Radical Ion Probes. 7. Behavior of a "Hypersensitive" Probe for Single Electron Transfer in Reactions Not Involving Electron Transfer

Article abstract of DOI:10.1021/jo970584w

1-Methyl-5,7-di-tert-butylspiro[2.5]octa-4,7-dien-6-one (8) and 1,1-dimethyl-5,7-di-tert-butylspiro-[2.5]octa-4,7-dien-6-one (1) react with thiophenoxide ion to produce cyclopropane ring-opened products. Thermodynamic considerations effectively rule out any possibility that single electron transfer is involved in theses reactions; the process PhS^- + substrate → PhS^. + substrate^.- is endothermic by over 50 kcal/moll Nucleophilic attack occurs both at the least- and most-hindered carbons of the cyclopropyl group, and the product ratio (R(1°/2°) from 8 and A(1°/3°) from 1, where 1°, 2°, and 3° refer to the regioisomeric phenyl sulfides formed from these substrates) is found to vary with solvent. In dipolar, aprotic solvents, nucleophilic attack occurs preferentially at the least-hindered carbon of the cyclopropyl group (A(1°/2°) and A(1°/3°) ≈ 4-5), consistent with an S_N2 mechanism. In protic solvents, products arising from nucleophilic attack at the more-substituted carbon of the cyclopropyl group become increasingly important, consistent with the onset of a carbocationic (S_N2(C+)) pathway. The strengths and weaknesses of 1 and 8 as probes for single electron transfer are discussed in the context of these results.

Full text of DOI:10.1021/jo970584w

5550
J . Org. Chem. 1997, 62, 5550-5556
Ra d ica l Ion P r obes. 7. Beh a vior of a “Hyp er sen sitive” P r obe for
Sin gle Electr on Tr a n sfer in Rea ction s Not In volvin g Electr on
Tr a n sfer
J ames M. Tanko* and Larry E. Brammer, J r.
Department of Chemistry, Virginia Polytechnic Institute and State University,
Blacksburg, Virginia 24061-0212
Received April 1, 1997X
1-Methyl-5,7-di-tert-butylspiro[2.5]octa-4,7-dien-6-one (8) and 1,1-dimethyl-5,7-di-tert-butylspiro-
[2.5]octa-4,7-dien-6-one (1) react with thiophenoxide ion to produce cyclopropane ring-opened
products. Thermodynamic considerations effectively rule out any possibility that single electron
transfer is involved in theses reactions; the process PhS- + substrate f PhS• + substrate•- is
endothermic by over 50 kcal/mol! Nucleophilic attack occurs both at the least- and most-hindered
carbons of the cyclopropyl group, and the product ratio (R(1°/2°) from 8 and R(1°/3°) from 1, where
1°, 2°, and 3° refer to the regioisomeric phenyl sulfides formed from these substrates) is found to
vary with solvent. In dipolar, aprotic solvents, nucleophilic attack occurs preferentially at the least-
hindered carbon of the cyclopropyl group (R(1°/2°) and R(1°/3°) ≈ 4-5), consistent with an SN2
mechanism. In protic solvents, products arising from nucleophilic attack at the more-substituted
carbon of the cyclopropyl group become increasingly important, consistent with the onset of a
carbocationic (SN2(C+)) pathway. The strengths and weaknesses of 1 and 8 as probes for single
electron transfer are discussed in the context of these results.
I. In tr od u ction
reaction. The key elements of this hypothesis are
highlighted in Scheme 1. If 1 reacts with a nucleophile
via SET, rearrangement of the resulting radical anion
(2 f 3) will lead to a product (4) in which the nucleophile
is attached to the most-hindered carbon. In contrast, for
the polar pathway nucleophilic attack is expected to occur
preferentially at the least-hindered carbon, leading to 5.1
Single electron transfer (SET) has emerged as an
important mechanistic pathway for a number of organic
transformations. Many processes previously believed to
proceed solely via polar (two electron) pathways are now
believed to involve some component of SET. Most
experimental approaches for diagnosing whether electron
transfer may be involved in a particular organic trans-
formation exploit the unique properties of the paramag-
netic intermediates (i.e., free radicals and radical ions)
that are formed.
By examining reactions of 1 with nucleophiles which
have been shown independently to react with carbonyl
compounds via SET, we were able to confirm the first
half of Scheme 1, namely, the behavior of this system in
bona fide SET processes.2 However, in order to be
generally useful as mechanistic probe for SET, it is
equally important to understand how this compound
behaves in reactions where SET is not occurring.
