Photo-Arbuzov rearrangements of 1-arylethyl phosphites: Stereochemical studies and the question of radical-pair intermediates

Article abstract of DOI:10.1021/jo001545e

The direct UV irradiation of the 1-arylethyl phosphites 7, 8, and 9 was carried out in acetonitrile, benzene, and cyclohexane, as was the triphenylene-sensitized reaction of 9. Dimethyl 1-phenylethyl phosphite, 7, gives the photo-Arbuzov rearrangement pro

Full text of DOI:10.1021/jo001545e

980
J . Org. Chem. 2001, 66, 980-990
P h oto-Ar bu zov Rea r r a n gem en ts of 1-Ar yleth yl P h osp h ites:
Ster eoch em ica l Stu d ies a n d th e Qu estion of Ra d ica l-P a ir
In ter m ed ia tes
Worawan Bhanthumnavin and Wesley G. Bentrude*
Department of Chemistry, University of Utah, 315 South 1400 East, Salt Lake City, Utah 84112
bentrude@chemistry.utah.edu
Received October 31, 2000
The direct UV irradiation of the 1-arylethyl phosphites 7, 8, and 9 was carried out in acetonitrile,
benzene, and cyclohexane, as was the triphenylene-sensitized reaction of 9. Dimethyl 1-phenylethyl
phosphite, 7, gives the photo-Arbuzov rearrangement product, dimethyl 1-phenylethylphosphonate
(10), in 67% average yield and minor amounts (2%) of 2,3-diphenylbutane (11a ) in quantum yields
of 0.32 and 0.02, respectively. The photorearrangement of optically active, predominantly (R)-1-
phenylethyl phosphite 7 (R/S ) 97/3; 94% ee), at 35-40 °C proceeds with a high degree of
stereospecificity at the stereogenic migratory carbon to give predominantly (R)-10 (R/S ) 86/14,
72 ( 2% ee). Use of the nitroxide radical trap TEMPO affords phosphonate 10, presumably all
cage product, from predominantly (R)-7 (R/S ) 97/3; 94% ee) in 64% yield (80% ee, R/S ) 90/10).
By contrast, the 1-(4-acetylphenyl)-ethyl phosphite, predominantly (S)-8 (S/R ) 98/2, 96% ee), on
direct irradiation gives the corresponding phosphonate (12) in only 20% yield along with dimer
11b in 40% accountability yield. Phosphonate 12 is nearly racemic (R/S ) 52/48). Direct irradiation
of predominantly (R)-9 (R/S ) 98/2, 96% ee), a 1-(1-naphthyl)ethyl phosphite, results in a product
distribution similar to that from predominantly (R)-7, but with a somewhat higher degree of
retention of configuration in the product phosphonate 13 (R/S ) 93/7, 86 ( 3 ee). By contrast, the
triplet triphenylene-sensitized photorearrangement of largely (R)-9 (R/S ) 98/2, 96% ee) leads to
product distributions similar to those from direct irradiation of predominantly (S)-8 and is
accompanied by almost total loss of stereochemistry in its product phosphonate, 13 (R/S ) 51/49).
The partial loss of stereochemistry on direct irradiation of 7 and 9 provides evidence for radical
pair formation. Furthermore, these stereochemical results are diagnostic of the multiplicity of the
initial radical pair formed. Values for kcomb/krot for the proximate free radical pairs from 7 and 9,
derived experimentally, are severalfold larger than those for the proximate singlet pair from Ph2Cd
CdN-CHPhMe, corrected to 35 °C. The possibility that kcomb is increased for the pairs from 7 and
9 is proposed.
In tr od u ction
substituted and on whether the photoreaction of the
1-naphthylmethyl phosphite is triplet sensitized. These
The photo-Arbuzov rearrangements of arylmethyl phos-
phites (1) to the corresponding arylmethylphosphonates
(2) upon direct irradiation with UV light have been
reported by this laboratory1-3 and found to be syntheti-
cally useful (Ar ) phenyl).4,5 Product,3 CIDNP,6 and
CIDEP7 studies of 1 (Ar ) Ph, p-MeCOC6H6CH2, 1-naph-
thylmethyl) show that these rearrangements proceed
largely via radical pair intermediates [3,4] whose mul-
tiplicities depend on whether the phenyl ring is p-acetyl-
(1) Omelanczuk, J .; Sopchik, A. E.; Lee, S.-G.; Akutagawa, K.;
Cairns, S. M.; Bentrude, W. G. J . Am. Chem. Soc. 1988, 110, 6908-
6909.
(2) Cairns, S. M.; Bentrude, W. G. Tetrahedron Lett. 1989, 30, 1025-
1028.
(3) Ganapathy, S.; Soma Sekhar, B. B. V.; Cairns, S. M.; Akutagawa,
K.; Bentrude, W. G. J . Am. Chem. Soc. 1999, 121, 2085-2096.
(4) Bentrude, W. G.; Mullah, K. B. J . Org. Chem. 1991, 56, 7218-
7224.
radical pairs [3,4] are of general interest as examples of
so-called proximate pairs, relatively unusual intermedi-
ates that are formed without an intervening molecule
(e.g., CO2, N2). A classic example is the radical pair
generated by the thermolysis of Ph2CdCdN-CHPhMe
that combines to form Ph2C(CN)CHPhMe.8 Singlet proxi-
mate pairs are expected to undergo a high degree of
recombination within the solvent cage as shown for 1 (Ar
(5) Mullah, K. B.; Bentrude, W. G. Nucleosides Nucleotides 1994,
13, 127-153.
(6) Koptyug, I. V.; Sluggett, G. W.; Ghatlia, N. D.; Landis, M. S.;
Turro, N. J .; Ganapathy, S.; Bentrude, W. G. J . Phys. Chem. 1996,
100, 14581-14583.
(7) Koptyug, I. V.; Ghatlia, N. D.; Sluggett, G. W.; Turro, N. J .;
Ganapathy, S.; Bentrude, W. G. J . Am. Chem. Soc. 1995, 117, 9486-
9491.
(8) Lee, K.-W.; Horowitz, N.; Ware, J .; Singer, L. A. J . Am. Chem.
Soc. 1977, 99, 2622-2627.
10.1021/jo001545e CCC: $20.00 © 2001 American Chemical Society
Published on Web 01/06/2001

Photo-Arbuzov Rearrangements of 1-Arylethyl Phosphites
Sch em e 1a
J . Org. Chem., Vol. 66, No. 3, 2001 981
a
Pro-R or pro-S orientation of 14.
) Ph,1,2 1-naphthyl3,9) with predominant retention of
configuration if a stereogenic migratory carbon center is
involved (5f 6, direct irradiation, eq 1).10
tes are likely to be efficient photoinitiators of polymeri-
zation of alkenes, especially electron-rich ones.
We report here investigations of the stereochemistry
of certain photo-Arbuzov rearrangements for which pro-
duction of predominately triplet or, alternatively, singlet
pairs can be controlled. To attempt to demonstrate a
correlation between stereochemistry at the migratory
carbon and the multiplicity of the radical pair, the
optically active phosphites (R)-2-(1-phenylethyloxy)-1,3,2-
dioxaphosphorinane ((R)-7), (S)-2-(1-(4-acetylphenyl)-
ethyloxy)-1,3,2 dioxaphosphorinane ((S)-8), and dimethyl
(R)-1-(1-naphthyl)ethyl phosphite ((R)-9) were employed.
Furthermore, the dimethoxyphosphinoyl radicals (3)
from photolysis of the arylmethyl phosphites (1) are
closely related structurally to diphenylphosphinoyl radi-
cals [Ph2P(O)•]. The latter are generated on ultraviolet
irradiation of acylphosphine oxides which have been used
extensively as photoinitiators of radical polymerization.11
Diarylphosphinoyl radicals have been shown to be highly
reactive in radical addition to alkenes and in abstraction
reactions12,13 with rates dependent on radical structure.
The demonstrated6,7 generation from certain 1 of the
highly pyramidal (31P hyperfine splitting constant 697
G7,14), and therefore highly reactive,12,13 dimethoxyphos-
phinoyl radicals (3) suggests that 1 and related phosphi-
A generalized picture of this approach is seen in Scheme
1 for singlet pairs [3,14] from 7 and 9 and in Scheme 2
for triplet pairs [3,14].
(9) Bentrude, W. G.; Bhanthumnavin, W.; Ganapathy, S. Phospho-
rus, Sulfur, Silicon 1996, 109-110, 537-540.
(10) Bhanthumnavin, W.; Bentrude, W. G.; Arif, A. M. J . Org. Chem.
1998, 63, 7753-7758.
Indeed, the results reported provide evidence for dis-
crete singlet radical pair intermediates in the direct
irradiation of phosphites 7 and 9. Furthermore, it is
demonstrated that the stereochemistry at migratory car-
bon can be correlated with presumed excited-state mul-
tiplicity and that of the radical pairs from optically active
7-9 and thus provides a useful assignment criterion,
along with product distributions and CIDNP and CIDEP
measurements. Indeed, these systems are ideal for
measurement of the competition between cage combina-
(11) For recent reviews see Dietliker, K. In Radiation Curing in
Polymer Science and Technology; Fouassier, J .-P., Rabek, J . F., Eds.;
Elsevier Applied Science: New York, 1993; Vol. II, p 155. Schnabel,
Wl. In Laser in Polymer Science and Technology: Applications;
Fouassier, J . P.; Rabek, J . F., Eds.; CRC: Boca Raton, 1991; Vol. 2, p
95.
(12) Sluggett, G. W.; McGarry, P. F.; Koptyug, I. V.; Turro, N. J . J .
Am. Chem. Soc. 1996, 118, 7367-7372.
(13) J ockusch, S.; Turro, N. J . J . Am. Chem. Soc. 1998, 120, 11773-
11777.
(14) Davies, A. G.; Griller, D.; Roberts. B. P. J . Am. Chem. Soc. 1972,
94, 1782. Roberts, B. P.; Singh, K. J . Organometallic Chem. 1978, 159,
31.

982 J . Org. Chem., Vol. 66, No. 3, 2001
Bhanthumnavin and Bentrude
Sch em e 2a
a
Indicates pro-S or pro-R orientation.
