Notable sulfur atom effects on the regio- and stereoselective formation of oxetanes in Paterno-Buchi photocycloaddition of aromatic aldehydes with silyl O,S-ketene acetals

Article abstract of DOI:10.1021/ja993997i

Paterno-Buchi photocycloaddition of silyl O,S-ketene acetals (SKA) 1 and aromatic aldehydes 2 was investigated in detail. The photoreaction of aldehydes 2a-d with β,β-dimethyl-O,S-SKA 1a-e has been found to give, regio- (ca. 70/30 to 90/10) and stereoselectively (ca. 70/30 to 90/10), trans-3-siloxyoxetanes 3 independent upon the aldehyde, the substituents SR¹ and SIR₃, and reaction medium (solvent and salts). The triplet 2- oxatetramethylene 1,4-diradical T-1,4-DR is reasonably proposed as a common intermediate. The regioselectivity is rationalized by (1) the relative stability of 1,4-diradicals and (2) the relative nucleophilicity of sp²- carbons in O,S-SKA. The trans-selectivity is explained by the sulfur atom effects in O,S-SKA 1, which control the approach direction of the electrophilic oxygen of triplet nπ* aldehydes to the nucleophilic alkene. The fast ISC process of the triplet 1-alkylthio-1-siloxy-2-oxatetramethylene 1,4-diradical T-1,4-DR in competition with the bond rotation has been proposed. The substrate-dependent formation of 3fb from E- or Z-1f supports the hypothesis. The S-directed regio- and diastereoselectivity are found for the first time in the present study.

Full text of DOI:10.1021/ja993997i

J. Am. Chem. Soc. 2000, 122, 4005-4010
4005
Notable Sulfur Atom Effects on the Regio- and Stereoselective
Formation of Oxetanes in Paterno`-Bu¨chi Photocycloaddition of
Aromatic Aldehydes with Silyl O,S-Ketene Acetals
Manabu Abe,* Kiyotada Fujimoto, and Masatomo Nojima
Contribution from the Department of Materials Chemistry, Graduate School of Engineering, Osaka
UniVersity, Suita 565-0871, Osaka, Japan
ReceiVed NoVember 12, 1999. ReVised Manuscript ReceiVed February 22, 2000
Abstract: Paterno`-Bu¨chi photocycloaddition of silyl O,S-ketene acetals (SKA) 1 and aromatic aldehydes 2
was investigated in detail. The photoreaction of aldehydes 2a-d with â,â-dimethyl-O,S-SKA 1a-e has been
found to give, regio- (ca. 70/30 to 90/10) and stereoselectively (ca. 70/30 to 90/10), trans-3-siloxyoxetanes 3
independent upon the aldehyde, the substituents SR1 and SiR3, and reaction medium (solvent and salts). The
triplet 2-oxatetramethylene 1,4-diradical T-1,4-DR is reasonably proposed as a common intermediate. The
regioselectivity is rationalized by (1) the relative stability of 1,4-diradicals and (2) the relative nucleophilicity
of sp2-carbons in O,S-SKA. The trans-selectivity is explained by the sulfur atom effects in O,S-SKA 1, which
control the approach direction of the electrophilic oxygen of triplet nπ* aldehydes to the nucleophilic alkene.
The fast ISC process of the triplet 1-alkylthio-1-siloxy-2-oxatetramethylene 1,4-diradical T-1,4-DR in
competition with the bond rotation has been proposed. The substrate-dependent formation of 3fb from E- or
Z-1f supports the hypothesis. The S-directed regio- and diastereoselectivity are found for the first time in the
present study.
Introduction
Scheme 1
Scheme 2
Regio- and stereoselectivity on the formation of oxetanes in
photochemical [2+2] cycloaddition, the so-called Paterno`-
Bu¨chi reaction,1 of carbonyl compounds with alkenes have
attracted considerable attention from the synthetic and mecha-
nistic points of view. Numerous inter- and intramolecular
reactions have been reported in the past four decades, which
reveal in general that a 1,4-diradical (2-oxatetramethylene-1,4-
diyl) is the key intermediate formed in the reaction of electro-
philic oxygen of the excited state of carbonyls with electron-
rich alkenes (Scheme 1).2
From the synthetic point of view, the regio- and stereose-
lective formation of oxetanes is still a challenging subject.2b,3
In this regard, we have recently reported the regioselective
formation of 2-siloxyoxetanes in the photochemical reaction of
acyclic and cyclic silyl O,O-ketene acetals (O,O-SKA) with
* Corresponding author. Fax: +81-6-6879-7928. E-mail: abe@
ap.chem.eng.osaka-u.ac.jp.
aromatic carbonyl compounds via the electron-transfer reaction
(Scheme 2).4 During the course of our continuous investigation
on the Paterno`-Bu¨chi reaction, we have found the regio- and
stereoselective formation of 3-siloxyoxetanes in the photore-
action of benzaldehyde (Ar ) Ph, R2 ) H) with silyl O,S-ketene
acetals (O,S-SKA) (Scheme 2).
The sharp contrast results on the regioselectivity (2-siloxy
vs 3-siloxy) in the formation of siloxyoxetanes have prompted
us to investigate the notable sulfur atom effects on the oxetane
formation in detail. Herein, we report our study on the
cycloaddition of aromatic aldehydes 2 with O,S-SKA 1.