One very popular approach is to incorporate structural
features into the reactants which serve as intramolecular
traps for paramagnetic intermediates. Assuming that
the rearrangement is a unique feature of the radical or
radical ion, the detection of structurally rearranged
products may infer that an electron transfer mechanism
was involved. The problem with this approach is that
rearranged products may also result from mechanisms
other than SET. When such is the case, the differentia-
tion between SET and polar pathways becomes a very
knotty problem.
II. Resu lts a n d Discu ssion
A. In itia l Obser va tion s. Reaction of 1 (0.02 M) with
thiophenoxide ion (PhS- K+) in DMSO (in the presence
of 18-crown-6 to minimize ion-pair formation) led to
formation of ring opened products, 1° and 3° sulfides 6
and 7 (eq 1). This reaction occurred in e5 min and in
nearly quantitative yield. The product ratio 6/7 (R(1°/
3°)) was unaffected by changes in the concentration of
PhS-: At both 0.02 and 0.2 M PhS-, R(1°/3°) ) 3.6.
In 1994, we suggested that spiro[2.5]octa-3,6-dien-5-
one 1 may be an excellent probe for the detection of SET
in reactions of nucleophiles with carbonyl compounds.1
The estimated reduction potential of 1 (-2.2 V vs SCE)
is similar to that of aromatic ketones and enones.
Moreover, the radical anion resulting from one-electron
reduction of 1 undergoes ring opening (predominantly to
3° distonic radical ion 3) with a rate constant >107 s-1
This system is especially intriguing because it may
provide a means for differentiating between polar vs SET
pathways based upon the observed regiochemistry of the
.
(1)
X Abstract published in Advance ACS Abstracts, J uly 15, 1997.
(1) Tanko, J . M.; Brammer, L. E., J r.; Hervas’, M.; Campos, K. J .
Chem. Soc., Perkin Trans. 2 1994, 1407.
(2) Tanko, J . M.; Brammer, L. E., J r. J . Chem. Soc., Chem. Commun.
1994, 1165.
S0022-3263(97)00584-7 CCC: $14.00 © 1997 American Chemical Society

Radical Ion Probes
J . Org. Chem., Vol. 62, No. 16, 1997 5551
Sch em e 1
Sch em e 2
Sch em e 3
are the result of direct electron transfer as outlined in
Scheme 1. The redox potential of the PhS•/PhS- couple
is reported to be 0.1 V (vs SCE).3 The reduction potential
of 1 is estimated to be -2.2 V vs SCE (also in DMSO).
Thus, direct electron transfer between PhS- and 1 (or 8)
is endothermic by over 50 kcal/mol!
We considered the possibility that 7 and 10 might be
the result of a chain process involving radical anions, i.e.,
the SRN1 reaction4 (Scheme 3). Presumably this process
could be initiated either fortuitously via action of labora-
tory light or, perhaps, via spontaneous electron transfer
between PhS- and substrate. (The fact that electron
transfer from PhS- to 1 is extremely sluggish might be
overcome if the chain length were sufficiently long.)
Similarly, reaction of monomethyl derivative 8 (under
the same conditions) led to nearly the same ratio of 1°
and 2° sulfides (9 and 10, eq 2): R(1°/2°) ) 3.8.
(2)
However, the SRN1 mechanism was ruled out for the
reaction of 1 and PhS- on the following grounds: The
reaction is not photoinitiated and occurs readily in the
dark resulting in the same ratio of products (R(1°/3°) )
3.2) in the same yield and in the same period of time.
Furthermore, the reaction is not inhibited by O2, PhSH,
or PhSSPh, and the product ratios were nearly identical
(R(1°/3°) ) 3.2, 3.6, and 3.7, respectively) in the presence
of these materials. (O2,5 PhSH,6 and PhSSPh7 react with
B. Com p etin g SN2/SN1 P a th w a ys. It is likely that,
for the reaction of both 1 and 8 with PhS-, the primary
sulfides are the result of direct nucleophilic attack at the
least-hindered carbon of the cyclopropyl group (i.e., an
SN2 process). However, the origin of the products 7 and
10, where PhS- appeared to react at the more-hindered
carbon was less obvious. Competing SN1/SN2 pathways
(Scheme 2) appeared to be ruled out by two observa-
tions: (a) R(1°/3°) and R(1°/2°) were nearly identical
(presumably R(1°/3°) would be less than R(1°/2°) because
1 can form a 3° carbocation, vs a 2° carbocation for 8),
and (b) the product ratio is unaffected by nucleophile
concentration (higher nucleophile concentrations were
expected to favor the SN2 product).