Ta ble 1. P r od u ct Yield s a n d Ster eoch em ica l Resu lts for
tion (kcomb) and rotation (krot), because we are able to show
that combination of 3 and 14 to reform phosphite does
not occur. For phosphite 7, use of the radical scavenger
TEMPO allows estimates to be made of relative rate
th e P h otor ea ction of 7 (R/S ) 97/3)a
accountability
yield (%)b
conversion
of 7 (%)
R/ S ratio
of 10d
11a c
constants for cage combination (kcomb) and rotation (krot
)
solvent
10
(See Scheme 1) including a very high kcomb/krot value of
11. A minimum value for kcomb/ krot of 17 for the radical
pair from phosphite 9 is estimated. Both are considerably
larger that the same ratio (2.5, CCl4) reported for the
proximate radical pair from Ph2CdCdN-CHPhMe.8
acetonitrile
acetonitrile
acetonitrile
cyclohexane
cyclohexane
cyclohexane
benzene
25
37
78
27
40
57
15
24
48
64
69
63
63
62
59
67
70
66
4.4
4.6
4.0
3.8
4.2
3.8
4.0
3.8
4.2
86/14
88/12
87/13
88/12
86/14
87/13
86/14
85/15
84/16
Resu lts
benzene
benzene
P r ep a r a tion of 7-9. Isola tion of 10-13. Deter m i-
n a tion of En a n tiom er ic Ra tios. Highly pure phos-
phites 7-9 were prepared in routine fashion from the
requisite phosphoramidites and either the optically active
alcohols of known configuration and high optical purity
or the racemic alcohols. Irradiation of racemic 7-9
afforded phosphonates 10, 12, and 13 and radical dimer
11 that were isolated routinely and characterized spec-
a
b
Ca. 0.01 M 7; Rayonet reactor at 35-40 °C. By GC analysis.
Based on the amount of consumed 7. c Yield of 11a doubled to
account for the stoichiometry of its formation from 7. Determined
by chiral HPLC on a CHIRALCEL OD column.
d
closely to those of the starting alcohols. Unfortunately,
the phosphate enantiomers from 8 were unresolved on
chiral HPLC. Unreacted optically active 7 and 9, at
various photoconversions, were similarly oxidized, puri-
fied, and analyzed on the chiral HPLC column to deter-
mine the R/S ratio of unreacted phosphite which re-
mained unchanged over the period of the reaction. (See
tabulated data in the Supporting Information).
Phosphonates 10, 12, and 13, formed on photolysis of
optically active 7-9, were isolated by HPLC, and their
R/ S ratios were determined on a CHIRALCEL OD
HPLC column. The predominant enantiomer was as-
signed the R configuration for 10 and 13 and S config-
uration for 12, based on the primarily retentive stereo-
chemistry determined for the photoreaction of eq 1 (X-
ray crystallography)10 and in an earlier, preliminary
study of the photorearrangement of (R)-7.1 Yields of
photolysis products (Table 1) were determined by quan-
titative GC measurements (authentic materials cali-
brated versus an internal standard). It is important to
note that over a wide range of phosphite conversion no
variation in the stereochemical results was noted.
troscopically and by elemental analysis. To determine the
ratio of enantiomers (R/ S), each optically active phos-
phite was oxidized to the phosphate by tert-butyl hydro-
peroxide. This reaction is known to be near-quantitative
and to occur with retention of configuration at phospho-
rus.15 The phosphates were purified by HPLC followed
by HPLC analysis on a CHIRALCEL OD column and S
enantiomers baseline-resolved). The ratios of phosphate
enantiomers (R/ S) from phosphites 7 and 9 corresponded
P h otolysis of Op tica lly Active P h osp h ite 7. Phos-
phite 7 (ca. 0.01 M) in deoxygenated acetonitrile, benzene,
and cyclohexane was irradiated at 35-40 °C through
quartz at 254 nm (Rayonet reactor). In Table 1 are
recorded phosphonate yields, determined by GC, and
(15) Bentrude, W. G.; Hargis, J . H. J . Am. Chem. Soc. 1970, 92,
7136-7144.

Photo-Arbuzov Rearrangements of 1-Arylethyl Phosphites
J . Org. Chem., Vol. 66, No. 3, 2001 983
Ta ble 2. P r od u ct Yield s a n d Ster eoch em ica l Resu lts for
Ta ble 3. P r od u ct Yield s a n d Ster eoch em ica l Resu lts
Obta in ed fr om Dir ect Ir r a d ia tion a n d
Tr ip h en ylen e-Sen sitized P h otor ea ction s of P h osp h ite 9
(R/S ) 98/2)a
th e P h otor ea ction of 8 (S/R ) 98/2)a
accountability
yield (%)b
conversion
of 8 (%)
S/ R ratio
of 12d
accountability
yield (%)b
solvent
12
11bc
conversion
of 9 (%)
R/ S ratio
of 14d
acetonitrile
acetonitrile
acetonitrile
cyclohexane
cyclohexane
cyclohexane
18
40
97
20
55
93
20
19
19
19
20
18
42
38
36
40
38
40
52/48
51/49
50/50
51/49
54/46
52/48
conditions
solvent
14
11cc
direct hν acetonitrile
direct hν acetonitrile
direct hν acetonitrile
direct hν cyclohexane
direct hν cyclohexane
direct hν cyclohexane
14
39
92
15
36
95
14
45
92
65
72
60
69
63
61
21
20
19
4.6
4.6
3.8
2.2
2.0
3.4
40
91/9
92/8
94/6
94/6
95/5
a
b
Ca. 0.01 M 8. Rayonet reactor at 35-40 °C. By GC analysis.
92/8
Based on the amount of consumed 8. c Yield of 11b doubled to
account for the stoichiometry of its formation from 8. Determined
sens.
sens.
sens.
acetonitrile
acetonitrile
acetonitrile
52/48
51/49
51/49
d
40
42
by chiral HPLC on a CHIRALCEL OD column.
a
Ca. 0.01 M 9. 0.02 M triphenylene. Rayonet reactor at 35-40 °C.
stereochemical data obtained by HPLC (CHIRACEL)
from isolated phosphonate 10 at various conversions.
Major amounts of 10 (60-70%) are seen. Rather small
quantities of dimer 11a (2%) are found which result from
random radical coupling (Scheme 1) of about 4% of the
theoretically formed 1-phenylethyl radicals which escape
from the original solvent cage. From 7 with R/S ratio 97/
3, the enantiomeric composition (R/S) of product phos-
phonate 10 (chiral HPLC) ranges from 84/16 to 88/12 (73
( 2% ee, avg error). The quantum efficiencies of the
formation of phosphonate 10 and dimer 11a were deter-
mined in acetonitrile to be 0.32 and 0.02, respectively.
P h otor ea r r a n gem en t of Op tica lly Active 8. Direct
irradiation of the ketone chromophore (n-π*) of optically
active phosphite 8 (ca. 0.01 M, freeze-pump-thaw
degassed, Pyrex tubes) was carried out with light from a
450 W medium-pressure mercury lamp, filtered through
a uranium filter sleeve (λ > 320 nm). Table 2 shows the
corresponding phosphonate 12 to be formed in ap-
proximately 20% yield in both acetonitrile and cyclohex-
ane. Approximately 40% of the 1-arylethyl radicals
potentially formed on direct irradiation of 8 are accounted
for as dimer 11b (20% yield; see Scheme 2 to be discussed
later). Additional products were detected and identified
by GC and GC/MS, but not quantified. These materials
are analogous to those found in the reaction of p-
acetylbenzyl dimethyl phosphite3,7 and are derived from
random free radical coupling. Product phosphonate 12
from optically active 8 (S/R ) 98/2) is seen to be only
slightly nonracemic at carbon (average S/R ratio, 51/49).
Dir ect Ir r a d ia tion of Op tica lly Active 1-Na p h -
th yleth yl P h osp h ite 9. Phosphite 9 (ca. 0.01 M in
deoxygenated solution) was irradiated at 35-40 °C with
light from 300 nm Rayonet lamps through Pyrex in both
acetonitrile and cyclohexane solvents. In Table 3 product
distributions similar to those from photorearrangement
of phosphite 7 are seen. In both solvents 60-70% yields
of phosphonate 13 resulted. 2,3-Dinaphthylbutane (11c)
(meso/d,l ) 1/1) accounts for 2-5% of the potentially
formed 1-naphthylethyl radicals (Scheme 1). Compared
to phosphite 7, a somewhat higher degree of retention of
configuration at the migratory carbon is observed. Thus
phosphite 9 (R/S ) 98/2) gives the enantiomeric phos-
phonates in R/ S ratios ranging from 91/9 to as high as
95/5 (86 ( 3% ee, avg error).
By GC analysis. Based on the amount of consumed 9. c Yield of
b
11c doubled to account for the stoichiometry of its formation from
9. Determined by chiral HPLC on a CHIRALCEL OD column.
d
wavelengths (> 340 nm), make triphenylene an ap-
propriate triplet sensitizer for the 1-naphthylethyl chro-
mophore (450 W medium-pressure mercury lamp, ura-
nium glass filter, 320 nm cutoff). Phosphite (R)-9 in
deoxygenated acetonitrile (R/S ) 98/2, ca. 0.01 M) also
containing triphenylene (ca. 0.02 M) was sealed in Pyrex
tubes under vacuum. The reaction occurred quite slowly,
presumably because the light was filtered, to give the
results in Table 3. At all conversions, the corresponding
phosphonate is formed in 20 ( 1% yield, along with 2,3-
dinaphthylbutane (11c) in about 20% yield (40% account-
ability of 1-naphthylethyl radicals, potentially generated
from 9). Product phosphonate 13 is formed with a slight
excess of the R isomer present at all three conversions
(chiral HPLC, multiple injections at each conversion). A
control sample of phosphite 9 (ca. 0.01 M) did not undergo
reaction on irradiation in the absence of triphenylene.
P h otolysis of Op tica lly Active P h osp h ite 7 in th e
P r esen ce of TEMP O. The stable radical TEMPO is
often used to scavenge free radical pairs that escape the
initial solvent cage and appeared applicable to the study
of the photolysis of 7. However, irrradiation of TEMPO
can lead to hydrogen abstraction by its excited state.17
This side reaction was minimized by doing the trapping
studies in cyclohexane where GC showed the formation
of minor amounts of abstraction product (TEMPO-H) to
be less than in acetonitrile. Although an excess of
TEMPO was employed, its concentration was held rela-
tively low to minimize the “anti-scavenger effect” of
TEMPO noted by Turro et al.18 A detailed accounts of
radical-trapping approaches used during photolysis of
related arylmethyl phosphites was published earlier.3
Solutions of phosphite 7 (ca. 0.01 M, R/S ) 97/3) were
irradiated in cyclohexane at 254 nm in the presence of a
range of concentrations of TEMPO (0.002-0.022 M).