(1) (a) Paterno`, E.; Chieffi, G. Gazz. Chim. Ital. 1909, 39, 341. (b) Bu¨chi,
G.; Inman, C. G.; Lipinsky, E. S. J. Am. Chem. Soc. 1954, 76, 4327.
(2) (a) Arnold, D. R. AdV. Photochem. 1968, 6, 301. (b) Carless, H. A.
In Synthetic Organic Photochemistry; Horspool, W. M., Ed.; Plenum
Press: New York, 1984; p 425. (c) Johnston, L. J.; Scaiano, J. C. Chem.
ReV. 1989, 89, 521. (d) Griesbeck, A. G.; Mauder, H.; Stadtmu¨ller, S. Acc.
Chem. Res. 1994, 27, 70. (e) Shimizu, N.; Ishikawa, M.; Ishikura, K.;
Nishida, S. J. Am. Chem. Soc. 1974, 96, 6456. (f) Turro, N. J.; Wriede, P.
A. J. Am. Chem. Soc. 1970, 92, 320. (g) Shroeter, S. H.; Orlando, C. M. J.
Org. Chem. 1969, 34, 1181. (h) Dalton, J. C.; Tremont, S. J. J. Am. Chem.
Soc. 1975, 97, 6916. (i) Schore, N. E.; Turro, N. J. J. Am. Chem. Soc.
1975, 97, 2482. (j) Wilson, R. M.; Wunderly, S. W. J. Chem. Soc., Chem.
Commun. 1974, 461. (k) Wilson, R. M.; Wunderly, S. W. J. Am. Chem.
Soc. 1974, 96, 7350. (l) Adam, W.; Kliem, U.; Lucchini, T. Tetrahedron
Lett. 1986, 27, 2953.
(4) (a) Abe, M.; Ikeda, M.; Shirodai, Y.; Nojima, M. Tetrahedron Lett.
1996, 37, 5901. (b) Abe, M.; Shirodai, Y.; Nojima, M. J. Chem. Soc., Perkin
Trans. 1 1998, 3253. (c) Abe, M.; Ikeda, M.; Nojima, M. J. Chem. Soc.,
Perkin Trans. 1 1998, 3261.
(3) (a) Porco, J. A.; Schreiber, S. L. In ComprehensiVe Organic Synthesis;
Trost, B. M., Ed.; Pergamon Press: New York, 1991; Vol. 5, p 168. (b)
Bach, T. Synthesis 1998, 683. (c) Bach, T. Liebigs Ann./Recl. 1997, 1627.
10.1021/ja993997i CCC: $19.00 © 2000 American Chemical Society
Published on Web 04/14/2000

4006 J. Am. Chem. Soc., Vol. 122, No. 17, 2000
Abe et al.
Table 2. Solvent and Salts Effects on the Formation of
Scheme 3
Siloxyoxetanes 3 and 5 in the Photoreaction of Benzaldehyde
Derivatives 2 with O,S-KSA 1aa
products and yields (%)
trans-3/
cis-3
entry
2
solvent
salts
trans-3 cis-3
5
6
3/5
1b 2a CH3CN none
30
23
23
50
46
48
44
41
37
22
39
11 15 15 73/27
11 15 16 68/32
10 14 13 71/29
12 15 21 80/20
73/27
69/31
72/28
81/19
84/16
91/9
2
3
2a CH2Cl2 none
2a C6H6
none
4b 2b CH3CN none
5c 2b CH3CN Et4NClO4
9
13
12
10 15 83/17
6
2b CH2Cl2 none
6
5
15 79/21
16 78/22
7c 2b CH2Cl2 Bu4NPF6
92/8
8
2b C6H6
none
11 18 17 79/21
20 22 21 65/35
15 11 16 63/37
74/26
72/28
70/30
80/20
9b 2c CH3CN none
10 2c CH2Cl2 none
11 2c C6H6
none
17 22
0 69/31
a The photoreaction through a Pyrex filter (>290 nm) of O,S-KSA
1a (100 mM) and benzaldehyde derivatives 2 (50 mM) was performed
at 0 °C for 4 h. b The data were taken from Table 1. c The photoreaction
was performed in the presence of the salts (500 mM).
Table 1. Paterno`-Bu¨chi Coupling of O,S-KSA 1 and Aromatic
Aldehydes 2 in CH3CN at 0 °Ca
products and yields (%)b
cis-3), but also the regioselectitivy (3/5). In the photoreaction
with O,S-SKA 1c (R1 ) Me, SiR3 ) TIPS) (entry 6), alcohol
7cb (SR1 ) Me, SiR3 ) TIPS, Ar ) p-CNC6H4) was isolated
as a minor product (2%). For the other photoreactions, a trace
amount of the alcohol seems to be formed, but the structural
assignment was discarded due to the low yield. The formation
of alcohol 7 suggest the hydrogen abstraction by the triplet nπ*
carbonyl on the allylic hydrogen of O,S-SKA 1.
entry
1
2
trans-3 cis-3
5
6
trans-3/cis-3c 3/5d
1
2
3
4
5
6f
7
8
1a 2a
1a 2b
1a 2c
1a 2d
1b 2b
1c 2b
1d 2b
1e 2b
30
50
37
18
32
38
58
42
11
12
20
5
15
15
21
21
0
30
22
25
25
73/27
80/20
65/35
78/22
82/18
67/33
88/12
91/9
73/27
81/19
72/28
15
22
trace
13
14
9
e
-
8
75/25
80/20
88/12
73/27
18
8
4
17
As mentioned above, interestingly, the photoreaction of O,S-
SKA 1 gave preferentially trans-3-siloxyoxetanes 3 independent
of the aldehydes 2, R1 and SiR3 substituents, which is in sharp
contrast to the results of the predominant formation of 2-sil-
oxyoxetanes in the photoreaction with O,O-SKA (Scheme 2).