(3) Andrieux, C. P.; Hapiot, P.; Pinson, J .; Save´ant, J .-M. J . Am.
Chem. Soc. 1993, 115, 7783.
(4) For examples of ArS- participating in SRN1 reactions involving
aliphatic radicals, see: (a) Meijs, G. F. J . Org. Chem. 1986, 51, 606.
(b) Ahbala, M.; Hapiot, P.; Houmam, A.; J ouini, M.; Pinson, J .; Save´ant,
J .-M. J . Am. Chem. Soc. 1995, 117, 11488. For examples of ArS-
participation in the aromatic SRN1 reaction, see: Bunnett, J . F.; Creary,
X. J . Org. Chem. 1974, 39, 3174.
(5) Rate constants for reaction of alkyl radicals with O2 in solution
are greater than 109 M-1 s-1, see: Maillard, B.; Ingold, K. U.; Scaiano,
J . C. J . Am. Chem. Soc. 1983, 105, 5095.
(6) Rate constants for R• + PhSH f RH + PhS• are ca. 108 M-1 s-1
for R ) 1°, 2°, or 3° alkyl. See: Franz, J . A.; Bushaw, B. A.; Alnajjar,
M. S. J . Am. Chem. Soc. 1989, 111, 268.
C. Com p et in g SN2/SE T P a t h w a ys? It is very
unlikely that 3°-sulfide (from 1) and 2°-sulfide (from 8)

5552 J . Org. Chem., Vol. 62, No. 16, 1997
Tanko and Brammer
Ta ble 1. P r od u ct Ra tios a n d Yield s for th e Rea ction of 1
a n d 8 w ith P h S- in Sever a l Solven ts
Sch em e 4
reaction of 1
reaction of 8
solvent
R(1°/3°)a
yield (%)
R(1°/2°)b
yield (%)
THF
5.34
3.91
4.68
3.93
4.75
3.05
1.35
0.96
0.65
0.43
84
84
96
80
86
80
91
84
72
56
4.81
4.77
4.17
4.37
4.70
3.75
2.47
2.30
1.75
1.46
63
87
97
86
87
80
89
90
97
95
EtOAc
acetone
pyridine
DMF
DMSO
t-BuOH
i-PrOH
EtOH
alkyl radicals efficiently and are expected to trap 3°
distonic radical ion 3, thereby inhibiting the reaction and
diminishing the yield of 3° sulfide 7 and resulting in a
larger observed R(1°/3°).) The observations that these
reagents neither perturbed the product ratio nor inhib-
ited the reaction is a good indication that competing SN2/
MeOH
S
RN1 process is not involved.
D. Reeva lu a tion of Resu lts: Th e SN2(C+) Mech -
a
b
Ratio of 1° sulfide 6 to 3° sulfide 7. Ratio of 1° sulfide 9 to 2°
sulfide 10.
a n ism . The general mechanism for nucleophilic substi-
tution via carbocationic pathways is depicted in Scheme
4. Using the steady-state approximation, the overall rate
law for this reaction is:
k1k2[RX][Nu-]
k-1[X-] + k2[Nu:-]
d[RX]
dt
-
)
Typically, k-1[X-] is small relative to k2[Nu:-], and the
rate law reduces to k1[RX], i.e., the classic SN1 reactions.
However, when the opposite situation prevails (k-1[X-]
> k2[Nu:-]), the rate law reduces to an overall second-
order rate law: K1k2[RX][Nu:-]. This latter situation is
quite rare and is referred to as the SN2(C+) process (“2”
because the reaction is second-order overall, “C+” be-
cause a carbocation is involved).8-10
All of our observations for the reactions of 1and 8 with
PhS- are consistent with competing SN2/SN2(C+) path-
ways for product formation: (a) Both processes are first
order in nucleophile, thus explaining why R(1°/2°) and
R(1°/3°) are unaffected by changes in PhS- concentration.
(b) The carbocation generated from 1 (1+) is expected to
be more stable than that generated from 8 (8+). How-
ever, for the same reason, 1+ is also expected less reactive
toward nucleophiles than 8+. Thus the more favorable
equilibrium constant for carbocation formation from 1
(K1) may be offset by a smaller rate constant for reaction
with PhS- (k2), thereby explaining why R(1°/2°) and R(1°/
3°) are so similar.