Yields and enantiomeric ratios of phosphonate 10 are
shown in Table 4. Results obtained in the absence of
TEMPO are also included for comparison. The amounts
of 2,3-diphenylbutane, the product formed from random
coupling of 1-phenylethyl radicals (Scheme 1), decreases
to zero at 0.015 M TEMPO concentration. Concomitantly,
the yield of product phosphonate 10 decreases slightly
Tr ip let-Sen sitized Rea ction s of P h osp h ite 9. The
easily accessible triplet state (T1) of triphenylene (ET
)
67 kcal/mol16), and its UV absorption at relatively long
(17) J ohnston, L. J .; Tencer, M.; Scaiano, J . C. J . Org. Chem. 1986,
51, 2806-2808.
(18) Step, E. N.; Buchachenko, A. L.; Turro, N. J . J . Am. Chem. Soc.
1994, 116, 5462-5466.
(16) Murov, S. L.; Carmichael, I.; Hug, G. L. Handbook of Photo-
chemistry, 2nd ed.; Marcel Dekker: New York, 1993.

984 J . Org. Chem., Vol. 66, No. 3, 2001
Bhanthumnavin and Bentrude
Ta ble 4. P h otolysis of 7 (R/S ) 97/3) in th e P r esen ce of
at concentrations such as those used in the stereochem-
ical studies, were shown by chiral HPLC to have un-
changed R/ S ratios over irradiation periods similar to
those employed in the stereochemical studies. Further-
more, when these phosphonates were subjected to the
entire workup procedure, their enantiomeric composi-
tions were unchanged.
TEMP Oa
accountability
yield (%)b
conversion
of 7 (%)
R/ S ratio
of 10d
7/TEMPO
10
11a c
e
e
e
15
24
48
23
20
16
17
20
19
22
20
67
70
66
68
66
68
67
67
64
65
64
4.0
3.8
4.2
3.8
3.0
2.0
0.8
0
86/14
85/15
86/14
85/15
86/14
88/12
89/11
90/10
89/11
90/10
90/10
1:0.4
1:0.6
1:0.8
1:1.0
1:1.5
1:1.8
1:2.0
1:2.2
Discu ssion
Tables 1-4 display data from the irradiation of phos-
phites 7-9 that demonstrate the essential invariance of
the results as a function of conversion. These data are
condensed and summarized in Table 5 for ease of
comparison. In addition, the last few columns of Table 5
list stereochemical results derived, as set forth later, from
the R/ S ratios of reactants and products found in Tables
1-4. Optically active phosphites 7-9 will be seen to give
product distributions and stereochemistries at the ste-
reogenic carbon, strongly dependent on whether the
photorearrangement proceeds from a triplet or singlet
excited state. These results will be discussed by reference
to Schemes 1 and 2. Scheme 1 depicts the reaction
alternatives and stereochemical options for a singlet
radical pair [3,14] formed from a singlet excited state
(14S). Scheme 2 is an analogous treatment for radical
pairs that are initially triplet. The essentially planar
1-arylethyl radicals (14) are either pro-R or pro-S,
depending on whether rotation (krot) has occurred. Com-
bination of pairs [3,14] within the initial solvent cage
(kcomb) affords phosphonate with either retained or in-
verted configuration at the stereogenic carbon center of
the carbon free radical with the degree of retention at
0
0
0
a
Ca. 0.01 M 7, cyclohexane solvent. Rayonet reactor at 35-40
°C. By GC analysis. Based on the amount of consumed 7. c Yield
b
of 11a doubled to account for the stoichiometry of its formation
d
from 7. Determined by chiral HPLC on a CHIRALCEL OD
column. e No TEMPO added.
and levels off at approximately 65%. Moreover, as the
concentration of the trap increases, a higher degree of
retention of configuration at the migrating carbon is
observed in the product phosphonate, and the reproduc-
ible enantiomeric composition of 10 increases to a con-
stant value (90/10 R/S; 80% ee). These results indicate
that total scavenging of cage-free radicals has occurred.
GC and GC/MS analysis showed the presence of
TEMPO-H and TEMPO-cyclohexyl resulting from hydro-
gen abstraction from the cyclohexane solvent by elec-
tronically excited TEMPO. However, neither of the
products of trapping of 1-phenylethyl radical or phos-
phinoyl (MeO)2P(O)• by TEMPO was detected by GC/MS
analyses. It not surprising that products from radical
trapping by TEMPO could not be detected by GC and
GC/MS analyses on irradiation of phosphite 7. The
reaction for the 4-keto analogue of TEMPO with reso-
nance-stabilized carbon-centered radicals is thermally
reversible even at ambient temperatures, and the 1,1-
diphenylethyl radical/TEMPO adduct decomposes slowly
even at 20 °C (k1 ) 4.4 × 104 M-1 s-1, k-1 ) 3.5 × 104
M-1 s-1).19,20 In a hot GC inlet and column, these
decomposition processes surely become even more sig-
nificant. Literature reports21,22 of trapping of phosphinoyl
radicals by TEMPO do not discuss the thermal stability
of the trapping products.
From the photolysis of phosphite 7 in the presence of
TEMPO, one might expect to observe the product (MeO)2P-
(O)-TEMPO by 31P NMR, as is reported in the photolysis
of 1 with Ar ) p-MeCOC6H4 at δ 6.7.3 (The 31P NMR
chemical shift for the analogous trapping product using
a closely related nitroxyl radical is 4.94.22) However, a
distinct peak in the expected region that could be clearly
assigned to (MeO)2P(O)-TEMPO was not found because
of low peak intensities, low signal-to-noise, and the
presence of several peaks in the expected chemical shift
region.
carbon dependent on the relative values of kcomb and krot
.
Cage combination (kcomb) to form phosphonate competes
with diffusive separation (kdiff) which generates cage-free
pairs that on combination yield phosphonate of racemic
(R/S) composition.
Before discussing stereochemistry in detail, it must be
strongly emphasized (as detailed in the Results section)
that the stereochemical integrities of both 7 and 9 are
completely preserved during photoreaction, whether on
direct irradiation of 7 and 9 or on triplet sensitization of
9. The enantiomeric excess of phosphite measured on the
phosphate, at a range of conversions, remain within ca.
2-3% of that prior to irradiation. (See Results section
and Supporting Information, Tables 1 and 2). This
finding will be presumed to apply also to phosphite 8,
whose phosphate enantiomers overlapped on chiral HPLC.
The phosphonates themselves similarly retain their
stereochemical integrity on photoirradiation. Thus, the
loss of configuration at stereogenic carbon in phospho-
nates 10, 12, or 13 must occur following formation of the
presumed radical pairs [3,14] of Schemes 1 and 2.
Furthermore, the stereochemical integrity of the start-
ing phosphites means that the radical pair [3,14] does
not recombine to regenerate phosphite. Thus, on direct
irradiation of 7 and 9, some loss of initial phosphite
stereochemistry is seen in the product phosphonates 10
and 13 (Table 5). Presuming that this results (as dis-
cussed subsequently) from combination of caged radical
pairs [3,14] following rotation of the prochiral 1-arylethyl
species (14), recombination of 3 and 14 to reform phos-
phite, were it to occur, must do so with some loss of
stereochemistry at carbon. This conclusion applies as well
to triplet sensitization of the reaction of 9 where stere-
Ster eoch em ica l Con tr ol Exp er im en ts. Solutions of
optically active phosphonates 10, 12, and 13, irradiated
(19) Grattan, D. W.; Carlsson, D. J .; Howard, J . A.; Wiles, D. M.
Can. J . Chem. 1979, 57, 2834-2842.
(20) Howard, J . A.; Tait, J . C. J . Org. Chem. 1978, 43, 4279-4283.
(21) Baxter, J . E.; Davidson, R. S. Makromol. Chem., Rapid Com-
mun. 1987, 8, 311-314.
(22) Busfield, W. K.; Grice, I. D.; J enkins, I. D. Aust. J . Chem. 1995,
48, 625-634.

Photo-Arbuzov Rearrangements of 1-Arylethyl Phosphites
J . Org. Chem., Vol. 66, No. 3, 2001 985
Ta ble 5. Su m m a r y of P h otoch em ica l Resu lts
accountability yield (%)
cage yield
phosphonate (%)
% net
a
compound
conditions
phosphonate
dimer
100yd
retentionc
kcomb/krot
7
7
8
9
9
direct hν
with TEMPO
direct hν
direct hν
sens.
67 ( 2b
64
19 ( 1
65 ( 4
20 ( 1
4
0
e
65
e
e
e
88 ( 2b
74
82
4
88
2
6c 7d
93
9c 13d
0.1c
38 ( 2
3 ( 1
40
f
95 ( 2b
15c 19d
0.04c
f
a
b
Yield of dimer doubled to account for the stoichiometry of its formation from starting materials. Average deviation from average.
c Method I. Method II. e TEMPO not added. Cage yield not applicable. f Method II not applicable.
d
ochemical integrity is almost totally lost in the product
phosphonate (13). To the extent that initially formed
triplet radical pairs might live long enough in the solvent
cage to intersystem cross and recombine, considerable
loss of stereochemistry in phosphite 9 formed on recom-
bination would be expected. However, this affect would
be attenuated in that only a small yield of cage products
of any type is expected from triplet pairs, as is discussed
later. That the pair [3,14] does not combine to regenerate
phosphite will be assumed to apply as well to the
photorearrangement of 8 whose phosphates were insepa-
rable by chiral HPLC analysis.
pair [3,(R)-14] formed from (R)-7 or (R)-9 (Scheme 1) but
also by photolysis of the 2-3% of the minor enantiomers,
(S)-7 and (S)-9. However, to a close approximation, one
still can apply eq 2 by simply correcting the amount of
(S)-10 or (S)-13 reported in the Tables 1, 3, and 4 for the
2-3% of phosphonate formed from S phosphite as
demonstrated below. Equation 3 then yields the ratio
k
comb/krot. Furthermore, net retention at carbon, an em-
pirical quantity not based on a presumed mechanism, is
derived in the usual fashion from eq 4:
% net retention ) (Robsd - Sobsd)/(Robsd + Sobsd) (4)
An a lysis of Ster eoch em ica l Resu lts. The stereo-
chemical results (Tables 1-4) can be analyzed in terms
of the radical-pair [3,14] of Scheme 1 by two approaches
(methods I and II) that give comparable results. Let RSM
and SSM be the percentage of starting phosphite in the R
and S configurations, respectively, and Robsd and Sobsd
represent the percentage of the observed product phos-
phonate in the R and S configurations. Method I is
modeled after the published work on the stereochemically
analogous thermal rearrangement via singlet radical
pairs of optically pure (S)-MePhCH-NdCdCPh2 re-
ported by Singer et al.8 The steady-state kinetic treat-
ment on the radical pair from the ketenimine was based
on a scheme entirely parallel to Scheme 1 with the
assumption that the rotation of radical 14 (1-phenylethyl
in both systems) is reversible. Thus, starting with stereo-
chemically pure (R)-phosphite, the steady-state approxi-
mation for the concentration of radical pair [3,14] is
applied to give eq 2 and from it eq 3:
In this approach (method I) for phosphite (R)-7, the
observed ratio Robsd/ Sobsd of 90/10 for the TEMPO-
scavenged photolysis (Table 4) is changed to 90/8 by
subtraction of the approximately 2% of (S)-10 generated
from (S)-7 to give a percentage ratio of 91/9. From eq 3,
kcombkrot ) 9. Percent net retention ) 100(91-9)/100 )
82. These parameters for photolysis of phosphites 7-9
are recorded in Table 5.