Solvent and Salts Effects. The mechanism of a Paterno`-
Bu¨chi type reaction has been found to be largely dependent upon
the combination of the substrates.5 The electron-transfer pathway
is the most promising mechanism, if the electron transfer is
exothermic, i.e. ∆Get ()Eox - Ered)6 < 0. To understand the
mechanism of our Paterno`-Bu¨chi photocycloaddition of
O,S-SKA 1 and aldehydes 2, the oxidation potential of O,S-
SKA 1a (SiR3 ) TBDMS, SR1 ) SMe) was measured by using
cyclic voltanmmetry (CV). The irreversible oxidation wave (Ep)
was observed at 1.22 V vs SCE. Judging from the known
reduction potentials4b (E*red) of the triplet excited state of the
aromatic carbonyl compounds (E*red ) ET - Ered ) 1.16 V for
2a (Ar ) Ph), 1.51 V for 2b (Ar ) p-CNC6H4), and 0.86 V for
2c (Ar ) p-MeOC6H4)), the electron transfer between O,S-SKA
1a and p-cyanobenzaldehyde 2b seems to be possible in
acetonitrile used for the photoreactions. Since electron-transfer
reactions have been found to be largely influenced by the
reaction medium,7 i.e. solvent (dielectric constant ꢀ) and the
presence of salts, the medium effects on the product distribution
were examined in the photoreaction with O,S-SKA 1a (R1 )
Me, SiR3 ) TBDMS) (Table 2).
a The photoreactions were performed with a high-pressure Hg lamp
through a Pyrex filter (>290 nm) at 0 °C for 4 h. b Isolated yields (%)
after column chromatography on silica gel. c The ratios were obtained
by the H NMR (270 MHz) peak areas. d The values were based on
1
the isolated yields of 3 and 5. e The ratio was not determined due to
the low yield of 5ad. f The adduct 7cb (Ar ) p-CNC6H4, SR1 ) SMe,
SiR3 ) TIPS) was isolated in 2% yield.
Results
Photoreaction of Silyl O,S-Ketene Acetals 1a-e with
Aromatic Aldehydes 2a-d. A degassed acetonitrile solution
of aromatic aldehydes 2a-d (0.05 M) and silyl O,S-ketene
acetals 1a-e (0.1 M) was irradiated with a high-pressure Hg
lamp (300 W) through a Pyrex filter (>290 nm) at 0 °C (Scheme
3 and Table 1).
The photoreactions of benzaldehyde derivatives 2a-d with
O,S-SKA 1a (SR1 ) Me, SiR3 ) TBDMS) were found to give
3-siloxyoxetane 3 (trans-3/cis-3 ) ca. 70/30-80/20; the con-
figuration was determined by the 1H NMR NOE measurements,
vide infra) as a major product together with the aldol-type adduct
5 and the corresponding pinacol 6 (entries 1-4). Interestingly,
the diastereoselectivities, trans-3/cis-3, were not largely de-
pendent upon the electronic and steric effects of the substituents
on the aryl group. The aldol-type adduct 5 may be formed from
the acid-labile 2-siloxyoxetane 4 under the isolation conditions
(silica gel), since the direct formation of the adduct 5 could not
be detected by the 1H NMR analyses (270 MHz) of the
photolysate. Our previous findings on the formation of aldol-
type adducts derived from 2-siloxyoxetanes strongly support
the transformation.4 Thus, the regioselectivity (3-siloxyoxetane
3/2-siloxyoxetane 4) was estimated from the product ratios of
3/5 (ca. 70/30 to 90/10) as shown in Table 1. In the photore-
action of p-cyanobenzaldehyde 2b (entry 2), a high mass balance
was observed (mass balance 98%). Thus, the substituent effects
of R1 and the SiR3 group in the photoreaction were examined
with p-cyanobenzaldehyde 2b (entries 5-8).
(5) (a) Eriken, J.; Plith, P. E. Tetrahedron Lett. 1982, 23, 481. (b) Mattay,
J.; Gersdorf, J.; Buchkremer, K. Chem. Ber. 1987, 120, 307. (c) Gersdorf,
J.; Mattay, J.; Go¨rner, H. J. Am. Chem. Soc. 1987, 109, 1203.
(6) (a) Weller, A. Pure Appl. Chem. 1968, 16, 115. (b) Rehm, D.; Weller,
A. Isr. J. Chem. 1970, 8, 259. (b) Weller, A. Z. Phys. Chem. N. F. 1982,
130, 129.
(7) (a) Kavarnos, G. J.; Turro, N. J. Chem. ReV. 1986, 86, 401. (b)
Santamaria, J. In Photoinduced Electron Transfer; Fox, M. A., Chanon,
M., Eds.; Elsevier: Amsterdam, 1988; Part B, Chapter 3.1. (c) Mizuno,
K.; Otsuji, Y. Yuki Gousei Kagaku Kyoukaishi 1988, 47, 916. (d) Mizuno,
K.; Ichinose, N.; Otsuji, Y. Chem. Lett. 1985, 455. (e) Niwa, T.; Kikuchi,
K.; Matsusita, N.; Hayashi, M.; Katagiri, T.; Takahashi, Y.; Miyashi, T. J.
