosphite 11 (0.011 M), tri-n-
propyl phosphate (internal standard) and 5% (v/v)
m-xylene (sensitizer), irradiated through quartz at 254
nm, showed 89% consumption of 11 after 144 h. Numer-
ous peaks were noted by GC; however, GC/MS failed to
detect the molecular ions of phosphonates 12 or 13 (202
amu). 31P NMR spectroscopy showed no evidence for their
formation. Under the same conditions, phosphites 14 and
15 were consumed and formed many products. However,
no GC/MS or 31P NMR evidence for phosphonate forma-
tion was seen.
En er gy Dia gr a m . The results of the above studies,
along with those from a previous full paper1d and
communication,3b prompted us to attempt to unify our
understanding of the effects of structural change on the
efficiency of these photorearrangements in terms of the
energy diagram of Scheme 3 which is based on a 1,2-
diradical model4 for the π-π* triplet. (Relative energies
in Scheme 3 are approximate.) Such a test of this model
is especially important since no direct spectroscopic
evidence for the intermediates of Schemes 1 and 2 exists.
Effects to be explained include: (1) the negative effect of
placement of phosphorus in relatively small (five-, six-,
or seven-membered) ring; (2) the negative effect of
methyl-substitution at the 1-allyl position; (3) the positive
effect of phenyl or alkyl substitution at the 2-allyl
position; and (4) the positive effect of placement of the
allyl double bond in a five-membered ring. Very signifi-
cantly, Scheme 3 also incorporates the mode 4 permuta-
tion set forth in Scheme 2 that is consistent with the
observed stereochemistry at phosphorus in the present
study.
The “1,2-biradical model” was first proposed by Cald-
well et al.4 to explain reactions emanating from ππ*
alkene triplets. In this view the carbon that bonds to
phosphorus in Scheme 3 to form the triplet 1,3-biradical
(e.g., 4 or 16) functions essentially as a free alkyl
monoradical. As will be seen, the structure-quantum
efficiency results indeed fit quite well to this model.
Shown, for purposes of illustration, are the spiro 1,3-
biradicals I and II, formed from a cyclic phosphite such
as 7 or 11, with phenyl substitution at the 2-allyl position.
For the sake of simplicity, product formation in Scheme
3 is assumed to occur after intersystem crossing to the
So 1,3-biradical. At the ground state 1,3-biradical stage,
the overall efficiency of the photorearrangement is de-
termined by the rate constant for reversion to So phos-
phite in competition with successive mode four (M4)
permutation (I a II) followed by â scission of II to give
allylphosphonate. If the permutation (reversible) and â
scission (irreversible) are both rapid, then allylphospho-
nate formation competes readily with reformation of
ground state (So) allyl phosphite resulting in a relatively
high overall quantum yield, φP.
The failure of phosphites 14 and 15 to give the
corresponding phosphonates is in contrast to the rear-
rangements of acyclic (MeO)2POCH2CHdCH2 and (MeO)2-
POCH2C(Me)dCH2 that generate good yields of (MeO)2P-
(O)CH2CHdCH2 and (MeO)2P(O)CH2C(Me)dCH2.3 This
is consistent with the greatly reduced quantum yields
(φP) encountered for the photorearrangement of phosphite
5 (phosphorus in a six-membered ring) compared to that
for 1.1c,d However, since the relative quantum yield for
photorearrangement of 19 was 15 times3b that of the
acyclic phosphite (MeO)2POCH2CHdCHMe, it was an-
For dimethyl phosphites (e.g., 1 and the dimethyl
analogues of 11, 14, and 15) methyl substitution at the
alkene terminus that bonds to phosphorus weakens the
bond to phosphorus in the triplet and in the singlet 1,3-
biradical increasing its ease of reversal to triplet (T1) or
ground state (So) phosphite (decreased φP3b). Phenyl (I
and II),1c,d methyl, or ring carbon substituents3b at the
2-allyl position stabilize the 1,3-biradical and, further-
more, increase the rate of â scission of the So biradical to
product allyl phosphonate by stabilizing the double bond
(11) See Supporting Information for preparation of 10 and separa-
tion of its diastereomers. (Sopchik, A. E.; Hager, D. H. J . Org. Chem.
2000, 65, 2778-2785.)
(12) Zander, M. Z. Naturforsch., Teil A 1984, 39A, 1145.
(13) Gut, I. G.; Wood, P. D.; Redmond, R. W. J . Am. Chem. Soc. 1996,
118, 2366.