Despite its rarity, it is not unreasonable to suspect that
the SN2(C+) mechanism might be operating in these
systems. For “typical” substrates, external return (re-
combination of R+ and X- to regenerate R-X) is a
bimolecular process; k-1[X-] is small relative to k2[Nu:-]
because the concentration of X- in solution is small.11 For
our system however, k-1 (Scheme 2) is a unimolecular
process...the leaving group is still attached to the car-
bocation. Thus it is not at all unreasonable to expect that
intramolecular ring closure would be faster than inter-
molecular nucleophilic attack (k-1 > k2[PhS-], Scheme
2).
F igu r e 1. Variation in the log of the product ratio (6/7) for
reaction of 1 with PhS- as a function of the ionizing power of
the solvent.
E. Evid en ce for Ca r boca tion s a s In ter m ed ia tes
in th e Rea ction of 1 a n d 8 w ith P h S-. To fully explore
whether carbocations were involved in the reactions of 1
and 8 with PhS-, these reactions were examined in
several solvents. In all instances, PhSK was utilized in
the presence of 18-crown-6 so as to complex K+ and
minimize ion-pairing effects. Mass balances in these
experiments were typically >80%.12 The results of these
experiments are summarized in Table 1.
In Figures 1 and 2, log(R(1°/3°)) and log(R(1°/2°)) are
plotted as a function of the ionizing power of the solvent,
expressed in terms of log(kneo), where kneo is the rate
constant for ionization of neophyl tosylate (12) in the
same solvent (eq 3).13
(3)
There are several good reasons for selecting log(kneo
)
as the solvent parameter to correlate these results: (a)
Ionization of 12 proceeds with anchimeric assistance
yielding spiro cation 13. As a result, these solvent
parameters are a good measure of a solvent’s ionizing
(7) Undecyl radical reacts with PhSSPh with a rate constant of 2 ×
105 M-1 s-1, see: Curran, D. P.; Martin-Esker, A. A.; Ko, S.-B.;
Newcomb, M. J . Org. Chem. 1993, 58, 4691.
(8) Gelles, E.; Hughes, E. D.; Ingold, C. K. J . Chem. Soc. 1954, 2918.
(9) Hartshorn, S. R. Aliphatic Nucleophilic Substitution, Cambridge
University Press: London, 1973; pp 12-13.
(10) Kinoshita, T.; Ueda, H.; Takeuchi, K. J . Chem. Soc., Perkin
Trans. 2 1993, 603.
(12) Compound 1 has a tendency to polymerize in MeOH and EtOH,
and as a consequence, the mass balances were only 56 and 72%,
respectively, in these two solvents.
(13) Smith, S. G.; Fainberg, A. H.; Winstein, S. J . Am. Chem. Soc.
1961, 83, 618.
(11) For a classical SN1 process, the addition of X- to the solution
results in a depression in reaction rate (i.e., the common-ion effect).
See ref 9.

Radical Ion Probes
J . Org. Chem., Vol. 62, No. 16, 1997 5553
Ta ble 2. Effect of P h S- Con cen tr a tion on th e P r od u ct
Ra tios Obser ved for Rea ction s of 1 a n d 8 in
Rep r esen ta tive Solven ts
reaction of 1
reaction of 8
solvent
acetone
[PhS-], Ma
R(1°/3°)b
[PhS-], Mc
R(1°/2°)d
0.002
0.02
5.1
4.9
0.001
0.005
0.01
0.001
0.005
0.01
0.001
0.005
0.01
4.1
4.2
4.4
3.9
4.1
4.1
2.1
3.1
3.1
DMSO
0.02
0.20
3.6
3.6
i-PrOH
EtOH
0.02
0.20
0.60
0.62
a
b
0.02 M 1. Ratio of 1° sulfide 6 to 3° sulfide 7. c 0.10 M 8.
F igu r e 2. Variation in the log of the product ratio (9/10) for
reaction of 8 with PhS- as a function of the ionizing power of
the solvent.
d
Ratio of 1° sulfide 9 to 2° sulfide 10.