Phys. Chem. 1993, 97, 11960. (f) Hubig, S. M.; Sun, D.; Kochi, J. K. J.
Chem. Soc., Perkin Trans. 2 1999, 781. (g) Griesbeck, A. G.; Buhr, S.;
Fiege, M.; Schmickler, H.; Lex, J. J. Org. Chem. 1998, 63, 3847.
As shown in entries 5-8, the substituents, R1 and SiR3, did
not significantly affect only the diastereoselectivity (trans-3/

S Atom Effects on Formation of Oxetanes
J. Am. Chem. Soc., Vol. 122, No. 17, 2000 4007
Scheme 4
Scheme 5
Table 3. Paterno`-Bu¨chi Coupling of O,S-KSA E- or/and Z-1f
with p-Cyanobenzaldehyde 2b in CH3CN at 25 °Ca
diastereomer ratios of 3fbb
As shown in Table 2, the solvent polarity (CH3CN, ꢀ 37.5;
CH2Cl2, ꢀ 8.9; benzene, ꢀ 2.3) did not affect the product ratios,
i.e. trans-3/cis-3 and 3/5 (entries 1-3, 4, 6, and 8-11).
Furthermore, the salts (Et4NClO4, Bu4NPF6) gave no effects
(entries 5 and 7). The negligible solvent and salt effects on the
diastereoselectivity (trans-3/cis-3) and the regioselectivity
(3/5) suggest that the electron-transfer mechanism between the
triplet excited state of aldehydes 2 and O,S-SKA 1 is unlikely
at least for the formation of oxetanes observed.
DPA-Photosensitized Reaction of p-Cyanobenzaldehyde
2b with O,S-SKA 1a. To ensure the mechanism for the
formation of siloxyoxetanes 3 and 4, the 9,10-diphenylan-
thracene (DPA, ES ) 3.15 eV, Eox ) 1.22 V, Ered ) -1.94 V
vs SCE)8-sensitized photoreaction (>400 nm) of O,S-SKA 1a
with p-cyanobenzaldehyde 2b (Ered ) -1.38 V vs SCE) was
performed (eq 1). Under such conditions, the electron-transfer
trans,trans- cis,trans- trans,cis- cis,cis-
entry
1
1f
3fb
3fb
3fb
3fb
E-1f
96
4
0
0
(E/Z ) 90/10)
2
3
Z-1f
50
72
4
6
20
10
26
12
(E/Z ) 10/90)
E,Z-1f
(E/Z ) 47/53)
a A solution of 2 (0.05 M) and O,S-KSA E- or/and Z-1f (0.5 M)
was irradiated by using a Xenon lamp (350 ( 5 nm) at room
temperature (ca. 25 °C). All of the reactions were stopped at ca. 50%
consumption of the aldehyde 2b. b The isomer ratios, normalized to
100%, were determined by 1H NMR (270 MHz) peak area, error (5%.
or/and Z-1f have been chosen and reacted with p-cyanobenzal-
dehyde 2b (eq 2, Table 3).
reaction would be the most promising process as judged by their
redox potentials, generating the cation radical of 1a and anion
radical of 2b (Scheme 4).9 After irradiation for 15 h (conversion
of aldehyde 2b, 100%), the aldol-type adduct 5ab (Ar )
p-CNC6H4) was the only isolable product (79%). We could not
detect any trace amounts of 3-siloxyoxetane 3ab. Thus, the
predominant formation of 2-siloxyoxetane 4ab should be
expected under electron-transfer conditions. This result is quite
consistent with the exclusive formation of 2-siloxyoxetanes
derived from O,O-SKA (Eox ) 0.90 V vs SCE)4b,10 with much
lower oxidation potential than O,S-SKA 1 (Eox ) 1.22 V vs
SCE), as shown in Scheme 2.
Since the photochemical isomerization between O,S-SKAs
E- and Z-1f was observed, the photoreactions by using Xenon
lamp (hV ) 350 ( 5 nm) were performed in the presence of
large excess amounts of E- or/and Z-1f (10 equiv) and stopped
at ca. 50% conversion of the aldehyde 2b, to minimize the
change of the isomer ratios (E-/Z-1f) during the course of the
Under the direct irradiation of aromatic aldehydes 2, 3-si-
loxyoxetanes 3 were major products (Table 1). Thus, 3-siloxy-
oxetanes 3 should be derived from the triplet 1,4 diradicals
(T-1,4-DR) formed directly or via exciplex between the nπ*
triplet excited state of aldehydes 2 and O,S-SKA 1 (Scheme
5), not by the coupling between the radical ion pair. Because if
the direct electron transfer is occurred, the resulting radical ion
pair would give the 2-siloxyoxetanes 4 as shown in Scheme 4.
