Anomalous C-C bond cleavage in sulfur-centered cation radicals containing a vicinal hydroxy group

Article abstract of DOI:10.1021/jo035001z

1,3-Dithianyl cation radicals having α-hydroxy-neopentyl or similar groups in position 2, which are generated via oxidative photoinduced electron transfer, undergo anomalous fragmentation necessitating refinement of the accepted mechanism. Experimental and computational data support a rationale in which proton abstraction from the hydroxy group in the initial cation radical does not cause a Grob-like fragmentation, but rather produces a neutral radical species, the alkoxy radical, that undergoes fragmentation in either direction, i.e., cleaving the C-C bond to dithiane or to the tertiary alkyl group.

Full text of DOI:10.1021/jo035001z

SCHEME 1
An om a lou s C-C Bon d Clea va ge in
Su lfu r -Cen ter ed Ca tion Ra d ica ls
Con ta in in g a Vicin a l Hyd r oxy Gr ou p
Zaiguo Li and Andrei G. Kutateladze*
Department of Chemistry and Biochemistry,
University of Denver,
Denver, Colorado 80208-2436
akutatel@du.edu
Received J uly 11, 2003
Abstr a ct: 1,3-Dithianyl cation radicals having R-hydroxy-
neopentyl or similar groups in position 2, which are gener-
ated via oxidative photoinduced electron transfer, undergo
anomalous fragmentation necessitating refinement of the
accepted mechanism. Experimental and computational data
support a rationale in which proton abstraction from the
hydroxy group in the initial cation radical does not cause a
Grob-like fragmentation, but rather produces a neutral
radical species, the alkoxy radical, that undergoes fragmen-
tation in either direction, i.e., cleaving the C-C bond to
dithiane or to the tertiary alkyl group.
observation of fundamental interest, necessitating a
refinement of the accepted mechanism.
All our previous observations were in keeping with the
mechanism shown in Scheme 1a, until we tested photo-
induced fragmentation in the tert-butyl derivative 1,
which in addition to the expected pivalaldehyde 2 (“nor-
mal” cleavage) produced dithiane-2-carboxaldehyde 3 as
a major product in 1:6.2 ratio, Scheme 1b. Isolation of
considerable amounts of R-(tert-butyl)benzhydrol, Ph2C-
(OH)tBu, indirectly confirmed the formation of tert-butyl
radical as the second product of the fragmentation.
Clearly the quasi-Grob electron pushing rationale, i.e.,
the benzophenone anion radical is pushing and the
heteroatom-centered cation radical is pulling, needed
refinement, because compound 1 was primarily cleaving
the wrong bond.
Oxidative electron transfer-induced fragmentations in
vicinal amino alcohols and diols have been known for a
long time and their mechanism has been exhaustively
studied.1 The accepted mechanistic rationale includes
photoinduced electron transfer to an ET-sensitizer, e.g.
benzophenone, followed by a mesolytic cleavage in the
generated cation radical, assisted by the benzophenone
anion radical deprotonating the vicinal hydroxy group.
It was noted earlier that the C-C bond cleavage step is
reminiscent of the Grob fragmentation in closed shell
systems.2
First we tested if the formation of 3 were a result of
the abstraction of the dithiane’s C-2 proton yielding the
enol form of 3:
Similar photoinduced fragmentation of 2-(R-hydroxy-
benzyl)-1,3-dithianes was recently reported by us, Scheme
1a, and its mechanism was investigated utilizing classical
physical organic methods such as the Hammett substitu-
ent effect and kinetic isotope effect studies3 and also a
laser flash photolysis study.4 In this note we report an
anomalous fragmentation in sulfur-centered cation radi-
cals containing vicinal hydroxy groupssan experimental
Deuteration of 1 changed the ratio of products 2 to 3
slightly, from 1:6.2 to 1:4.9, indicating that there is no
primary kinetic isotope effect on formation of 3. In fact,
completely blocking position 2 in the dithiane ring with
a methyl group did not shut off the anomalous fragmen-
tation, although the product ratio changed significantly,
with pivalaldehyde 2 becoming the major product (2:5 )
1.8:1).
(1) For review see: Gaillard, E. R.; Whitten, D. G. Acc. Chem. Res.
1996, 29, 292.
These experiments ruled out the competitive O-H vs
S2C-H deprotonation, leaving us with one possibility. It
appears that deprotonation of the hydroxy group in the
generated cation radical does not result in the formation
(2) (a) Ci, X.; Whitten, D. G. J . Am. Chem. Soc. 1987, 109, 7215. (b)
Ci, X.; Whitten, D. G. J . Am. Chem. Soc. 1989, 111, 3459.
(3) McHale, W. A.; Kutateladze A. G. J . Org. Chem. 1998, 63, 9924.
(4) Vath, P.; Falvey, D. E.; Barnhurst, L. A.; Kutateladze, A. G. J .
Org. Chem. 2001, 66, 2886.
10.1021/jo035001z CCC: $25.00 © 2003 American Chemical Society
Published on Web 09/26/2003
8236
J . Org. Chem. 2003, 68, 8236-8239