Sch em e 5
power and are not “contaminated” by the solvent’s
nucloephilicity (as is the case with parameters such as
the Y scale). (b) log(kneo) values are available for a
broader range of solvents (ranging from dipolar, aprotic
to protic) than is the case for Y values,14 which are limited
mostly to protic solvents.15
As Figures 1 and 2 reveal, there are two regions in the
plots of log(R) vs log(kneo). In region I (which is consti-
tuted by the dipolar aprotic solvents) the product ratio
is independent of the ionizing power of the solvent. Using
the results for THF, EtOAc, acetone, pyridine, and DMF,
the slope of the plot of log(R) vs log(kneo) is -0.011 and
-0.016 for 1 and 8, respectively. In region II (constitut-
ing the more polar aprotic solvents and protic sovents),
the product ratio varies linearly with the ionizing power
of the solvent. Using the results for DMSO, t-BuOH,
Ta ble 3. Ra tio of P r od u cts F or m ed in Com p etition
Exp er im en ts P ittin g 8 vs CH3CH2Br for P h S- in Sever a l
Solven ts
product ratio product ratio
solvent
[PhS-], Ma
14:10
14:9
R(1°/2°)
acetone
i-PrOH
DMSO
0.001
0.005
0.01
0.001
0.005
0.01
0.001
0.005
0.01
7.4
7.8
33
41
52
0.94
0.75
0.88
28
27
36
4.5
5.2
5.2
2.1
2.1
2.1
3.95
4.1
4.1
i-PrOH, EtOH, and MeOH, the slope of log(R) vs log(kneo
)
10.0
0.49
0.35
0.42
7.1
is -1.07 and -0.401 for 1 and 8, respectively.
Because of the observed curvature in the log(R) vs
log(kneo) plots for 1 and 8 (Figures 1 and 2), it seemed
prudent to verify that the effect of PhS- concentration
on the product ratios was the same in both regions I and
II. (The preliminary results discussed above were con-
fined to DMSO solvent, which appears to be the break
point in these plots.) Toward this end, the effect of PhS-
concentration on R was studied for both the reaction of
1 and 8 in representative solvents in both regions I and
II (Table 2).
The results in Table 2 reveal that R(1°/2°) and R(1°/
3°) are unaffected by [PhS-] in both regions I and II,
consistent with the fact that the competing processes
resulting in apparent nucleophilic attack at the least- and
most-hindered postions of 1 and 8 were the same reaction
order in PhS- (presumably first order).
6.7
8.0
a
For all experiments, [14] ) [CH3CH2Br] ) 0.10 M.
(Scheme 5). The results of these experiments are sum-
marized in Table 3. (These experiments also provide
values of R(1°/2°), which are consistent with the data in
Tables 1 and 2.)
The product ratios 9:14 and 10:14 remained constant
over an order of magnitude variation in PhS- concentra-
tion. Making the reasonable assumption that the reac-
tion CH3CH2Br + PhS- f CH3CH2SPh + Br- is an SN2
process in all solvents and, thus, first order in PhS-, the
results in Table 3 confirm that the processes leading to
9 and 10 are both first order in PhS- in these solvents.
By inference, the same presumably holds true for the
formation of 6 and 7 from 1.
G. Effect of Solven t/Cou n ter ion on th e Str u ctu r e
of Dien on e 8. The reactivity of 1 and 8 toward nucleo-
philes, either via direct displacement or carbocationic
pathways, is attributable to direct conjugation between
the cyclopropyl and carbonyl (Scheme 6). In these
compounds, the cyclopropyl group is effectively locked in
the bisected conformation, where this interaction is
maximal.
F . Deter m in a tion of th e Rea ction Or d er in P h S-.
Determination of the precise reaction order in PhS- in
reactions with 8 was accomplished by competition ex-
perments.16 In these experiments, 8 was allowed to
compete with CH3CH2Br for PhS- in several solvents
(14) Grunwald, E.; Winstein, S. J . Am. Chem. Soc. 1948, 70, 846.
For a review, see: Bentley, T. W.; Llewellyn, G. In Progress in Physical
Organic Chemistry, Vol. 17; Taft, R. W., Ed.; Wiley: New York, 1990;
pp 121-158.
(15) log(kneo) values were not available for 2-propanol and tert-butyl
alcohol. These values were extrapolated from the slope of a graph of
Y values vs log(kneo) values.
(16) Attempts to directly study the kinetics of reaction between PhS-
and either 1 or 2 by UV proved troublesome because the reactions were
over in less than 5 min, and the results were unacceptably erratic.