Paterno`-Bu1chi Coupling of p-Cyanobenzaldehyde 2b and
O,S-SKA E- or/and Z-1f. To understand the trans-selective
formation of 3-siloxyoxetane 3 shown above, O,S-SKA E-
reaction. Unfortunately, the four possible diastereomers of
3-siloxyoxetanes 3fb, trans,trans-, cis,trans-, trans,cis-, and
cis,cis-, could not be perfectly separated by using HPLC
analysis. Thus, the relatively high conversions (50%) of
aldehyde 2b were needed to determine the product ratios by
1H NMR (270 MHz) analyses. Fortunately, all of the benzylic
protons Ha were well distinguished in d6-benzene, thus, the
determination of the accurate ratios (error (5%) was feasible
(Table 3, the configurational determination will be discussed
below). As shown in Table 3, the obtained ratios were largely
dependent upon the configuration of the starting material 1f.
When the E-isomer E-1f (E/Z ) 90/10) was used for the
(8) Murov, S. L.; Carmichael, I.; Hug, G. L. In Handbook of Photo-
chemistry; Marcel Dekker: New York, 1993.
(9) The fluorescence of DPA was actually quenched by p-cyanobenzal-
dehyde 2b, but not by O,S-SKA 1a. For the generation of radical ion pairs
by means of such a photosensitization, see; (a) Majima, T. Pac, C.; Sakurai,
H. J. Am. Chem. Soc. 1980, 102, 5265. (b) Pac, C. Pure Appl. Chem. 1986,
58, 1249.
(10) Fukuzumi, S.; Fujita, M.; Otera, J.; Fujita, Y. J. Am. Chem. Soc.
1992, 114, 10271.

4008 J. Am. Chem. Soc., Vol. 122, No. 17, 2000
Abe et al.
photoreaction, only trans,trans- and cis,trans-3fb (75% yield)
were detected by 1H NMR, and the isomer ratio was 96/4 (entry
1, final ratio of E-/Z-1f ) 88/12). Namely, the configuration of
the starting material E-1f was perfectly retained in the products.
The trans,trans-isomer could be isolated as a pure form and
fully characterized. The triplet quenching experiment was also
performed in the presence of 1,3-pentadiene (0.5 M, ET ) 59
kcal/mol)8 under similar irradiation conditions. The formation
of the oxetane 3fb (ca. 10%) was suppressed, and the starting
materials 1f and 2b were recovered in reasonable yields. Thus,
the 3-siloxyoxetanes 3 are mainly formed by the triplet excited
state of aldehydes 2 (ET ) ca. 70 kcal/mol).8
Alternatively, in the photoreaction with Z-1f (E/Z ) 10/90),
all of the four possible isomers (total yield, 72%) were observed
(entry 2, trans,trans-, cis,trans-, trans,cis-, and cis,cis-3fb )
50/4/20/26, final ratio of E-/Z-1f ) 15/85). The trans,cis- and
cis,cis-isomers could be isolated after several trials of column
chromatography on silica gel. Unfortunately, the cis,trans-isomer
could not be obtained as a pure form, but with the cis,cis-isomer
(cis,cis/cis,trans ) 2/3).
Finally, the competitive reaction of E- and Z-1f (E/Z )
47/53) was performed (entry 3). The product ratio obtained in
the reaction was found to be 72/6/10/12 (total yields, 68%, final
ratio of E/Z ) 51/49). The results suggest that the reaction of
aldehyde 2b with E-1f is much faster than that with the
Z-isomer. If we assume that all of the trans,trans- and
cis,trans-3fb come from the E-1f, the reaction with the E-isomer
is about four times faster than that with the Z-isomer
(trans,trans- + cis,trans-3fb/trans,cis- + cis,cis-3fb ) 78/22).
Thus, we feel that the production of trans,trans- and cis,trans-
3fb in the photoreaction with Z-1f (initial ratio, E/Z ) 10/90;
final ratio, E/Z ) 15/85) mainly derived from E-1f, which is
already included as a minor component in the starting materials.
The substrate-dependent reactions clearly suggest that the pro-
duct ratios were not controlled by the relative thermodynamic
stability of the 3-siloxyoxetanes 3fb formed in the photoreac-
tion.11
Figure 1.
Similarly, we could determine the configuration of the four
diastereomers of 3fb. Namely, the isomer with the clear NOE
enhancements between Ha and Hb (3%), and also Hb and StBu
(6%), was determined as trans,trans-3fb. In the same way, the
stereochemical determination for trans,cis- and cis,cis-3fb was
done as depicted in eq 2. Thus, the rest of isomer could be
determined as the cis,trans-isomer.
Discussion
Regioselectivity: 3-Siloxy- vs 2-Siloxyoxetane Formation.
As mentioned above, we have concluded that 3-siloxyoxetanes
3 may be formed via triplet 1,4-diradical species. The conclusion
is quite reasonable from experimental results: (1) the regiose-
lectivity, 3/5, was not dependent upon the reduction potential
of the excited carbonyl compounds 2 (see, entries 2-4 in Table
1) and (2) the insensible solvent and salt effects on the
regioselectivity, 3/5 in Table 2, (3) the selective formation of
2-siloxyoxetane 4 in the DPA-sensitized electron-transfer reac-
tion (eq 1), and (4) the formation of oxetanes 3, 4 were
effectively suppressed by the presence of 1,3-pentadiene (triplet
quencher).
The question quickly arose; Why was the 3-siloxyoxetane 3
selectively formed via triplet 1,4-diradical T-1,4-DR? In this
regard, we should consider the following two effects:12 (1) the
relative stability of the regioisomeric diradicals, T-1,4-DR-1
vs T-1,4-DR-2 and (2) the relative nucleophilicity of the C1
and C2 carbons of O,S-SKA 1. To elucidate the first point,
semiempirical calculations (PM3/UHF, Triplet)13,14 were per-
formed for the two triplet states of diradicals, T-1,4-DR-1 vs
T-1,4-DR-2 (SiR3 ) TMS, SR1 ) SMe) as models (Figure 1).