SCHEME 2
TABLE 1. Qu a n tu m Yield s of F r a gm en ta tion in
Aceton itr ile
of a charge separated “Grob-like” precursor -O-C-C-
S+•-R but rather produces a neutral oxygen-centered
radical (or a species behaving as one), via intramolecular
electron transfer. The O-radical undergoes subsequent
fragmentation in either direction, with the partition
correlating with the stability of the produced radicals,
Scheme 2.5 We also compared the results of the fragmen-
tation in methylene chloride and acetonitrile, which
constitutes almost an order of magnitude difference in
dielectric constant, and did not see any difference in the
partitioning within experimental error. The combined
experimental and computational results seem to indicate
that a charge-separated species does not exists or at best
is in fast equilibrium with the alkoxy radical.
The overall mode of fragmentation is somewhat similar
to cleavage in 4-methoxybenzyl alcohol radical cations
reported by Baciocchi and Steenken.6 At the same time,
as was shown by Asmus and co-workers in numerous
studies,7 intermolecular reaction of oxy-radicals (e.g.,
hydroxy, generated by pulse radiolysis) with sulfides
eventually leads to the formation of sulfur-centered
cation radicals, not vice versa. What is unusual in our
case is that the intramolecular charge transfer in R-hy-
droxyalkyl dithianes seems to occur in the opposite
direction, with the net result of the S-centered cation-
radical oxidizing the alkoxide.
Additional examples presented in Table 1 show that
cleaving the bond to the methyldithiane moiety is by far
the most efficient mode of fragmentation in this series.
Comparison of fragmentation ratios in 1 and 4 and also
of an unbiased competition between methyl dithianyl and
dithianyl departure in bis-adduct 9 shows that methyl
stabilization increases Φb by a factor of 11 to 14. If the
rate-determining step is the fragmentation in oxy-radical
1•, then the “normal” to “abnormal” ratio should simply
correlate with the relative stabilities of the radicals
produced as a result of such fragmentation.8 The change
in fragmentation ratios for 1 and 4 (from 1:6.2 to 1.8:1)
a
Partial quantum yields of cleavage “a” (to the left) and “b” (to
the right), respectively
(5) â-Alkylthio-ethoxy radicals were implicated in the dissociation
of hydroxyl radical adducts of (alkylthio)ethanol derivatives: Scho¨ne-
ich, C.; Bobrowski, K. J . Am. Chem. Soc. 1993, 115, 6538.
(6) (a) Baciocchi, E.; Bietti, M.; Steenken, S. J . Am. Chem. Soc. 1997,
119, 4078. (b) Baciocchi, E.; Bietti, M.; Steenken, S. Chem. Eur. J . 1999,
5, 1785.
corresponds to an approximately 1.4 kcal/mol relative
change in free energy of activation. We used Pasto’s9
isodesmic reaction, CH3-nXn + CH4 f XnCH4-n + CH3,
as the definition to compute the additional resonance
stabilization energy for 2-methyldithianyl radical at the
•
•
(7) Bonifacic, M.; Schaefer, K.; Moeckel, H.; Asmus, K. D. J . Phys.
Chem. 1975, 79, 1496. Asmus, K.-D. Acc. Chem. Res. 1979, 12, 436.
J . Org. Chem, Vol. 68, No. 21, 2003 8237