2790 J . Org. Chem., Vol. 65, No. 9, 2000
Notes
on a 20 m × 0.25 mm fused silica capillary column (RSL-150)
with tri-n-propyl phosphate as internal standard.
being formed (increased φP). Cyclization to the triplet 1,3-
biradical also is favored when the alkene moiety is in a
small ring (19). The increased lifetime4 of the planar
π-π* T1 optimizes the probability of cyclization to the
triplet 1,3-biradical, rather than reversion to the ground
state (So phosphite), and reversion to the π-π* T1 is less
energetically favorable (increased φP3b).
The permutation I a II pictured by the structures in
Scheme 3 is for a spiro 1,3-biradical such as that from 5
or 7 which is postulated to be slow,1c,d as demonstrated
experimentally for spiro phosphoranyl monoradicals;6a,b
its barrier is represented by the dotted line. A low
quantum yield for 5 f phosphonate formation results (φP
) 0.002-0.003).1c,d Nonetheless, if phosphite 7 does
undergo photo-Arbuzov rearrangement via initial forma-
tion of triplet biradical cis-7(T1), eq 2, followed by
permutational isomerization, the process analogous to I
f II (16 f 17) is consistent with, though not required
by, the observed stereochemistry (cis-7 f cis-8 and
trans-7 f trans-8.)
For monocylic (i.e., nonspiro) phosphoranyl radicals,
ESR shows the M4 process to be much more rapid6 (solid
line of Scheme 3). This allows, for example, the efficiency
of trapping by permutational isomerization of 20 (formed
from 1), in competition with reformation of 1, to be
greatly improved via an efficient cascade analogous to I
(CH2 ap) f II (CH2 eq) f allylphosphonate (For phos-
phite 2, φP ) 0.2-0.3)1c,d
P h otoin d u ced Tr ip let En er gy Tr a n sfer In itia ted Rea r -
r a n gem en t of 5-ter t-Bu tyl-2-(2-p h en yla llyloxy)-1,3,2-d ioxa -
p h osp h or in a n e (7). For example, under argon by glovebag
techniques in a 100.0 mL volumetric flask, tri-n-propyl phos-
phate (110.4 mg, 0.493 mmol), 8/92 cis/trans phosphite 7 (151.4
mg, 0.857 mmol), and triphenylene (114.0 mg, 0.500 mmol) were
diluted to volume with acetonitrile. (The cis/trans ratio of
diastereomers of 7 had been determined by 31P NMR.) Three
quartz tubes (13 mm × 80 mm) were flushed with argon, capped
with rubber septa, and transferred to the glovebag. Via syringe,
5.0 mL of the reaction solution were added to each tube. The
tubes were then irradiated for 96 h with light from the 350 nm
lamps of a Rayonet photochemical reactor. Reactions were
directly monitored by syringe sampling at 0, 24, 48, 72, and 96
h. At each conversion, the moles of each diastereomer of
unreacted 7 were determined by 31P NMR, by reference to the
internal standard, on aliquots of reaction solution transferred
to an NMR tube in a glovebag under argon. Moles of the
individual diastereomers of 8 formed were determined in the
same way. Very rarely it was necessary to concentrate the
samples under a slow stream of argon prior to quantitation by
31P NMR. The yields of the individual diastereomers of 8 (tri-
n-propyl phosphate internal standard) also could be determined
by GLC on a sample carefully concentrated under argon, using
a 20 m × 0.32 mm RSL-150 capillary column. In that event the
individual diastereomers of remaining phosphite 7 were quan-
titated by GC in by slow syringe addition under argon to the
reaction concentrate of approximately 1.1 equiv of a of tert-
BuOOH in an organic solvent to convert 7 to the phosphate.
Close agreement was found in results determined by the two
methods. The data of Table 1 are based on the 31P NMR method.
The combined yield of both isomers of 8 at >99% conversion of
7 (96 h) was 55% (GC). The reaction mixture was purified by
radial chromatography (100% ethyl acetate) to give 81 mg (0.28
mmol, 54% isolated yield) of 5-tert-butyl-2-oxo-2-(2-phenylallyl)-
1,3,2-dioxaphosphorinane (8) as a mixture of diastereomers. The
structures of isolated 8 were confirmed by comparison of their
spectral parameters to those of independently synthesized 8 (see
Supporting Information).