Consequently, the heat of formation (∆Hf) of T-1,4-DR-1, which
leads to 3-siloxyoxetane, was found to be lower than that of
T-1,4-DR-2 at such a level of theory, ∆∆Hf ) ca. 3.0 kcal/
mol. The calculation results suggest that the diradical T-1,4-
DR-1 (optimized structure, see below) is more stable than the
regioisomeric diradical T-1,4-DR-2.
Configurational Determination of 3-Siloxyoxetanes 3. First
of all, the stereochemical determination of the 2,2-dimethyl-3-
siloxyoxetanes 3 formed from O,S-SKA 1a-e is discussed on
the basis of both the NOE measurements and the comparison
of the chemical shifts of SiR3 and SR1 groups. The isomer with
clear NOE enhancements (in the range of 3-7%) between Hc
and the protons of SR1 was determined as the trans-isomer.
Alternatively, in the cis-isomer the clear NOE enhancements
between Hc and the protons of silyl group were found in the
range of 2-5%. Furthermore, in the trans-isomer, the upfield
chemical shift (ca. 0.5 ppm) of the protons of the SiR3 group
and the low-field shift (ca. 0.3 ppm) of those of the SR1 group
were observed in comparison with those in the cis-isomer (see
the Experimental Section in the Supporting Information). Thus,
the stereochemical determinations for all of the 3-siloxyoxetanes
3 were feasible.
Next, to ensure the relative nucleophilicity of C1 and C2
carbons, the HOMO coefficients were calculated for O,S-SKA
1b (Figure 2). The relatively large coefficient at C2 (+0.54)
was calculated, compared to that at C1 (+0.32). The obtained
results are consistent with experimental observations, i.e. the
Mukaiyama aldol reaction.15 Additionally, a large HOMO
coefficient at the sulfur atom (-0.65) with the opposite phase
was also found, while the oxygen of the siloxy group has an
almost negligible coefficient. The significant difference of the
relative nucleophilicity suggests that the electrophilic oxygen
of the triplet excited carbonyls 2 would attack preferentially
the C2 carbon of O,S-SKA 1 to generate 1,4-diradicals, e.g.
T-1,4-DR-1.
(12) For the concept for regioselectivity of Paterno` -Bu¨chi reactions,
see; (a) Turro, N. J. Modern Molecular Photochemistry; Benjamin/
Cummings Publishing Co., Inc.: Menlo Park, 1978; p 432. (b) Sengupta,
P.; Chardra, A. K.; Nguyen, M. T. J. Org. Chem. 1997, 62, 6404.
(13) Stewart, J. J. P. J. Comput. Chem. 1989, 10, 209.
(14) PM3 calculations were run on a CAChe system, Sony/Tektonix Co.
(15) (a) Mukaiyama, T.; Banno, K.; Narasaka, K. J. Am. Chem. Soc.
1974, 96, 7503. (b) Mukaiyama, T. Org. React. 1982, 28, 203.
(11) Stereoselective formation of siloxyoxetanes controlled by product
stability, see; (a) Bach, T.; Tetrahedron Lett. 1991, 32, 7037. (b) Bach, T.;
Jo¨dicke, K. Chem. Ber. 1993, 126, 2457. (c) Bach, T. Tetrahedron Lett.
1994, 35, 5845. (d) Bach, T. Liebigs Ann. 1995, 855. (e) Bach, T.; Jo¨dicke,
K.; Kather, K.; Fro¨hlich, R. J. Am. Chem. Soc. 1997, 119, 2437.

S Atom Effects on Formation of Oxetanes
J. Am. Chem. Soc., Vol. 122, No. 17, 2000 4009
Figure 2.
Figure 3.
Scheme 6
If our hypothesis as shown in Scheme 6 is right, the ISC
from the triplet state T-1,4-DR to its singlet state S-1,4-DR
should be faster than both the C1-C2 and C2-O bond rotations
(Scheme 6). The situation is unusual since, in general, the rate
constant of C-C bond rotation in ethane derivatives is around
1010-12 s-1 19
, while the rate constant for the spin-forbidden ISC
process is reported to be 106-8 s-1 for carbon chained 1,3- and
1,4-diradicals.20 Thus, a C-C bond rotation should be much
faster than the ISC process, which is a reason for stereo-random
formation of oxetanes via triplet diradicals.21 However, the
general consideration may not be applied to our case, if the
spin-orbit coupling (SOC) mechanism22 for the ISC process
is crucial for our system. The SOC mechanism has been
reasonably accepted for short distance 1,n-diradicals, i.e. n )
3, 4. Salem and Rowland have addressed theoretically the
following points expected to influence the SOC mechanism:
(1) the perpendicularity of two singly occupied orbitals, (2) the
distance between the radical cites, and (3) ionic contributions
to the singlet wave function.23 These factors may accelerate the
ISC process.24
On the basis of theoretical considerations, the ISC process
of our 2-oxatetramethyene 1,4-diradicals (T-1,4-DR) should be
much faster than that of pure hydrocarbon 1,4-diradicals for
the following reasons: (1) the spatial orientation, perpendicular
geometry,2e,25 of the two singly occupied orbitals in T-1,4-DR
is already the best for the ISC process by the spin-orbit
coupling (SOC) mechanism and (2) the ISC process may be
accelerated by the zwitterionic character (ZI) of the singlet
diradical S-1,4-DR (Figure 3), which may be induced by the
adjacent substituents (SR, OSiR3, Ar).