In Method II it is noted first that on direct irradiation
of 7 and 9, the degree of retention of configuration at the
stereogenic carbon in phosphonates 10 and 13 is very
high. Therefore, as a simplifying approximation, it can
be assumed that the initially formed pro-R radical 14 will
either combine to give R phosphonate or undergo only a
single rotation before radical combination to yield S
phosphonate. The same assumption of course applies to
the [3, (S)-14] radical pairs from phosphite of S config-
uration. In this approach neither the R nor the S form
of the starting phosphite is ignored, and Scheme 1 would
be expanded to include direct generation of the pair [3,(S)-
14] from (S)-7 and (S)-9. Clearly, the assumption that
kcomb . krot is not valid for the direct photolysis of
phosphite 8 or the triplet sensitized photolysis of 9 for
which stereochemistry is nearly completely lost. Method
II, therefore, is not applicable to those photoreactions.
By Method II if y equals the fraction of the initial
radical pair [3,14] from RSM or SSM that combines before
rotation of 14 to give product phosphonate with retention
of configuration, then 1 - y is the fraction of RSM or SSM
that yields product phosphonate with inverted configu-
ration at the stereogenic carbon. It is then obvious that
the total amount of each enantiomer to be formed (Robsd
or Sobsd) is given by eqs 5 and 6.
R
obsd/ Sobsd ) (kcomb + krot)/krot
(2)
and from eq 2
kcomb/ krot ) Robsd/Sobsd - 1
(3)
Equation 2 is simpler than its equivalent published in
ref 8 in its derivation in that formation of phosphonate
by combination of cage-free pairs is neglected; i.e., the
assumption is made that kdiff is negligible compared to
the sum kcomb + krot. This simplification is needed since
we do not have the otherwise necessary accurate measure
of the fraction of products, f, that escape the solvent cage
(f ) kdiff/ (kcomb + kdiff)) even for (R)-7, because of the lack
of total product accountability discussed elsewhere in this
paper. The neglect of kdiff appears justified by the small
amount of 1-arylethyl radical dimers (11a and 11c) found
and the small change in both yield of phosphonate 10
and overall stereochemistry noted when TEMPO is added
to the photolysis of (R)-7.
Robsd ) RSMy + SSM(1 - y)
Sobsd ) SSMy + RSM(1 - y)
(5)
(6)
These equations are readily solved to give eq 7 from
which y can be obtained. The quantity y, the fraction of
initially formed reaction pairs that combine with retention
Unfortunately, in the phosphite photoreaction systems,
pair [3, (S)-14] is formed not only by rotation within the

986 J . Org. Chem., Vol. 66, No. 3, 2001
Bhanthumnavin and Bentrude
of configuration at the chiral center of starting phosphite,
is given in Table 5 for phosphites 7 and 9 on a percentage
basis as 100y:
arise from the somewhat larger steric size of the 1-naph-
thylethyl radical which gives rise to larger rotational
correlation time.
Approximately 35% of the phosphinoyl radicals poten-
tially formed in the photorearrangements of 7 and 9 is
unaccounted for. Numerous side products are formed in
small quantities, as shown by 31P NMR spectroscopy, but
were not examined further because of their number and
our primary interest in reaction stereochemistry. Forma-
tion of unidentified products of coupling of the radical
pair by attack of the phosphinoyl radical at the ring
carbons of the naphthyl ring may occur. Indeed, a recent
paper reports23 the coupling at the ortho and para
carbons as well as the benzylic carbon for the radical pair
[Ph2P(O)• •CH2Ph] that is formed photolytically from
Ph2P(O)CH2 Ph. The formation of side products and the
trapping of radical diffusion products by benzene solvent
on photolysis of 1 (Ar ) Ph) was discussed in an earlier
paper.3 The failure to observe a prominent peak assign-
able to (MeO)2P(O)-TEMPO in the 31P NMR spectrum of
products from irradiation of phosphite 7 solutions with
TEMPO added supports the conclusion that cage-free
radicals are not generated in large amounts, as is also
indicated by low yields of products from the dimerization
of 1-arylmethyl radicals. Disproportionation of the radical
pair [3,14] is a further cage process available to the
1-arylethyl phosphites but not to arylmethyl phosphites
(1). However, 31P NMR or GC evidence for formation of
(MeO)2P(O)H in measurable amounts, likely >5%, was
not found.
y ) (Robsd - SSM)/(RSM - SSM) )
(Robsd - RSM - 100)/(2RSM - 100) (7)
For example, under TEMPO-scavenging conditions, the
photoreaction of phosphite 7, 97% of which is in the R
configuration at the stereogenic carbon, yields phospho-
nate 10 that is 90% R in composition, corresponding to y
) 0.93 (eq 7). The value of 100y (93) is different from the
% net retention obtained from eq 4 (method I), a number
obtained without regard to mechanism.
For method II the relationship between the rate
constant for rotation (krot) and the rate constant for
recombination (kcomb) simply corresponds (eq 8) to the
fraction of radical pairs within the initial solvent cage of
the starting material that undergoes recombination
before rotation has occurred divided by the fraction that
undergoes recombination after rotation:
kcomb/krot ) y/(1 - y)
(8)
Therefore, for the radical system discussed above from
phosphite 7 with added TEMPO (y ) 0.93), kcomb/krot
0.93/(1.00-0.93) ) 13 (Table 5). Thus, radical pair [3, 14]
is 13 times more likely to combine than to undergo
rotation followed by combination. Methods I and II are
)
seen in Table 5 to give very similar results for kcomb/ krot
.
Dir ect P h otolysis of P h osp h ites 7 a n d 9. As shown
in Tables 1, 3, 4, and 5, photorearrangements of the
1-arylmethyl phosphites 7 and 9, triggered by direct
irradiation, give the corresponding phosphonates in 60-
70% yields. When TEMPO is added (Tables 4), the yield
of phosphonate 10, exclusively cage product, is slightly
reduced to 65% compared to 67% in its absence. Only
3-4% of the potentially formed 1-arylethyl radicals (14,
Schemes 1 and 2) are accounted for as 2,3-diarylbutanes
(Tables 1, 3, and 4). Thus, the majority of phosphonates
10 and 13 is formed within the reaction cage. The
measured quantum yield, φP, for formation of phospho-
nate 10 from 7 of 0.32 is close to that for phosphonate
generation from 1 (Ar ) Ph) of 0.43.3 These results show
the photo-Arbuzov rearrangements of arthylmethyl and
1-arylethyl phosphites to be reasonably efficient processes.
A very high degree of combination of the initial pair
with retention of configuration at the stereogenic carbon
(100y, method I) is seen: for 7, 88 ( 2% without
scavenger added and 93% with sufficient levels of added
TEMPO (g0.015 M); for 9, 95 ( 2% unscavenged. The
corresponding percent net retention values (method II),
74, 82, and 88 for phosphites 7 and 9, show the same
systematic increase in stereospecificity with structure as
the 100y parameters.
From eqs 2 and 8, kcomb/ krot for the radical pair from 9
is estimated to be 15 and 19 (methods I and II, respec-
tively, avg 17). These are minimum values, since a small
portion of phosphonate 13 (no added TEMPO) arises via
combination of free pairs to yield racemic 13. On the basis
of comparisons of averaged kcomb/krot values for 7 and 9
from unscavenged photoreactions (7 and 17, respectively),
and the assumption that kcomb is essentially the same for
the radical pairs from 7 and 9, it appears that the
1-naphthylethyl radical undergoes rotation more slowly
than its 1-phenylethyl counterpart. This difference may
Com p a r ison s to Oth er P h oto-Ar bu zov System s.
The published study10 of the stereochemistry of the
photorearrangement of trans-5 showed a similar but
slightly lower degree of retention at stereogenic carbon
(100y ≈ 80) in an unscavenged reaction. Unscavenged
accountability yields of phosphonate from direct irradia-
tion of the benzyl phosphite analogous to 9 (1, Ar ) Ph)
range from 67 to 81% in acetonitrile, cyclohexane, and
benzene,3 and accountabilities of benzyl radicals as
bibenzyl are generally 6-8%. The naphthyl analogue of
7 and 9 (1, Ar ) 1-naphthyl) was found24 to give 71%
and 74% accountability yields of phosphonate in cyclo-
hexane and benzene, respectively, that are reduced by
5-10% by the addition of tert-BuBr which scavenges
the cage-free phosphinoyl radicals to form (MeO)2P(O)-
Br.24
Th e Qu estion of Ion -P a ir In ter m ed ia tes. The pos-
sibility that these singlet photoprocesses involve ion pairs
was addressed in a previous publication3 on the photo-
chemistry of phosphite 1 with Ar ) Ph and Ar )
p-MeCOC6H4. It was discounted on the basis of the lack
of effect on product distribution and trapping results of
changing the solvent from benzene to highly polar
acetonitrile. Unpublished results25 showing the lack of
systematic effect of substituents on the group Ar (Ph and
1-naphthylmethyl) of 1 and the failure to trap ion pairs
with added MeOH were also cited. The lack of effect of
solvent polarity on these reactions (Tables 1, 3, and 4) is
further emphasized by the consistency of the stereochem-
ical outcome in polar acetonitrile as well as in nonpolar
cyclohexane and benzene. The lifetimes of ion pairs, and
(23) Zhao, N.; Neckers, D. C. J . Org. Chem. 2000, 65, 2145-2150.
(24) Bentrude, W. G.; Ganapathy, S.; Somasekhar, B. B. V. Unpub-
lished results.
(25) Bhanthumnavin, W. Unpublished results from this labora-
tory.