B3LYP/6-31g* level and found it to be 0.8 kcal/mol, which
indeed is in a reasonably good agreement with experi-
ment.
Formation of conjugated carbonyl improves the ef-
ficiency of photocleavage by a factor of 2 (cf. 6 and 9).
These results point to the late transition state of the C-C
bond cleavage.
Strikingly, separation of the hydroxy group from the
dithiane moiety with an extra methylene had virtually
no effect on Φa (cf. 1 and 7), although the through-bond
communication between the sulfur-centered cation radi-
cal and the oxygen atom was expected to diminish.
Conceivably this deficiency is offset by a direct interac-
tion, i.e., via a two-center three-electron bond. While
stable four-membered structures of this type have never
been observed,10 five-membered oxathiolane intermedi-
ates have previously been postulated.5
F IGURE 1. AM1 geometries of 11+• with intra- (left) and
inter-dithiane three-electron bonds.
cyclic dithia compounds reported by Asmus and co-
workers.12 They showed that the σfσ* band correspond-
ing to the intramolecular three-electron S-S bond in the
1,3-dithiane cation radical has λmax of 600 nm, or 47.7
kcal/mol, whereas the cation radical of 1,5-dithia-cyclooc-
tane, capable of forming a more stable 5-membered cycle
with the three-electron S-S bond, showed a considerable
blue shift of this band to 400 nm (or 71.5 kcal/mol), with
the difference of 23.8 kcal/mol.
In conclusion, dithianyl cation radicals having R-hy-
droxy-neopentyl or similar groups in position 2 undergo
an unusual fragmentation that necessitates refinement
of the previously proposed mechanism. Our experimental
and computational data support a rationale in which
deprotonation of the initial cation radical produces a
neutral radical species, most probably the alkoxy radical
that undergoes fragmentation in either direction. Ab
initio computations indicate that the deprotonation is
coupled with intramolecular electron transfer, although
experimentally we cannot rule out the possibility that it
is a two-step process, i.e., deprotonation followed by the
intramolecular electron transfer in the initially formed
charge separated species. In such a case both channels
of fragmentation, the Grob-like fragmentation in the
radical zwitterionic intermediate and the fragmentation
in the O-centered neutral radical, may be operational and
competing.
The rationale shown above could be an attractive
alternative to the mechanism presented in Scheme 2.
However, we did not find any computational evidence for
S-O three-electron bond formation in cation radicals of
1,2-thioalcohols 1, 4, 6, and 8-11 that can only form four-
membered rings, which, as mentioned above, is in keep-
ing with experimental observations.
Bis-dithianes 11 and 12 showed a significant increase
in quantum yields of fragmentation in part due to a
statistical factor of 2: both directions of fragmentation
lead to the formation of the table 2-methyl-1,3-dithian-
2-yl radical.
In addition to the statistical factor, upon close exami-
nation one can see that the efficiency Φa of the meth-
yldithiane cleaving off in both bis-dithianyl systems 10
and 11 (0.224 and 0.225, respectively) is higher than that
for 4 (0.167) or 6 (0.122). We suggest that this is due to
additional stabilization of the respective cation radicals
10+• and 11+•, which makes electron transfer less revers-
ible. AM1 computations show that two kinds of S-S
three-electron bonds could exist in 11+•, i.e., the intra-
dithiane (internal two-center three-electron bond forming
a three-membered ring, reported earlier11) and the inter-
dithiane S-S bond (i.e., the five-membered ring), Figure
1. The latter is calculated to be 21.8 kcal/mol more stable,
which is in remarkably good agreement with the reported
effects of ring size on the stability of cation radicals from
These results may have implications for the general
mechanism of fragmentation in heteroatom-centered
cation radicals containing vicinal hydroxy groups, al-
though at this point we are not ready to generalize
whether the same argument is applicable to mesolytic
cleavage in cation radicals of vicinal amino alcohols.
Exp er im en ta l Section
Gen er a l P r oced u r e for th e P r ep a r a tion of Dith ia n e
Ad d u cts w ith Ald eh yd es or Ep oxid es. To 1,3-dithiane or
2-methyl-1,3-dithiane (5 mmol), dissolved in 20 mL of THF and
stirred at -25 °C under a nitrogen atmosphere, was added 1.1
molar equiv of n-butyllithium (1.6 M in hexane). The mixture
was stirred at -25 °C for 3 h, the temperature was then reduced
to -78 °C, and the corresponding aldehyde or epoxide (5 mmol)
in 5 mL of THF was added. The mixture was stirred at -78 °C
for an additional 3 h and stored at -25 °C overnight. Saturated
ammonium chloride was added, and the aqueous phase was
extracted twice with dichloromethane. The combined organic
solution was washed twice with water and dried over Na2SO4.
The solvent was removed under vacuum and the residue was
purified by column chromatography.
(8) For a recent paper correlating radical stabilities in fragmenta-
tions of alkoxy radicals see: Nakamura, T.; Watanabe, Y.; Suyama,
S.; Tezuka, H. J . Chem. Soc., Perkin Trans. 2 2002, 1364 and references
therein.
(9) Pasto, D. J .; Krasnansky, R.; Zercher, C. J . Org. Chem. 1987,
52, 3062.
(10) (a) Hiller, K.-O.; Asmus, K.-D. Int. J . Radiat. Biol. 1981, 40,
597. (b) Asmus, K.-D.; Go¨bl, M.; Hiller, K.-O.; Mahling, S.; Mo¨nig, J .
J . Chem. Soc., Perkin Trans. 2 1985, 641.
Deuterated 1-d was obtained according to this general
procedure, starting with 2,2-dideuterio-1,3-dithiane, which in
(11) See: Maity, D. K. J . Am. Chem. Soc. 2002, 124, 8321 and
references therein.
(12) Asmus, K.-D.; Bahnemann, D.; Fischer, C.-H.; Veltwisch, D. J .
Am. Chem. Soc. 1979, 101, 5322.
8238 J . Org. Chem., Vol. 68, No. 21, 2003