As judged by the above considerations, the regioselective
formation of oxetanes 3 may arise from these two points.16
Trans-Selective Formation of 3-Siloxyoxetanes 3. Judging
from the substrate-dependent formation of 3-siloxyoxetanes 3fb
depicted in eq 2 and Table 3, the trans-selective formation of
3-siloxyoxetanes 3 is not controlled by the thermodynamic
preference of the trans-isomer.11 Thus, we propose the hypo-
thetical mechanism for the trans-selective formation of 3-si-
loxyoxetanes 3 as shown in Scheme 6.
Namely, the perpendicular approach12a,17,18 of the electrophilic
oxygen of the triplet excited state of 2 preferentially occurred
from the sulfur-atom side of the O,S-SKA 1 to give the triplet
1,4 diradical T-1,4-DR. The sulfur (S)-directed approach of the
electrophilic oxygen is reasonably explained by the electrostatic
attraction between the sulfur atom and the electrophilic oxygen
(see the large HOMO coefficients at the sulfur atom (Figure
2)). Then, the intersystem crossing (ISC) to the singlet state
S-1,4-DR quickly occurred to afford the trans-isomer of
3-siloxyoxetane 3. The S-directed approach can be reasonably
supported by the fact that much higher diastereoselectivity
(trans,trans-/cis,trans-3fb ) 96/4) than the â,â-dimethyl sub-
stituted case, e.g. trans-/cis-3ab ) 80/20, was observed. Namely,
the direction of the S-directed approach of the carbonyl oxygen
for E-1f (R2 ) H, R3 ) Et) is also sterically favored. The
combined electrostatic and steric effects may increase the
diastereoselectivity derived from O,S-SKA E-1f.
The significantly short lifetime, ca. 1-5 ns,26 of 2-oxatet-
ramethylene-1,4-diyls has been reported in comparison with that
of the tetramethylene-1,4-diradicals, ca. 500 ns20b by flash
photolysis studies. Caldwell has suggested that a 102-103
acceleration for 2-oxatetramethylene-1,4-diradicals relative to
pure hydrocarbon biradicals would be expected in the ISC
process.20c He has also mentioned that the large rate acceleration
(19) Horn, B. A.; Herek, J. L.; Zeweil, A. H. J. Am. Chem. Soc. 1996,
118, 8755.
(20) (a) Adam, W.; Grabowski, S.; Wilson, R. M. Acc. Chem. Res. 1990,
23, 165. (b) Barton, D. H. R.; Charpiot, B.; Ingold, K. U.; Johnston, L. J.;
Motherwell, W. B.; Scaiano, J. C.; Stanforth, S. J. Am. Chem. Soc. 1985,
107, 3607. (c) Caldwell, R. A. In Kinetics and Spectroscopy of Carbenes
and Biradicals; Platz, M. S., Ed.; Plenum Press: New York, 1990; p 77.
(d) Kita, F.; Adam, W.; Jordan, P.; Nau, W. M.; Wirz, J. J. Am. Chem.
Soc. 1999, 121, 9265.
(21) For stereo-random formation of oxetanes, see; Turro, N. J.; Wriede,
P. A. J. Am. Chem. Soc. 1970, 92, 320.
(22) (a) Bonacic-Koutecky, V.; Koutecky, J.; Michl, J. Angew. Chem.,
Int. Ed. Engl. 1987, 26, 170. (b) Zimmerman, H. E.; Kutateladze, A. G.;
Maekawa, Y.; Mangette, J. E. J. Am. Chem. Soc. 1994, 116, 995. (c) Minaev,
B. F.; Jonsson, D.; Norman, P.; Agren, H. Chem. Phys. 1995, 194, 19.
(23) Salem, L.; Rowland, C. Angew. Chem., Int. Ed. Engl. 1972, 11, 92.
(24) For the semiempirical formula for the magnitude of SOC, see; (a)
Furlani, T. R.; King, H. F. J. Chem. Phys. 1985, 82, 5577. (b) Carlacci, L.;
Doubleday, C., Jr.; Furlani, T. R.; King, H. F.; Mclever, J. W., Jr. J. Am.
Chem. Soc. 1987, 109, 5323.
(25) (a) Griesbeck, A. G.; Stadtmu¨ller, J. Am. Chem. Soc. 1990, 112,
1281. (b) Griesbeck, A. G.; Stadtmu¨ller, S. Chem. Ber. 1990, 123, 357.
(26) (a) Freilich, S. C.; Peters, K. S. J. Am. Chem. Soc. 1981, 103, 6255.
(b) Caldwell, R. A.; Majima, T.; Pac, C. J. Am. Chem. Soc. 1982, 104,
629.
(16) For the regioselective formation of oxetanes from methyl vinylsul-
fides, see; (a) Morris, T. H.; Smith, E. H.; Walsh, R. J. Chem. Soc., Chem.
Commun. 1987, 964. (b) Khan, N.; Morris, T. H.; Smith, E. H.; Walsh, R.