Photo-Arbuzov Rearrangements of 1-Arylethyl Phosphites
J . Org. Chem., Vol. 66, No. 3, 2001 987
their ability to lose stereochemical integrity before com-
bination, might be expected to respond to changes in
solvation.
for radical coupling reactions, which are already of the
order of those for diffusion, may seem unusual. However,
the so-called proximate free radical pairs under discus-
sion are formed in singlet reactions which are on the
mechanistic border between radical pair formation and
truly concerted processes. Thus, in addition to effects of
hybridization, very small changes in the initial proximity
of pairs, perhaps engendered by steric factors, may give
rise to changes in kcomb. Though steric effects on radical
combination are well-known, we are unaware of studies
correlating variations in radical hybridization with rates
of coupling (kcomb) of caged pairs. An increased kcomb
means that, as was argued earlier, the yield of cage
products from photolysis of 7 and 9 should be higher than
the experimental yields of phosphonates 10 and 12.
Com p a r ison s to Rela ted Ra d ica l-P a ir System s. A
classic study of a singlet proximate radical pair system
(no intervening molecule, e.g., CO2, N2) is the thermolysis
at 60 °C of Ph2CdCdNCHPhMe which was done both
with and without added scavengers.8 The coupling prod-
uct Ph2C(CN)CHPhMe was formed in g95% chemical
yield (unscavenged). Scavenging established cage yields
of nitrile of 62 and 75% in acetonitrile and CCl4,
respectively. The reported8 kcomb/krot values of 2.5 (CCl4)
and 1.3 (acetonitrile) for the ketenimine system sug-
gest that if kcomb is similar for the proximate pairs
from phosphite 7 and the ketenimine, then krot for the
1-phenylethyl radical (14) in the ketenimine system is
somewhat larger than it is in the radical pair [3,14]
formed from phosphite 7 (kcomb/krot ) 11, avg). An in-
creased krot value might result at least in part from the
ketenimine reaction being studied in a solvent of lower
viscosity and/or from the higher temperature (60 °C) at
which the system was investigated. Indeed, viscosity
differences between CCl4 and acetonitrile are most prob-
ably responsible for the approximate 2-fold decrease in
the ratio kcomb/krot determined at 60 °C in the investiga-
tion of Ph2CdCdNCHPhMe, as suggested by the authors
of that paper.8
The ratio kcomb/krot (11, avg value) for the radical pair
from phosphite 7 (TEMPO added) was measured at 35-
40 °C in cyclohexane, which has a viscosity (1.02 cp, 17
°C) close to that of CCl4 (1.04 cp, 20 °C). However, at 60
°C, where the study of Ph2CdCdNCHPhMe was done,
the viscosity of CCl4 is reduced to 0.58 cp8 which
consequently would increase krot (essentially rotational
diffusion) for the 1-phenylethyl radical. The viscosity
related expected increase in kcomb/krot at 35 °C, compared
to the value at 60 °C (2.5) (kcomb assumed invariant with
temperature), can be estimated from the expression
(η-RT1/η-RT2), where η is the viscosity at T1 or T2, and R is
a constant, normally one or less.26 (In other words, krot is
inversely proportional to viscosity to the power R.26) The
maximal effect of temperature decrease from 60 °C, at
which the radicals from Ph2CdCdNCHPhMe were stud-
ied, to 35° C where the investigations of phosphite 7 were
done, results for R ) 1. Using the reported value for η
for CCl4 at 60 °C (0.58 cp) and an interpolated value27 at
35 °C (0.81 cp), the reported ratio kcomb/krot for the
1-phenylethyl radicals from Ph2CdCdNCHPhMe would
be increased to (0.81/0.58)(2.5) or 3.5. This small increase
from 2.5 to 3.5 still allows for a severalfold increase in
kcomb for the 1-phenylethyl radical from photolysis of
phosphite 7 (kcomb/krot ) 11, avg value).
Most importantly, the relatively high kcomb/krot values
found for the ketenimine system, and for 7 and 9, all
involving proximate radical pairs, contrast strongly to
those observed for the nonproximate 1-phenylethyl radi-
cal-pair generated by thermolysis (105 °C) in benzene of
meso and S,S-(-)-azobis-sym-1-phenylethane (kcomb/krot
0.07)29 and the likewise nonproximate pair from (S)-Ph2-
CHNdNCHMePh at 110 °C in benzene (kcomb/krot
)
)
0.06).30 In both of these nonproximate, singlet radical-
pair systems, approximately 70% of the radicals diffuse
from the initial solvent cage (30% cage reaction). The
reduction in relative kcomb resulting from the presence of
the intervening nitrogen molecule is strongly evident. It
is also seen in value for kdiff/kcomb of 2.5 obtained for the
azobis-sym-1-phenylethane system.29
Were the product accountabilities for the photorear-
rangements of 7 and 9 higher, values for krot/kdiff and kcomb
/
kdiff could be calculated. Reasonable values would result
only if it were assumed that the yields of dimer 11
represent the total amount of noncage product formed
with the remaining products, perhaps 95%, resulting
from reaction within the solvent cage. A high degree of
combination relative to rotation is observed for pair [3,14]
from phosphite 7 (kcomb/krot ) 11, avg) and 9 (kcomb/krot
)
17, avg). Therefore, kcomb/kdiff values much larger than
those seen for the proximate radical pairs formed from
the Ph2CdCdNCHPhMe8 and the nitrogen-separated
pairs from azobis-sym-1-phenylethane29 would be antici-
pated.
In summary, the results of product studies, effects of
added scavengers, the failure to detect 31P CIDNP and
CIDEP phenomena,6,7,31 the stereochemical studies re-
ported here, and comparisons to related research on
proximate radical pairs,8 point to the formation on direct
irradiation of arylmethyl and 1-arylethyl phosphites
(including 7 and 9) of short-lived, singlet radical pairs,
[3,14], that predominantly undergo rapid recombination
within the initial solvent cage. One cannot rule out an
important concerted component of these photorearrange-
ments that features total retention of configuration at
migratory carbon. However, in the absence of evidence
for a concerted component, the available results are best
interpreted in terms of very short-lived radical pairs,
[3,14] (Scheme 1). This conclusion is in accord with
It seems likely, though still quite speculative, that kcomb
is increased for the pair [3,14], because it involves a
radical, (MeO)2P(O)•, which has a SOMO with a high
degree of s character. Changes in rate constants of
second-order reactions of a series of phosphinoyl radicals,
X2P(O)•, have been correlated with variations in SOMO
s character.12,13,28 Measurable variations in rate constants
(26) E.g., viscosity-dependent krot of caged radicals have been
correlated by η-0.75 (Koenig, T.; Owens, J . M. J . Am. Chem. Soc. 1973,
95, 8484) and by η-0.50 (J ohnson, R. A.; Seltzer, S. J . Am. Chem. Soc.
1973, 95, 938).
(27) Values of η for CCl4 at 25 °C (0.91 cp) and 50 °C (0.66 cp) are
reported in Lide, D. R. Handbook of Organic Solvents; CRC Press: Boca
Raton, 1995 from which an interpolated value at 35 °C of 0.81 cp can
be obtained.
(28) Gatlik, I.; Rzadek, P.; Gescheidt, G.; G., R.; Hellrung, B.; Wirz,
J .; Dietliker, K.; Hug, G.; Kunz, M.; Wolf, J .-P. J . Am. Chem. Soc. 1999,
121, 8332-8336.
(29) Greene, F. D.; Berwick, M. A.; Stowell, J . C. J . Am. Chem. Soc.
1970, 92, 867-874.
(30) Kopecky, K. R.; Gillan, T. Can. J . Chem. 1969, 47, 2371-2386.
(31) Sluggett, G. W.; Landis, M. S.; Turro, N. J . Unpublished results.

988 J . Org. Chem., Vol. 66, No. 3, 2001
Bhanthumnavin and Bentrude
previous studies of photo-Arbuzov rearrangements of
closely related 1-arylmethyl phosphites (1)1,2,6,7,9,10,24
Dir ect Ir r a d ia tion of Op tica lly Active P h osp h ite
8. It has been shown that introduction of the p-acetyl
substituent in 1 (Ar ) p-MeCOC6H4) results on direct
irradiation in generation of triplet radical pairs.3,6,7 The
potential fates of the analogous triplet radical pairs [3,14]
from direct irradiation of (S)-8 are shown in Scheme 2.
Since triplet radical pairs are likely to be relatively long-
lived, the rotation step (krot) is expected to compete
readily with kcomb and kdiff and be reversible as shown.
Yields of products formed in the initial cage should be
quite low as confirmed by the results seen in Tables 2
and 5. Large amounts of diffusion (kdiff, Scheme 2) yield
the product of dimerization of cage-free 1-arylethyl
radicals, 11b, in 20% yield that accounts for 40% of the
p-acetyl-1-phenylethyl radicals potentially formed.
The major portion of phosphonate 12 found (20% yield)
most likely results from random couplings of cage-free
radicals 3 and 14. Thus, by comparison, the photorear-
rangement of the closely related dimethyl p-acetylbenzyl
phosphite (1, Ar ) p-MeC6H4) in the absence of trapping
agents gives the phosphonate in 12% yield. However,
under optimal scavenging by added benzyl bromine or
thiophenol, only 3-5% of the phosphonate is formed,
presumably from combination of the radical pair analo-
gous to [3,14] prior to diffusion from the solvent cage.3
If phosphite 8 also gives a 3-5% yield of cage-formed
phosphonate (12), the very small degree of retention (4%
net retention, Table 5) at the migratory carbon observed
on direct irradiation of (S)-8 is expected (Scheme 2).
Phosphonate 12 from random recombination will be
totally racemic, and the initial triplet pairs (3-5%) that
undergo combination in the solvent cage also should
undergo considerable stereorandomization prior to in-
tersystem crossing and coupling.
of (R)-8. Thus, the stereochemical outcome from both the
direct irradiation of (R)-8 and triplet-sensitization of (R)-9
is diagnostic of the initial formation of triplet free radical
pairs.