turn was obtained by three consecutive lithiations of 1,3-dithiane
followed by quenching with D2O.
53.6, 47.9, 30.8, 29.7, 26.4, 26.0, 25.7, 24.0, 22.9. Anal. Calcd
for C10H20OS2: C, 42.51; H, 6.42. Found: C, 42.65; H 6.50.
Compounds 1,13 4,14 6,15 8 and 12,16 and 917 were described
previously.
Bis(2-m eth yl-1,3-d ith ia n -2-yl)-m eth a n ol (11). 11 was pre-
pared from lithiated 2-methyl-1,3-dithiane and 2-formyl-2-
methyl-1,3-dithiane.16 1H NMR (400 MHz, CDCl3) δ 4.30 (d, J
) 1.5 Hz, 1H), 3.60 (d, J ) 1.5 Hz, 1H), 3.17-3.01 (m, 4H), 2.85-
2.73 (m, 4H), 2.09-2.01 (m, 2H), 1.98-1.89 (m, 2H), 1.86 (s, 6H);
13C NMR (100 MHz, CDCl3) δ 77.6, 54.4, 27.6, 26.8, 25.4, 24.1.
Anal. Calcd for C10H20OS2: C, 44.55; H, 6.80. Found: C, 44.68;
H 6.80.
1-(1,3-Dith ia n -2-yl)-3,3-d im eth ylbu ta n -2-ol (7) (99%): 1H
NMR (CDCl3) δ 4.27 (dd, J 1 ) 10.3 Hz, J 2 ) 3.7 Hz, 1H), 3.57
(d, J ) 10.3 Hz, 1H), 2.98-2.80 (m, 4H), 2.17-1.70 (m, 5H), 0.9
(s, 9H); 13C NMR NMR (100 MHz, CDCl3) δ 75.9, 45.1, 37.4, 34.8,
30.6, 30.1, 26.2, 25.7. Anal. Calcd for C10H20OS2: C, 54.50; H,
9.15. Found: C, 54.64; H 9.18.
(1,3-Dit h ia n -2-yl)-(2-m et h yl-1,3-d it h ia n -2-yl)-m et h a n ol
(10). 10 was prepared from lithiated 1,3-dithiane and 2-formyl-
2-methyl-1,3-dithiane.16 1H NMR (400 MHz, CD3CN) δ 4.79 (d,
J ) 1.5 Hz, 1H), 4.10 (dd, J 1 ) 1.5 Hz, J 2 ) 4.4 Hz, 1H), 3.44 (d,
J ) 4.4 Hz, 1H), 3.07-2.56 (m, 8H), 2.09-1.98 (m, 2H), 1.83-
1.67 (m, 2H), 1.47 (s, 3H); 13C NMR (100 MHz, CDCl3) δ 75.4,
Ack n ow led gm en t. Support of this research by the
National Science Foundation (CHE-314344) is gratefully
acknowledged.
Su p p or tin g In for m a tion Ava ila ble: NMR spectra of
compounds 1, 1-D, 4, and 6-11 and computed geometries and
energies of 11+• and of dithiane-dithianyl radical and meth-
yldithiane-methyldithianyl radical pairs for the isodesmic
reaction. This material is available free of charge via the
Internet at http://pubs.acs.org.
(13) Kang, J .; Kim, J . I.; Lee, J . H. Bull. Korean Chem. Soc. 1994,
15, 865.
(14) Katzenellenbogen, J . A.; Bowlus, S. P. J . Org. Chem. 1973, 38,
627.
(15) Tietze, L. F.; Weigand, B.; Wuff, C. Synthesis 2000, 69.
(16) Seebach, D.; Corey, E. J . J . Org. Chem. 1975, 40, 231.
(17) Wan, Y.; Kurchan, A.; Kutateladze, A. J . Org. Chem. 2001, 66,
1894.
J O035001Z
J . Org. Chem, Vol. 68, No. 21, 2003 8239