P h otoin d u ced Tr ip let En er gy Tr a n sfer In itia ted Rea r -
r a n gem en t of 2-(Cyclop en t en -1-ylm et h oxy)-1,3,2-d ioxa -
p h osp h or in a n e (11). By a similar glovebag procedure under
argon, a 50.00 mL cyclohexane solution of m-xylene (2.5 mL),
tri-n-propyl phosphate (118 mg, 0.530 mmol), and 2-(cyclopen-
tenyl-1-methoxy)-1,3,2-dioxaphosphorinane (119 mg, 0.592 mmol)
was dispersed in three quartz tubes (13 mm × 80 mm). The
tubes were irradiated for 144 h with 254 nm light from a Rayonet
photochemical reactor. The reactions were sampled by GC at 0,
12, 24, 48, 96, and 144 h. Consumption of phosphite: 48 h (13%);
96 h (70%); 144 h (89%) (20 m × 0.32 mm RSL-150 capillary
column). An array of product peaks was detected by GC, none
of which contained a 202 amu characteristic M+ of the phos-
phonates (GC/MS) 12 or 13.
Exp er im en ta l Section
P r ep a r a tion of Com p ou n d s. Anhydrous solvents were
obtained by distillation under nitrogen: diethyl ether from
sodium/benzophenone; methylene chloride from calcium hydride
under argon; acetonitrile from calcium hydride. Photoreactions
in C6D6 and CDCl3 were deoxygenated by purging with argon.
Cyclohexane was spectral grade. Reagents were from Aldrich
Chemical Co. unless otherwise specified. Tri-n-propyl phosphate
was distilled prior to use. Triphenylene was recrystallized from
ethanol. Triethylamine and thymidine were used as received.
2-Phenylallyl alcohol,14 5-tert-butyl-2-chloro-1,3,2-dioxaphospho-
rinane,7 and thymidine 3′,5′-cyclic N,N-dimethylaminophos-
phoramidite15 were prepared by literature procedures. Unless
otherwise stated, distillations were performed with a short-path
apparatus. Radial chromatography employed 60 PF254 silica gel
containing gypsum (EM Science). Column chromatography was
performed on 60-200 mesh silica gel (EM Science).
P h otoin d u ced Tr ip let En er gy Tr a n sfer In itia ted Rea r -
r a n gem en ts of 5-ter t-Bu tyl-2-a llyloxy-1,3,2-d ioxa p h osp h o-
r in a n e (14) a n d 5-ter t-Bu t yl-2-(2-m et h yla llyloxy)-1,3,2-
d ioxa p h osp h or in a n e (15) were carried out in same way as
that for 11. (See Supporting Information.)
Tr ip let En er gy Tr a n sfer In itia ted Rea r r a n gem en t of
Th ym id in e Cyclic 2-P h en yla llyl-3′,5′-p h osp h ite (9). A 50.0
mL methylene chloride solution of triphenylene (55.1 mg, 0.241
mmol), tri-n-propyl phosphate (119.7 mg, 0.534 mmol), and
thymidine cyclic 2-phenylallyl-3′,5′-phosphite, 9 (101.9 mg, 0.252
mmol), was dispersed under argon in 1.5 mL portions into four
Pyrex NMR tubes (5 mm) fitted with 10/30 ground glass joints.
The tubes were degassed on a vacuum line by four freeze-
pump-thaw cycles (0.02 mmHg), flame sealed, and then irradi-
ated for 144 h with light from the 350 nm lamps of a Rayonet
photochemical reactor. The percent conversion, yields, and
diastereomer ratios were determined by 31P NMR by comparing
the integrations of the starting material and the product cyclic
phosphonates, in reference to the internal standard (tri-n-propyl
phosphate). The yield of phosphonate 10 was in the range of
7-12% based on converted starting phosphite 9 with conversions
ranging from 37 to 41%.
P h ysica l Meth od s. Melting points are uncorrected. J values
given in the 1H NMR spectral data refer to proton-proton
coupling unless otherwise stated. A 60 s repetition rate was
employed when monitoring the photoreactions by 31P NMR to
ensure the accuracy of the integrations. Microanalyses were
performed by Atlantic Microlab, Inc., Norcross, GA. GC-EIMS
(70 eV) analyses utilized a 30 m × 0.25 fused silica capillary
column. Reported intensities are percentages of the base peak
intensity. Other low resolution EIMS (70 eV) as well as HRMS
(EI, 70 eV) measurements utilized a standard inlet system.