Con clu sion s
For 1-arylethyl phosphites 7 and 9, the proposed
intermediacy of short-lived singlet radical pairs in the
formation of 1-arylethylphosphonates 10 and 12 on direct
UV irradiation is consistent with the relatively high
yields of phosphonates formed and with the very high
degree of stereospecificity (retention) observed at the
stereogenic migratory carbon. Nonetheless, the partial
loss of stereochemistry is discordant with the operation
of a purely concerted process. The degree of retention of
configuration at carbon is higher than that reported for
the combination of other proximate singlet radical pairs
in solution. Thus, the ratio kcomb/krot ) 11 (avg, TEMPO-
scavenged reaction, cyclohexane) determined for the
proximate radical pairs [3,14], generated photochemically
from phosphite (S)-7, is significantly higher than that
reported for the thermally generated singlet radical pairs
from ketenimine Ph2CdCdNCHPhMe, for which kcomb
/
krot ) 1.3 (acetonitrile) and 2.5 (CCl4) even when the
higher temperature of the ketenimine system is consid-
ered. This may be interpreted as consistent with a
severalfold increase in kcomb for the radical pair from
phosphite 7, presuming that krot for the 1-phenylethyl
radical, adjusted for the difference in reaction tempera-
tures, is nearly the same for the two pairs. By contrast,
the direct irradiation of phosphite 8 and sensitized
photoreaction of phosphite 9 evidently generates prima-
rily triplet radical pair intermediates and gives close to
total stereorandomization at the stereogenic carbon in
the product phosphonates formed by a combination of
cage and noncage radical combination processes. The
contrast in observed stereochemistry among the phos-
phites 7-9 is totally consistent with the outcome of
product distribution studies1-3,9,24 as well as CIDNP6,31
and CIDEP7,31 results for 7-9 and closely related phos-
phites. Thus, the stereochemical approach provides a
useful test for the generation of predominantly singlet
versus largely triplet radical pair populations in these
photo-Arbuzov reaction systems.
Stereochemical data in Table 5, derived by method I,
show a roughly 4% net retention for phosphite 8 in line
with the observed kcomb/krot ratio of 0.1, a minimum value.
This ratio is derived from phosphonate formed predomi-
nantly outside the solvent cage following complete ran-
domization of stereochemistry at the stereogenic carbon.
Were phosphonate formed strictly within the solvent
cage, the ratio of enantiomers would be higher than was
observed and lead to a larger value for the ratio kcomb
/
krot. Nonetheless, the inverse ratio, krot/kcomb, will be
assuredly be considerably higher than that for radicals
from phosphite 7 under TEMPO scavenged conditions
(11, avg value, Table 5).
Exp er im en ta l Section
Gen er a l P r oced u r es a n d Ma ter ia ls. Air-sensitive mate-
rials were transferred by syringe or cannula or in a glovebag
under an argon atmosphere. Spectrophotometric grade sol-
vents were used as received unless otherwise noted. Diethyl
ether, toluene, and tetrahydrofuran were dried over sodium/
benzophenone and freshly distilled prior to use. Chemicals
purchased from Aldrich Chemical Co. were (()-1-phenyletha-
nol, (R)-(+)-1-phenylethanol, (()-R-methyl-1-naphthalene metha-
nol, (-)-diisopinocampheylchloroborane, (-)-menthyl chloro-
formate. Tri-n-propyl phosphate was obtained from Lancaster
Synthesis. Unless otherwise stated, distillations were per-
formed with a short path apparatus. Flash chromatography
was conducted on silica gel 60, 230-400 mesh, obtained from
EM Science with the solvent system indicated. Microanalyses
were performed by Atlantic Microlab, Inc., Norcross, GA.
Sp ectr oscop ic a n d P h ysica l Da ta . Melting points are
Tr ip let-Sen sitized Rea ction s of P h osp h ite 9. Rela-
tively high yields of dimer 11c (ca. 40% accountability of
radicals 14 potentially formed), comparatively low yields
of phosphonate 13 (20%), and the very small degree of
retention of a configuration (% retention ≈ 2) observed
on triphenylene-sensitized reactions of 9 (Table 3) are
in line with the product distribution and stereochemical
outcome obtained on direct irradiation of phosphite 8,
which was designed to give product from the triplet
manifold upon excitation of the carbonyl functionality.
Similarly, the triplet thioxanthone-sensitized reaction of
dimethyl 1-naphthyl phosphite, 1 (Ar ) 1-naphthyl),
gives a 3-4% cage yield of the corresponding phospho-
nate.24 The mechanism of triphenylene-sensitized rear-
rangement of R-9 could be represented by a reaction
scheme involving generation of triplet radical pairs that
is completely analogous to Scheme 2 for direct irradiation
1
uncorrected. J values given in the H NMR spectral data refer
to proton-proton coupling unless otherwise stated. Ultraviolet
(UV) spectra were obtained in acetonitrile and cyclohexane on
a diode array spectrophotometer. Wavelength maxima (λmax
)
.
are reported in nm with extinction coefficients, ꢀ, in M-1 cm-1

Photo-Arbuzov Rearrangements of 1-Arylethyl Phosphites
J . Org. Chem., Vol. 66, No. 3, 2001 989
Qualitative and quantitative gas chromatographic analyses
were determined with a flame ionization detector (FID) and a
DB-1 capillary column(1% methyl silicone) (10 or 30 m × 0.25
mm × 0.25 mm) with tri-n-propyl phosphate as internal
standard. LR and HRMS were performed in the EI or the CI
mode (70 eV). GC/MS spectra were taken in the EI or the CI
mode. The column used was a 15 or 30 m × 0.25 mm × 0.25
mm DB-1 capillary column.
Gen er a l P r oced u r e for Deter m in a tion of En a n tio-
m er ic Com p osition of Ch ir a l Com p ou n d s. Column chro-
matographic purification (silica gel, chloroform) was done on
all mixtures before separation on HPLC. The prepurified
mixture was dissolved in 1-2 mL of the intended HPLC eluent
solvent system. The solution was filtered through a Millex-
HV13 (0.45 µm, 13 mm diameter) filter unit. The commonly
used solvent system for the separation of phosphates and
phosphonates from the mixture was 1.3-1.7% methanol in
chloroform. HPLC separations were obtained under isocratic
conditions (UV detector, 4.6 mm i.d. analytical, 10 mm i.d.
semipreparative, or 21.4 mm i.d. preparative Dynamax HPLC
column (100 A° spherical Microsorb packings in 5 µm particle
size). Optimum flow rates were 10-16 mL/min on the 21.4
mm i.d. column.
Chiral high performance liquid chromatographic analyses
were performed on a CHIRALCEL OD HPLC column (250 mm
× 4.6 mm i.d.), equipped with a 50 mm × 4.6 mm i.d. guard
column. Mobile phases were 95:5 to 90:10 of hexanes:2-
propanol, optimal flow rate 1.0 mL/min.
Oxid a tion of P h osp h ites 7 a n d 9. A dilute solution of the
starting phosphite in acetonitrile was treated with a 3.0 M
solution of tert-butyl hydroperoxide in 2,2,4-trimethylpentane
(small excess) in an ice bath. The solution was stirred for 10
min, and the solvent was removed by rotary evaporation. The
mixture was purified by HPLC and subsequently analyzed for
enantiomeric purity as described above. The unreacted starting
material from each photolytic mixture was oxidized in the
same manner, purified, and analyzed by chiral HPLC.
Qu a n tu m Yield Deter m in a tion s. The procedure was
previously described.7
Gen er a l P r oced u r e for Dir ect Ir r a d ia tion of P h os-
p h ites 7-9. All sample preparation was carried out in a
glovebag under an argon atmosphere. All solvent used is argon
saturated. A 0.01 M solution of phosphite was prepared
containing an equimolar amount of tri-n-propyl phosphate as
internal standard in a 10 mL volumetric flask and transferred
to a series of argon flushed quartz (phosphite 7) or Pyrex
(phosphite 8 and 9) tubes (70 mm × 12 mm o.d.). The tubes
were rubber septum capped, wrapped with Parafilm, that was
tightened with copper wired, and further wrapped with a piece
of aluminum foil. The solution was purged with (5 min) and
then irradiated in a Rayonet photochemical apparatus equipped
with 6 × 253.7 nm lamps (phosphite 7) or 6 × 300 nm lamps
(phosphite 9). Solutions of 8 were irradiated using light from
a 450 W medium-pressure Hg lamp filtered through a uranium
glass sleeve. The reactions were monitored by GC analysis
(duplicate or triplicate injections, results averaged).
Tr ip h en ylen e-Sen sitized P h otor ea ction of 9. Similarly,
under an argon atmosphere, a 10 mL 0.01 M acetonitrile
solution of 9 containing an equimolar amount of tri-n-propyl
phosphate internal standard and 2 equiv of triphenylene was
partitioned between three Pyrex tubes fixed to 10/30 joints,
degassed with four freeze-pump-thaw cycles (0.01 mmHg),
and flame-sealed under vacuum. The samples were irradiated
with light from a medium pressure 450-W Hg lamp and filtered
through a uranium glass filter sleeve. Product yields were
determined by GC after 7, 27, and 60 h after oxidation with
tert-butyl hydroperoxide in 2,4,6-trimethylpentane. The oxi-
dized unreacted 9 and the phosphonate 13 were isolated by
HPLC, and the stereochemistry of enantiomeric compositions
of both components were determined as described earlier.
P r ep a r a tion of 2-(1-P h en yleth yloxy)-1,3,2-d ioxa p h os-
p h or in a n e (7). At room temperature under argon atmo-
sphere, a solution of 1-phenylethanol (3.0 g, 25 mmol) in 10
mL of acetonitrile was added dropwise to a solution of N,N-
diethyl-1,3,2-dioxaphosphorinan-2-amine (4.4 g, 25 mmol) and
1-H-tetrazole (18 mg, 2.5 mmol) in 20 mL of acetonitrile. The
solution was stirred at room temperature for an additional 1
h. The solvent was removed in vacuo. A 15 mL portion of 1:1
diethyl ether:ethyl acetate was added to the residue. The
solution was filtered by Schlenk techniques under argon.
Solvent removal in vacuo and vacuum distillation gave 7 (3.5
g, 5.0 mmol, 62% yield) as a colorless liquid: bp 80-82 °C (0.05
mmHg) (99% purity by GC). 31P NMR (121 MHZ, CDCl3, {1H})
1
δ 130.09; H NMR (300 MHz, CDCl3) δ 1.44-1.52 (m, 1 H),
3
1.59 (d, 3 H, J HP ) 6.6 Hz), 2.32-2.48 (m, 1 H), 3.60-3.70
(m, 1 H), 3.74-3.84 (m, 1 H), 4.23-4.33 (m, 1 H), 4.49-4.58
(m, 1 H), 5.16 (dq, 1 H, 3J HP ) 8.7 Hz, 3J ) 6.5 Hz), 7.22-7.42
3
(m, 5 H); 13C NMR (75 MHz, CDCl3, {1H}) δ 25.24 (d, J CP
)
3.7 Hz), 28.68 (d, 3J CP ) 5.3 Hz), 59.28 (d, 2J CP ) 2.3 Hz), 59.32
2
2
(d, J CP ) 2.4 Hz), 71.69 (d, J CP ) 20.3 Hz), 126.02, 127.63,
3
128.61, 144.51 (d, J CP ) 3.2 Hz); UV (CH3CN): 248 (ꢀ 183),
254 (ꢀ 210), 258 (ꢀ 227), 264 (ꢀ 184); GC-EIMS (70 eV) m/z (rel
intensity) 226 [M]+ (6), 122 (29), 105 [CHCH3Ph]+ (100), 77
[Ph]+ (24); HRMS (EI) m/z [M]+ calcd for C11H15O3P: 226.0757,
found: 226.0747.