Bond lengths and atomic charges for these spirohexa-
dienones obtained by AM1 SCF-MO calculations17 are

5554 J . Org. Chem., Vol. 62, No. 16, 1997
Tanko and Brammer
Sch em e 6
in polar/protic solvents, both carbons are affected to nearly
the same extent. Moreover, when K+ or K+/18-crown-6
are present, the chemical shifts are virtually unaffected.
Thus it appears unlikely that there is any bond-
lengthening or partial ionization of the compound (i.e.,
17) in these solvents or in the presence of the positively
charged counterion.18
Ta ble 4. AM1-Ca lcu la ted Bon d Len gth s a n d Atom ic
Ch a r ges for Sp ir o[2.5]octa -4,7-d ien -6-on es
III. Con clu sion s
A. Dip ola r , Ap r otic Solven ts (Region I). In dipo-
lar, aprotic solvents spiro[2.5]octa-3,6-dien-5-ones 1 and
8 react with PhS- to produce ring-opened products via
an SN2 process. As expected, nucleophilic attack occurs
preferentially at the least-hindered carbon (by a factor
of approximately 4-5 for both 1 and 8). The only
“surprise” in these results is that there is any attack at
the more-hindered carbon at all. To the extent that the
numbers are believable (i.e., uncontaminated by the SN1
reaction) 3° carbons are generally (at least) 1000x less
reactive 1° in a classic SN2 reaction.9
However, this result is not unprecedented. Boger has
reported that 18 reacts via an SN2 mechanism with
CH3OH (Scheme 7).19 Nucleophilic attack occurs pref-
erentially at the primary carbon relative to tertiary of
the cyclopropyl group by a factor of 1.7:1. Compound 18
is a member of a class of compounds (the duocarmycins)
which are potent antitumor antibiotics and which possess
the identical structural feature (spiro[2.5]octa-4,7-dien-
6-one moiety) as 1 and 8. An X-ray crystal structure of
18 reveals that (like spiro[2.5]octa-3,6-dien-5-ones 1 and
8), the cyclopropyl group is effectively locked in a near
perfect bisected conformation.19 Thus both cyclopropyl
bonds are activated toward nucleophilic attack.20
Ta ble 5. Effect of Solven t a n d Cou n ter ion on 13C NMR
Sh ifts for th e Cyclop r op yl CH2 a n d CH of 8
δCH
δCH
2
solvent/counterion (ppm vs TMS) (ppm vs TMS) ∆ (ppm)a
benzene-d6
CDCl3
acetone-d6
DMSO-d6
DMSO-d6 + KClO4
DMSO-d6 + KClO4
(18-crown-6)
methanol-d4
27.6
28.1
28.9
28.3
28.3
28.4
27.6
27.9
28.1
27.8
27.8
27.8
0.0
-0.2
-0.8
-0.5
-0.5
-0.6
30.0
28.9
-1.1
a
∆ ) δCH - δCH
.
2
summarized in Table 4. With increased methyl substitu-
tion at C1, the C1-C3 bond length becomes slightly
longer than C2-C3, and the electron density becomes
diminished at C1 compared to C2. These observations
are consistent the resonance structures depicted in
Scheme 6 and help explain why C1 and C2 are both
reactive toward nucleophiles.
It is possible that the relative contributions of reso-
nance structures 15 and 16 may be affected by counterion
and solvent. Moreover, it is also possible that if interac-
tion between the 1 or 8 with the positively charged
counter ion associated with PhS- (K+/18-crown-6) were
sufficiently strong, this might lead to partial ionization
of the substrate (manifested by lengthening of the C1-
C3 bond and a buildup of positive charge at C1 relative
to C2).
In terms of the utility of 1 or 8 as probes for single
electron transfer, these results are consistent with our
earlier hypothesis (Scheme 1)1 regarding how these
compounds would behave when reacting with a nucleo-
phile via a polar pathway (SN2) processsnucleophilic
attack occurs preferentially at the least-hindered carbon
yielding 5. However, a fair amount of product also arises
13C NMR chemical shifts should be a sensitive measure
of the extent that solvent and/or counterion affects the
polarization of the cyclopropyl group in these compounds.
Chemical shifts for the cyclopropyl CH and CH2 of 8 (and
the difference between them) were determined in several
solvents and in the presence of K+ and K+/18-crown-6.
The results are summarized in Table 5 and suggest that
as solvent polarity increases, the chemical shifts move
downfield (very slightly). These results are consistent
with notion that the cyclopropyl carbons bear slightly
more positive charge in more polar and protic solvents
(i.e., the contributions of resonance forms 15 and 16 is
slightly enhanced in these solvents).