FABMS LSIMS measurements utilized a cesium ion gun. The
GLC yields were determined with a flame ionization detector
(14) Umbreit, M. A.; Sharpless, K. B. J . Am. Chem. Soc. 1977, 99,
5526.
(15) Bentrude, W. G.; Khan, M. R.; Saadein, M. R.; Sopchik, A. E.
Nucleosides Nucleotides 1989, 8, 1359.

Notes
5-ter t-Bu tyl-2-oxo-2-(2-p h en yla llyl)-1,3,2-d ioxa p h osp h o-
r in a n e (8) was routinely prepared from the Arbuzov reaction
of 5-tert-butyl-2-methoxy-1,3,2-dioxaphosphorinane.15 (See Sup-
porting Information for details).
J . Org. Chem., Vol. 65, No. 9, 2000 2791
3
J ) 9.6 Hz, J HP ) 8.7 Hz, 1 H), 4.14 (dddd, J ) 9.2 Hz, J ) 6.4
4
Hz, J ) 9.5 Hz, J HP ) 0.9 Hz, 1 H), 4.35 (ddd, J ) 6.4 Hz, J )
3
3
9.6 Hz, J HP ) 1.9 Hz, 1 H), 4.79 (d, J HP ) 10.1 Hz, 2 H), 5.38
(d, 2J ) 1.2 Hz, 1 H), 5.54 (d, 2J ) 1.2 Hz, 1 H), 6.09 (dd, J ) 4.6
Hz, J ) 6.6 Hz, 1 H), 7.01 (d, J ) 1.2, 1 H), 7.22-7.62 (m, 6 H)
10.0 (bs, 1 H); 13C NMR (125.7 MHz, CDCl3): δ 12.75, 36.32,
65.04 (d, 2J CP ) 19.8), 68.13 (d, 2J CP ) 4.6), 70.07 (d, 2J CP ) 4.5),
P r ep a r a tion of 5-ter t-Bu tyl-2-(2-p h en yla llyloxy)-1,3,2-
d ioxa p h osp h or in a n e (7). A solution of 5-tert-butyl-2-chloro-
1,3,2-dioxaphosphorinane15 (33.2 g, 0.168 mol) in 500 mL of
anhydrous ether, maintained at 0 °C, was stirred in a which
had been rinsed with triethylamine and dried at 110 °C. To it
was added, dropwise, a solution of triethylamine (18.7 g 0.185
mol) and 2-phenylallyl alcohol (23.4 g, 0.175 mol) in 150 mL of
dry ether over a 2 h period. The amine salts were filtered away
under argon, and the solvent was removed under reduced
pressure. The colorless residue oil was purified by chromatog-
raphy (radial chromatography, 50% ethyl acetate/hexane) which
resulted in 30.1 g (0.102 mol, 61% yield) of a mixture of two
isomers of phosphite 7; by 31P NMR, 4/96 (cis/trans). cis-7: 31P
NMR (121.4 MHz CDCl3) δ 124.8; 1H NMR (299.9 MHZ CDCl3):
δ 0.89 (s, 9 H), 2.05 (tt, 3J ) 3.9 Hz, 3J ) 12.0 Hz, 1 H), 3.85 (m,
3
73.12 (d, J CP ) 20.2), 82.84, 111.95, 114.66, 126.16, 128.25,
2
128.63, 135.24, 138.70, 144.54 (d, J CP ) 4.3 Hz), 163.94. Anal.
Calcd for C19H21N2O6P: C, 56.47; H, 5.23. Found: C, 56.41; H,
5.23. (Mixture of diastereomers: cis/trans ) 36/64).
P r ep a r a tion of 1-Hyd r oxym eth ylcyclop en ten e. A sus-
pension of aluminum chloride (6.5 g, 0.048 mol) and lithium
aluminum hydride (6.0 g 0.15 mol) in 500 mL of dry diethyl ether
was stirred at 0 °C for 1 h. To it was added, dropwise, a solution
of ethyl 1-cyclopentenecarboxylate16 (20 g, 0.15 mol) in 50 mL
of diethyl ether. After 2 h the reaction was quenched by the
dropwise addition of 12 mL of water. After 1 h the suspension
was filtered, and the organic phase was washed with saturated
NaCl (3 × 100 mL). The filtrate was dried over MgSO4 and
filtered. The solvent was evaporated. The residue was distilled
to give a colorless liquid (9.8 g, 0.097 mol, 70% yield, bp 60-62
°C at 4.0 mmHg [lit.17 bp 60-62 °C at 10 mmHg]): 1H NMR
(299.9 MHz CDCl3) δ 1.86-1.97 (m, 2H), 1.99 (bs, 1H), 2.27-
2.40 (m, 4H), 4.19-4.21 (m, 2H), 5.60-5.63 (m, 1H).