P r ep a r a t ion of (R)-2-(1-P h en ylet h yloxy)-1,3,2-d ioxa -
p h osp h or in a n e ((R)-7). By the procedure racemic 7, reaction
of N,N-diethyl-1,3,2-dioxaphosphorinan-2-amine (1.3 g, 7.4
mmol), (R)-(+)-1-phenylethanol (0.9 g, 7.4 mmol) (97/3 R/S by
HPLC on a CHIRALCEL OD), and 1-H-tetrazole (7.0 mg, 1
mmol) gave 1.1 g (4.9 mmol, 66% yield) of (R)-7 (98.8% purity
by GC, 97/3 R/S ratiodetermined on the phosphate).
P r ep a r a tion of 2-(1-P h en yleth yloxy)-2-oxo-1,3,2-d ioxa -
p h osp h or in a n e (7-O). Oxidation of racemic 7 (0.9 g, 4.1
mmol) by tert-butyl hydroperoxide (1.4 mL), as described
earlier, gave, after HPLC purification, 1.0 g of its oxide as an
oil (4.0 mmol, 97% yield, 99.8% purity by GC). 31P NMR (121
MHz, CDCl3, {1H}) δ -7.50; 1H NMR (300 MHz, CDCl3) δ
3
1.62-1.64 (m, 1 H), 1.68 (d, J ) 6.6 Hz), 2.10-2.26 (m, 1 H),
3.95-4.05 (m, 1 H), 4.11-4.24 (m, 1 H), 4.32-4.39 (m, 2 H),
5.53 (dq, 1 H, 3J HP ) 7.6 Hz, 3J ) 6.5 Hz), 7.29-7.43 (m, 5 H);
2
13C NMR (75 MHz, CDCl3, {1H}) δ 23.71 (d, CHCH3, J CP
)
4.7 Hz), 25.59 (d, 3J CP ) 6.7 Hz), 68.09 (d, 2J CP ) 6.8 Hz), 68.61
2
(d, J CP ) 6.7 Hz), 76.23 (d, J CP ) 5.2 Hz), 125.57, 128.06,
3
128.33, 141.05 (d, J CP ) 5.7 Hz); GC-EIMS (70 eV) m/z (rel
intensity) 242 [M]+, 139 (61), 121 [(OCH2CH2CH2O)PdO]+
(10), 105 [CHCH3Ph]+ (52), 104 (100), 77 [Ph]+ (33); HRMS
(EI) m/z [M]+ calcd for C11H15O4P: 242.0708, found: 242.0726.
Anal. Calcd for C11H15O4P: C, 54.55; H, 6.24. Found: C, 54.62;
H, 6.20.
2-(1-P h en yleth yl)-2-oxo-1,3,2-d ioxa p h osp h or in a n e (10)
was isolated by HPLC (1.1% methyl alcohol in chloroform) from
the products of photolysis of dioxaphosphorinane 7 as a white
solid. Recrystallization from ether/pentane gave colorless
needles (mp 90-91 °C). 31P NMR (121 MHz CDCl3, {1H}) δ
1
3
0.99; H NMR (300 MHz, CDCl3) δ 1.64 (dd, J HP ) 18.7 Hz,
3J ) 7.5 Hz), 1.67-1.79 (m, 1 H), 1.81-1.92 (m, 1 H), 3.31
(dq, 1 H, J HP ) 22.2 Hz, 3J ) 7.4 Hz), 3.95-4.14 (m, 2 H),
3
4.38-4.48 (m, 2 H), 7.25-7.36 (m, 5 H); 13C NMR (75 MHz,
CDCl3, {1H}) δ 15.03 (d, J CP ) 6.7 Hz), 26.21 (d, J CP ) 6.7
2
3
Hz), 38.05 (d, 1J CP ) 134.9 Hz), 66.21 (d, 2J CP ) 2.7 Hz), 66.31
2
(d, J CP ) 2.7 Hz), 127.21 (d, J CP ) 3.5 Hz), 128.51, 128.55 (d,
J CP ) 1.6 Hz), 128.65, 137.52 (d, 3J CP ) 7.5 Hz); UV (CH3CN):
242 (ꢀ 147), 250 (ꢀ 193), 258 (ꢀ 236), 264 (ꢀ 168), 268 (ꢀ 94);
GC-EIMS (70 eV) m/z (relative intensity) 226 [M]+ (25), 105
[CHCH3Ph]+ (100), 77 [Ph]+ (23); Anal. Calcd for C11H15O3P:
C, 58.41; H, 6.68. Found: C, 58.08; H, 6.37.
P r ep a r a tion of (S)-1-(4-(1-Hyd r oxyeth yl)p h en yl)eth a -
n on e. The procedure of Brown et al. was followed with a slight
modification.32-34 A reduction of 2-(4-acetylphenyl)-2-methyl-
1,3-dioxolane (see Supporting Information, 0.61 g, 2.9 mmol)
with diisopinocampheylchloroborane ((-)-DIP chloride) (1.1 g,
3.3 mmol) followed by a treatment with acetaldehyde (0.15 g,
(32) Brown, H. C.; Chandrasekharan, J .; Ramachandran, P. V. J .
Am. Chem. Soc. 1988, 110, 1539-1546.
(33) Brown, H. C.; Ramachandran, P. V.; Teodorovic, A. V.; Ran-
gaishenvi, M. V. J . Org. Chem. 1992, 57, 2379-2386.
(34) Brown, H. C.; Ramachandran, P. V. Acc. Chem. Res. 1992, 25,
16-24.

990 J . Org. Chem., Vol. 66, No. 3, 2001
Bhanthumnavin and Bentrude
3.5 mmol) gave the alcohol (0.33 g, 2.0 mmol, 70%) as a
colorless thick oil. H NMR (300 MHz, CDCl3) δ 1.54 (d, 3 H,
(77), 77 [Ph]+ (14), 43 [OdCCH3]+ (61). Anal. Calcd for
1
C
13H17O4P: C, 58.21; H, 6.39. Found: C, 58.19; H, 6.39.
J ) 6.4 Hz), 2.58 (s, 3 H), 4.96 (q, 1 H, J ) 5.6 Hz), 7.45 (d, 2
H, J ) 8.5 Hz), 7.92 (d, 2 H, J ) 8.3 Hz) (lit.35); 13C NMR (75
MHz, CDCl3, {1H}) δ 25.27, 26.58, 69.77, 125.41, 128.56,
136.07, 151.24, 197.98 (s, OdCCH3) (lit.35). The optical purity
of the alcohol was determined after its conversion to the cor-
responding diastereomeric carbonates following the method of
Westley and Halpern employing (-)-menthyl chloroformate.36
The enantiomeric ratio (R/ S) was 97.5/2.5 (GC).
P r ep a r a tion of Dim eth yl 1-(1-Na p h th yl)eth yl P h os-
p h ite (9). This compound was prepared from the reaction of
dimethyl N,N-diethylphosphoramidite (5.3 g, 32 mmol), R-meth-
yl-1-naphthalene methanol (4.3 g, 25 mmol), and 1-H-tetrazole
(280 mg, 4.0 mmol) following the procedure for preparation of
7. Pure 9 (98.5% by GC) was purified by flash chromatography
as a colorless oil (4.5 g, 17 mmol, 67%). 31P NMR (121 MHz,
1
CDCl3, {1H}) δ 132.07; H NMR (300 MHz, CDCl3) δ 1.71 (d,
3
3
P r ep a r a tion of 2-(1-(4-Acetylp h en yl)eth yloxy)-1,3,2 d i-
oxa p h osp h or in a n e (8). The reaction of N,N-diethyl-1,3,2-
dioxaphosphorinan-2-amine (1.3 g, 7.3 mmol), 1-(4-acetylphenyl)-
ethanol (1.2 g, 7.2 mmol), and 1-H-tetrazole (7.0 mg, 1.0 mmol)
using the method to prepare phosphite 7 afforded 8 (99%
purity by GC) as a pale yellow thick oil (1.2 g, 4.4 mmol, 62%)
after flash chromatography purification. 31P NMR (121 MHz,
3 H, J ) 6.6 Hz), 3.41 (d, 3 H, J HP ) 10.9 Hz), 3.44 (d, 3 H,
3J HP ) 10.5 Hz), 6.06 (dq, 1 H, J HP ) 8.8 Hz, 3J ) 6.4 Hz),
3
7.40-7.50 (m, 3 H), 7.69 (d, 1 H, J ) 7.1 Hz), 7.73 (d, 1 H, J
) 8.1 Hz), 7.82 (d, 1 H, J ) 7.6 Hz), 8.10 (d, 1 H, J ) 8.3 Hz);
13C NMR (75 MHz, CDCl3, {1H}) δ 24.87 (d, J CP ) 3.1 Hz),
3
2
2
48.82 (d, J CP ) 9.3 Hz), 49.06 (d, J CP ) 9.9 Hz), 68.57 (d,
2J CP ) 15.0 Hz), 123.05, 123.07, 125.33, 125.85, 127.84, 128.75,
129.70, 133.62, 139.46, 139.51; UV (CH3CN): 264 (ꢀ 3.7 × 104),
274 (ꢀ 6.0 × 104), 284 (ꢀ 7.1 × 104), 294 (ꢀ 4.8 × 104); GC-
EIMS (70 eV) m/z (relative intensity) 264 [M]+ (8), 155
[CHCH3C10H7]+ (100), 154 (45), 127 [C10H7]+ (15); HRMS [M]+
calcd for C14H17O3P: 264.0915, found: 264.0913.