J. Chem. Soc., Perkin Trans. 1 1991, 865.
(17) For the electrophilic attack, perpendicular approach, of excited
carbonyl to electron-rich alkenes, see; (a) Salem, L. J. Am. Chem. Soc.
1974, 96, 3486. (b) Bigot, B.; Devaquet, A.; Turro, N. J. J. Am. Chem.
Soc. 1981, 103, 6. (c) Palmer, I. J.; Ragazos, I. N.; Bernardi, F.; Olivucci,
M.; Robb, M. A. J. Am. Chem. Soc. 1994, 116, 2121.
(18) Griesbeck, A. G. In CRC Handbook of Organic Photochemistry
and Photobiology; Horspool, W. M., Song, P.-S., Eds.; CRC Press: New
York, 1995; p 550.

4010 J. Am. Chem. Soc., Vol. 122, No. 17, 2000
Abe et al.
(PM3/UHF).27 The conformational searches around ϑ1
(O-C2-C1-S) and ϑ2 (C1-C2-O-C3) dihedral angles are
shown in Figures 4 and 5. In reality, the optimized structure is
closely similar to the structure T-1,4-DR depicted in Scheme
6. Namely, the structure with (1) parallel alignment (ϑ1 ) ca.
+100°) between the C-O bond and the radical p orbital and
(2) gauche geometry (ϑ2 ) ca. (60°) was found to exist as an
energy minimum, with a rotational energy barrier of ca. 3-4
kcal/mol at this level of theory. The stereoelectronic effect, i.e.
the interaction between the SOMO of the alkoxyalkyl radical
and the adjacent C-O σ* orbital, can explain the parallel
alignment of the C-O bond with the p orbital.28 The results
clearly suggest that the conformation of diradicals T-1,4-DR
depicted in Scheme 6 has a lifetime. From these considerations,
our hypothesis, i.e. the rate of ISC process > C1-C2 and O-C2
bond rotations, may be possible.
Figure 4. Conformational search around O-C2-C1-S (ϑ1) in T-1,4-
DR-1.
Summary
We have examined the Paterno`-Bu¨chi photocycloaddition
of O,S-SKA 1 and aromatic aldehydes 2 in detail. The
regioselective formation of 3-siloxyoxetanes 3 has been reason-
ably explained by both the stability of the 1,4-diradical involved
and the interaction between the electrophilic oxygen of the triplet
nπ* aldehydes and the nucleophilic C2 carbon of the O,S-SKA.
The S-directed approach of the excited state of the aldehydes
and the successive ISC process have been proposed for the trans-
selective formation of 3-siloxyoxetanes 3. The notable sulfur
atom effects on the regio- and stereoselectivity in the formation
of siloxyoxetanes have been discovered for the first time. We
believe that our finding will open up the new concept of the
stereoselective formation of oxetanes.
Acknowledgment. This work was supported in part by the
grant-in-aid for Scientific Research on Priority Areas “Molecular
Physical Chemistry” and grant No. 10750610 from the Ministry
of Education, Science, Sports, and Culture of Japan. We thank
the Analytical Division of Faculty of Engineering, Osaka
University, for measuring elemental analyses and mass spectra.
Figure 5. Conformational search around O-C2-C1-C3 (ϑ2) in T-1,4-
DR-1.
Supporting Information Available: Experimental proce-
dure and spectral data (PDF). This material is available free of
charge via the Internet at http://pubs.acs.org.
JA993997I
may be due to the significant mixing of the oxygen lone pairs
and p orbitals at both radical termini. In our diradicals T-1,4-
DR, sulfur substituent also exists at radical termini. Thus, the
additional rate acceleration would be expected.
Next, we should be concerned with the conformational
stability of our diradicals T-1,4-DR with the perpendicular
orientation of the two p orbitals. If the conformation does not
exist as an energy minimum structure, our discussion seems to
make no sense. Thus, the triplet 2-oxatetramethylene 1,4-
diradical T-1,4-DR-1 has been chosen as a model compound
and calculated to estimate the relative stability in the possible
conformations by using the semiempirical MO method
(27) Excellent spin expectation S2 values of 2.04 were obtained for
the spin densities of these diradicals.
(28) (a) Dupius, J.; Giese, B.; Ru¨egge, D.; Fischer, H.; Korth, H.-G.;
Sustman, R. Angew. Chem., Int. Ed. Engl. 1984, 23, 896. (b) Korth, H.-G.;
Sustman, R.; Dupuis, J.; Gro¨ninger, K. S.; Witzel, T.; Giese, B. In
Substituent Effects in Radical Chemistry; Viehe, H. G., Janousek, Z.,
Merenyi, R., Eds.; D. Reidel Publishing Co.: Dordrecht, 1986; p 297. (c)
Giese, B.; Dupuis, J.; Gro¨ninger, K.; Hasskerl, T.; Nix, M.; Witzel, T. In
Substituent Effects in Radical Chemistry; Viehe, H. G., Janousek, Z.,
Merenyi, R., Eds.; D. Reidel Publishing Co.: Dordrecht, 1986; p 283. (d)
Fossey, J. In Substituent Effects in Radical Chemistry; Viehe, H. G.,
Janousek, Z., Merenyi, R., Eds.; D. Reidel Publishing Co.: Dordrecht, 1986;
p 87. (e) Kochi, J. K. In AdVances in Free-Radical Chemistry; Williams,
G. H., Ed.; Elek P.: London, 1975; Vol. 5, Chapter 4.