3
2 H), 4.15-4.23 (m, 2 H), 4.75 (d, J PH ) 8.7 Hz, 2 H), 5.46 (bs,
1 H), 5.56 (bs, 1 H), 7.30-7.42 (m, 3 H), 7.49-7.54 (m, 2 H); 13
C
NMR (75.4 MHz CDCl3): δ 27.36, 31.31 (d, 4J CP ) 1.6 Hz), 45.99
3
2
(d, J CP ) 4.6 Hz), 61.62, 64.67 (d, J CP ) 19.6 Hz), 113.79,
3
126.25, 127.99, 128.45, 138.45, 144.95 (d, J CP ) 5.1): trans-7:
31P NMR (121.4 MHz CHCl3) δ 132.4; 1H NMR (299.9 MHz
CDCl3): δ 1.01 (s, 9 H), 1.72-1.79 (m, 1 H), 3.87-4.02 (m, 2 H),
P r epar ation of 2-(Cyclopen ten -1-ylm eth oxy)-1,3,2-dioxa-
p h osp h or in a n e (11). Under an argon atmosphere, a solution
of 2-chloro-1,3,2-dioxaphosphorinane (500 mg, 2.55 mmol) in 50
mL of freshly distilled diethyl ether in a flask which had been
rinsed with triethylamine and dried at 110 °C, was stirred at 0
°C. A solution of triethylamine (309 mg, 3.06 mmol) and
1-hydroxymethylpentene (275 mg, 2.80 mmol) in 25 mL of dry
diethyl ether was added dropwise over 1 h. The amine salts were
filtered away under argon, and the solvent was removed under
reduced pressure. The clear liquid residue was purified by
chromatography (Chromatotron, 30% ethyl acetate-hexane) to
give 421 mg (1.63 mmol, 64% yield) of 11, a colorless oil: 31P
3
4.27-4.35 (m, 2 H), 4.79 (d, J PH ) 9.0 Hz, 2 H), 5.47 (bs, 1 H),
5.57 (bs, 1 H), 7.30-7.42 (3H), 7.49-7.54 (m, 2H); 13C NMR (75.4
MHz CDCl3): δ 28.32, 32.35, 45.05 (d, 3J CP ) 8.25 Hz), 61.15 (d,
2J CP ) 2.0 Hz), 64.75 (d, 2J CP ) 20.1 Hz), 113.85, 126.21, 128.00,
3
128.50, 138.44, 144.98 (d, J CP ) 4.9 Hz). GC EIMS (70 eV) m/z
(relative intensity) 294 [M+] (100), 279 [M - CH3]+ (1), 237 [M
- tert-Bu] (28); C16H23O3P; HRMS [M]+ (calcd) 294.13848, (obsd)
294.13970. Anal. Calcd for
C16H23O3P: C, 65.33; H, 7.88.
Found: C, 65.06; H, 7.89. (Mixture of isomers: cis/trans ) 4/96)
P r ep a r a tion of Th ym id in e Cyclic 2-P h en yla llyl 3′,5′-
P h osp h it e (9). A solution of thymidine 3′,5′-cyclic N,N-di-
methylaminophosphoramidite16 (0.30 g, 0.94 mmol), 1-H tetra-
zole (70 mg), and freshly distilled methylene chloride (30 mL)
was stirred under argon while 2-phenylallyl alcohol (134 mg,
1.0 mmol) was added via syringe at a rate of one drop every 5 s.