1
CDCl3, {1H}) δ 129.33; H NMR (300 MHz, CDCl3) δ 1.61 (d,
3
3 H, J ) 6.4 Hz), 1.67-1.79 (m, 1 H), 2.43 (m, 1 H), 2.60 (d,
3 H,3J ) 0.7 Hz), 3.65-3.75 (m, 1 H), 3.78-3.88 (m, 1 H), 4.27-
4.37 (m, 1 H), 4.50-4.60 (m, 1 H), 5.22 (dq, 1 H, J ) 8.7, 6.6
Hz), 7.51 (d, 2 H, J ) 8.3 Hz), 7.96 (d, 2 H, J ) 8.3 Hz); 13C
NMR (75 MHz, CDCl3, {1H}) δ 24.90 (d, 3J CP ) 3.6 Hz), 26.51,
P r ep a r a tion of Dim eth yl 1-(1-Na p h th yl)eth yl P h os-
p h a te (9-O). The oxidation of 9 (0.9 g, 3.4 mmol) using tert-
butyl hydroperoxide followed by HPLC purification (1% methyl
alcohol in chloroform) resulted in (0.9 g, 3.3 mmol, 96%) of
the oxide as a colorless oil. 31P NMR (121 MHz, CDCl3, {1H})
3
2
28.30 (d, J CP ) 5.2 Hz), 59.35 (d, J CP ) 1.0 Hz), 59.44 (d,
2J CP ) 1.6 Hz), 70.89 (d, J CP ) 20.2 Hz), 125.70, 128.45,
2
136.16, 149.07 (d, J CP ) 3.1 Hz), 197.60 (s); UV (CH3CN): 278
(ꢀ 1350), 284 (ꢀ 340), 288 (ꢀ 690), 314 (ꢀ 130); GC-EIMS (70
eV) m/z (relative intensity) 268 [M]+ (100), 253 [M - CH3]+
(22), 227 (84), 225 [M - OdCCH3]+ (11), 43 [OdCCH3]+ (36);
HRMS [M]+ calcd for C13H17O4P: 268.0864, found: 268.0870.
P r epar ation of (S)-2-(1-(4-Acetylph en yl)eth yloxy)-1,3,2-
d ioxa p h osp h or in a n e ((S)-8). The compound was prepared
from the reaction of (S)-1-(4-acetylphenyl)ethanol (0.10 g, 0.61
mmol), N,N-diethyl-1,3,2-dioxaphosphorinan-2-amine (0.11 g,
0.62 mmol), and 1-H-tetrazole (0.7 mg, 0.1 mmol). Pure 8
(98.7% purity by GC) was obtained after flash chromatography
(0.11 g, 0.40 mmol, 65%).
3
δ 1.09; 1H NMR (300 MHz, CDCl3) δ 1.82 (d, 3 H J ) 6.6 Hz),
3
3
3.58 (d, 3 H, J HP ) 11.2 Hz), 3.73 (d, 3 H, J HP ) 11.0 Hz),
6.25 (dq, 1 H, 2J HP ) 6.9 Hz, 3J ) 6.9 Hz), 7.46-7.57 (m, 3 H),
7.66 (d, 1 H, J ) 6.6 Hz), 7.80-7.89 (m, 2 H), 8.12 (d, 1 H, J
) 8.3 Hz); 13C NMR (75 MHz, CDCl3, {1H}) δ 23.84 (d, J CP
)
3
2
2
4.2 Hz), 54.11 (d, J CP ) 6.2 Hz), 74.44 (d, J CP ) 5.7 Hz),
122.98, 123.14, 125.28, 125.65, 126.29, 128.62, 128.85, 129.68,
133.69, 137.17 (d, J CP ) 5.2 Hz); UV (CH3CN): 262 (ꢀ 3.7 ×
104), 272 (ꢀ 6.0 × 104), 282 (ꢀ 7.0 × 104), 292 (ꢀ 4.8 × 104);
GC-EIMS (70 eV) m/z (relative intensity) 280 [M]+ (14), 155
[CHCH3C10H7]+ (41), 154 (95), 153 (100), 127 [C10H7]+ (45);
HRMS (EI) m/z [M]+ calcd for C14H17O4P: 280.0865, found:
280.0879. Anal. Calcd for C14H17O4P: C, 60.00; H, 6.11.
Found: C, 60.10; H, 6.15.
P r ep a r a tion of 2-(1-(4-Acetylp h en yl)eth yloxy)-2-oxo-
1,3,2-d ioxa p h osp h or in a n e (8-O). Compound 8 (1.4 g, 5.2
mmol) was oxidized with tert-butyl hydroperoxide (1.0 mL) as
described earlier. HPLC purification yielded a colorless thick
oil (1.2 g, 4.4 mmol, 62%). 31P NMR (121 MHz, CDCl3, {1H}) δ
Dim eth yl 1-(1-n a p h th yl)eth ylp h osp h on a te (13) was
isolated from the products of photolysis of 1-(1-naphthyl)ethyl
dimethyl phosphite (9) by HPLC (1.5% methanol in chloro-
form). 31P NMR (121 MHz, CDCl3, {1H}) δ 32.82; 1H NMR (300
1
129.33; H NMR (300 MHz, CDCl3) δ 1.57 (d, 3 H, 3J ) 6.4
3
Hz), 1.62-1.75 (m, 1 H), 2.40 (m, 1 H), 2.51 (d, 3 H, J ) 0.7
Hz), 3.60-3.70 (m, 1 H), 3.71-3.82 (m, 1 H), 4.21-4.31 (m, 1
H), 4.40-4.50 (m, 1 H), 5.34 (dq, 1 H, J ) 8.7, 6.6 Hz), 7.51 (d,
2 H, J ) 8.3 Hz), 7.96 (d, 2 H, J ) 8.3 Hz); 13C NMR (75 MHz,
3
MHz, CDCl3) δ 1.70 (dd, CHCH3, J HP ) 18.3 Hz, 3J ) 7.3
3
3
Hz), 3.34 (d, 3 H, J HP ) 10.5 Hz), 3.68 (d, 3 H, J HP ) 10.5
CDCl3, {1H}) δ 25.70 (d, 3J CP ) 4.0 Hz), 25.51, 29.30 (d, 3J CP
)
Hz), 4.11 (dq, 1 H, J HP ) 23.2 Hz, J HH ) 7.5 Hz), 7.43-7.55
(m, 3 H), 7.70-7.85 (m, 3 H), 8.08 (d, 1 H, J ) 8.6); 13C NMR
2
3
5.2 Hz), 59.54 (d, 2J CP ) 1.0 Hz), 59.64 (d, 2J CP ) 1.6 Hz), 69.74
2
2
(d, J CP ) 20.2 Hz), 126.50, 128.35, 134.15, 148.12 (d, J CP
)
(75 MHz, CDCl3, {1H}) δ 16.20 (d, J CP ) 5.2 Hz), 31.81 (d,
2
2
3.1 Hz), 198.60 (s); UV (CH3CN): 282 (ꢀ 1438), 288 (ꢀ 948),
292 (ꢀ 750), 318 (ꢀ 160); GC-EIMS (70 eV) m/z (relative
intensity) 284 [M]+ (100), 253 [M - CH3]+ (50), 225 [M - Od
CCH3]+ (23), 43 [OdCCH3]+ (25). Anal. Calcd for C13H17O5P:
C, 54.93; H, 6.03. Found: C, 54.87; H, 6.00.
1J CP ) 139.0 Hz), 52.58 (d, J CP ) 7.3), 53.07 (d, J CP ) 7.4),
122.75, 125.30 (d, J CP ) 3.6 Hz), 125.39, 125.83 (d, J CP ) 6.7
Hz), 126.04, 127.50 (d, J CP ) 3.6 Hz), 128.78 (d, J CP ) 1.0 Hz),
131.32 (d, J CP ) 7.3 Hz), 133.65 (d, J CP ) 1.2 Hz), 133.73 (d,
J CP ) 6.2 Hz); UV (CH3CN): 264 (ꢀ 3.9 × 104), 274 (ꢀ 6.2 ×
104), 284 (ꢀ 7.3 × 104), 294 (ꢀ 5.0 × 104); GC-EIMS (70 eV) m/z
(relative intensity) 264 [M]+ (15), 155 [CHCH3C10H7]+ (100);
HRMS (EI) m/z [M]+ calcd for C14H17O3P: 264.0915, found:
264.0924. Anal. Calcd for C14H17O3P: C, 63.63; H, 6.48.
Found: C, 63.89; H, 6.52.
2-(1-(4-Acetylp h en yl)eth yl)-2-oxo-1,3,2 d ioxa p h osp h o-
r in a n e (12). The title compound was isolated following
photolysis of 8 on HPLC (1.5% methanol in chloroform).
Recrystallization from ether/pentane yielded 12 as white
needles: mp 100-101 °C. 31P NMR (121 MHz, CDCl3, {1H}) δ
24.34; 1H NMR (300 MHz, CDCl3) δ 1.65 (dd, 3H, 3J HP ) 18.4,
2
3J ) 7.5 Hz), 1.87 (m, 2 H), 2.60 (s, 3 H), 3.38 (dq, 1 H, J HP
)
Ack n ow led gm en t. This research was supported by
grants from the National Science Foundation and the
National Institutes of Health.
3
22.5, J ) 7.4 Hz), 4.12 (m, 2 H), 4.46 (m, 2 H), 7.46 (m, 2 H),
7.94 (d, 2 H); 13C NMR (75 MHz, CDCl3, {1H}) δ 14.86 (d, 2J CP
) 5.2 Hz), 26.1 (d, 3J CP ) 7.8 Hz), 26.51, 37.88 (d, 1J CP ) 135.4
Hz), 66.19 (d, J CP ) 7.3 Hz), 128.42 (d, J CP ) 2.6 Hz), 128.72
(d, J CP ) 6.2 Hz), 135.88 (d, J CP ) 3.6 Hz), 143.00 (d, J CP
2
)
Su p p or tin g In for m a tion Ava ila ble: Experimental de-
tails and spectral data for preparations of precursor to phos-
phites 7-9. Tables showing the configurational stabilities of
unreacted phosphites 7 and 9 over the course of irradiation.
This material is available free of charge via the Internet at
http://pubs.acs.org.
7.3 Hz), 197.60 (s); UV (CH3CN): 280 (ꢀ 1428), 286 (ꢀ 938),
290 (ꢀ 740), 316 (ꢀ 150); GC-EIMS (70 eV) m/z (relative
intensity) 268 [M]+ (49), 254 (14), 253 [M - CH3]+ (100), 147
(35) Baldwin, B. W.; Morrow, C. J . Tetrahedron Asymmetry 1996,
7, 2871-2878.
(36) Westley, J . W.; Halpern, B. J . Org. Chem. 1968, 33, 3978-3980.
J O001545E