The reaction was stirred for a further 12 h. The solvent was
removed under reduced pressure (0.05 mmHg). The residue was
dissolved in 50% ethyl acetate-diethyl ether and passed through
a column of 60-200 silica gel under argon. The first 250 mL of
eluent was collected, and the solvent was removed under reduced
pressure affording 253 mg (0.62 mmol, 66% yield) of a white
solid; phosphite 9, cis/trans ) 36/64 (cis/trans), 31P NMR. cis-9:
31P NMR (202.4 MHz CDCl3) δ 122.97; 1H NMR (499.8 MHz
CDCl3):16 δ 1.91 (d, J ) 1.3 Hz, 3 H), 2.09 (ddd, J ) 2.3 Hz, J )
8.1 Hz, J ) 13.4 Hz, 1 H), 2.33 (ddd, J ) 8.6 Hz, J ) 11.0 Hz,
J ) 13.4 Hz, 1 H), 3.55 (ddd, J ) 9.3 Hz, J ) 10.2 Hz, J ) 4.6
1
NMR (121.4 MHz C6D6) δ 133.5; H NMR (299.9 MHz C6D6) δ
0.71-0.81 (m, 1 H), 1.71-1.79 (m, 2 H), 1.95-2.15 (m, 1 H),
2.19-2.30 (m, 4 H), 3.38-3.49 (m, 2 H), 4.21-4.32 (m, 4 H),
5.61-5.65 (m, 1 H); 13C NMR (75.4 MHz C6D6): δ23.40, 28.69
3
(d, J CP ) 5.2 Hz), 32.46 (CH2CH2CH2), 32.82, 59.35, 62.07 (d,
3
2J CP ) 19.6 Hz), 127.26, 141.73 (d, J CP ) 5.6 Hz). GC EIMS
(EI-70 eV) m/z (relative intensity) 202 [M]+ (2), 201 [M - 1]+
(2), 81 [M - C3H6PO3](36), 80 [M - C3H7PO3] (100); EI HRMS
C9H15O3P [M]+: Anal. Calcd for C9H15O3P: C, 53.45; H, 7.48.
Found: C, 53.45; H, 7.55 (mixture of isomers).
5-ter t-Bu tyl-2-allyloxy-1,3,2-dioxaph osph or in an e (14) an d
5-t er t -b u t y l-2-(2-m e t h y la lly lo x y )-1,3,2-d io x a p h o s p h o -
r in a n e (15) were prepared by the method used for 7 and 11.
(See Supporting Information.)
Hz, 1 H, H4′), 4.14 (dddd, J ) 11.0 Hz, J ) 8.1 Hz, J ) 9.3 Hz,
Ack n ow led gm en t. Support of this research by
grants from the National Science Foundation and Public
Health Service (GM) is gratefully acknowledged. We
thank Dr. Alan E. Sopchik for separation of the indi-
vidual isomers of phosphonate 10 and assignments of
1H and 13C NMR spectral parameters.
3
3J HP ) 1.4 Hz, 1 H), 4.21 (ddd, J ) 4.6 Hz, J ) 9.2 Hz, J HP
)
3
9.6 Hz, 1 H), 4.25 (ddd, J ) 10.2 Hz, J ) 9.2 Hz, J HP ) 2.5 Hz,
1 H), 4.77 (d, J HP ) 9.6 Hz, 2 H), 5.39 (d, 2J ) 1.0 Hz, 1 H),
3
5.52 (d, 2J ) 1.0 Hz), 6.07 (dd, J ) 8.6 Hz, J ) 2.3 Hz, 1 H),
6.79 (d, J ) 1.3 Hz), 7.22-7.62 (m, 5 H) 10.0 (bs, 1 H); 13C NMR
2
(125.7 MHz CDCl3): δ 12.82, 36.32, 65.68 (d, J CP ) 20.2 Hz),
2
3
66.38 (d, J CP ) 3.8 Hz), 68.66 (C(3’)), 75.03 (d, J CP ) 7.0 Hz),
81.95, 111.70, 115.02, 126.26, 128.25, 128.57, 134.89, 138.10,
Su p p or tin g In for m a tion Ava ila ble: Procedures for the
preparation and separation of the diastereomers of phospho-
nates 8 and 10, the attempted triplet-sensitized rearrange-
ments of 14 and 15, and the preparations of 14 and 15. This
material is available free of charge via the Internet at
http://pubs.acs.org.
2
144.59 (d, J CP ) 4.3 Hz), 150.49 (C(2)), 163.94 (C(4)). trans-9:
31P NMR (CHCl3) δ 129.74: 1H NMR (499.8 MHz CDCl3): δ 1.94
(d, J ) 1.2 Hz, 3 H), 2.27 (m, 2 H), 4.02 (dddd, J ) 9.9 Hz, J )
3
8.9 Hz, J ) 9.2 Hz, J HP ) 1.1 Hz, 1 H), 4.03 (ddd, J ) 9.5 Hz,
(16) Kuivila, H. G.; Patnode, P. P. J . Organomet. Chem. 1977, 129,
145.
J O9918302