(18) For both 1 and 8, the bulky tert-butyl groups at the ortho
position are expected to severely impede coordination of a Lewis acid
with the carbonyl oxygen. We suspect that most of the observed
changes in chemical shift observed for 8 are more a function of changes
in the dielectric constant of the medium, rather than any close
association of solvent or counterion with the oxygen.
(19) Boger, D. L.; J ohnson, D. S. Angew. Chem., Int. Ed. Engl. 1996,
33, 1438. See also: Boger, D. L. Acc. Chem. Res. 1995, 28, 20.
(20) An reviewer has suggested that this “SN2” reaction may be
occurring with significant electron-transfer character in the transition
state, e.g.,
However, the data also suggest that although the
cyclopropyl group is bearing slightly more positive charge
Increasing radical character at the cyclopropyl carbon favors attack
at the more hindered carbon, while an increase in covalent character
favors attack at the least hindered position. For an good discussion
of the SN2/ET mechanistic continuum, see: Pross, A. Theoretical and
Physical Principles of Organic Reactivity; Wiley: New York, 1995; pp
222-232.
(17) Dewar, M. J . S.; Zoebisch, E. G.; Healy, E. F.; Stewart, J . J . P.
J . Am. Chem. Soc. 1985, 107, 3902.

Radical Ion Probes
J . Org. Chem., Vol. 62, No. 16, 1997 5555
layer chromatography (TLC) was performed on precoated
Sch em e 7
polyester silica gel plates (Whatman) with
background.
a fluorescent
Ma ter ia ls. Acetone (J T Baker HPLC Grade), ethyl acetate
(Mallinckrodt), pyridine (Fisher), tert-butyl alcohol (Aldrich
99+%), 2-propanol (EM Scientific), ethanol (AAPER 100%),
and methanol (J T Baker HPLC Grade) were used as received.
THF (Mallinckrodt) was distilled from lithium aluminum
hydride before use. DMSO (Fisher) was distilled from CaH2
before use. DMF (EM Scientific) was stirred over anhydrous
copper(II) sulfate and activated neutral alumina under argon
for 3 days and then distilled prior to use. 1,1-Dimethyl-5,7-
di-tert-butylspiro[2.5]octa-4,7-dien-6-one (1) and 1-methyl-5,7-
di-tert-butylspiro[2.5]octa-4,7-dien-6-one (8) were prepared
according to published procedures.22
P ota ssiu m Th iop h en oxid e. To a solution of 50 mL of
100% ethanol and 1.12 g of potassium hydroxide (20.0 mmol)
was added with stirring 2.05 mL (20.0 mmol) of thiophenol
(Aldrich 97%). The solution was allowed to stir for an
additional 3 h after which the solvent was removed in vacuo.
The resulting white solid was stirred with 4 × 50 mL aliquots
of diethyl ether, cannulating the ether off each time. Removal
of the ether yielded pure potassium thiophenoxide.
Eth yl P h en yl Su lfid e (14). Iodoethane (5.6 g, 36.0 mmol)
(Aldrich 99%) was added to 10 mL of dry acetone. Potassium
thiophenoxide (5.36 g, 36.0 mmol) was added to the reaction
mixture with stirring. The reaction mixture was monitored
via HPLC. After disappearance of the starting material the
solvent was removed leaving a light yellow liquid. This liquid
was distilled (bp 50 °C, 2.0 mmHg consistent with lit.23 bp 204
°C) yielding 4 g (80%) of colorless liquid.
Gen er a l P r oced u r e for th e Rea ction of 1 a n d 8 w ith
P ota ssiu m Th iop h en oxid e. In a typical reaction 2.5 × 10-2
mmol of substrate was dissolved in 5 mL of dry solvent that
had been thoroughly purged with argon. The resulting solu-
tion was added to potassium thiophenoxide (0.0045 g, 3.0 ×
10-2 mmol) and 18-crown-6 (Aldrich, 99.5%) (0.0085 g, 3.0 ×
10-2 mmol) dissolved in 5 mL of the same solvent at room
temperature. The reaction mixture was allowed to stir for 1
h after which it was acidified with 1% H2SO4. The mixture
was subsequently extracted with 3× with ether. The combined
ether extracts were washed 3× with water, dried (MgSO4), and
evaporated. Yields were determined by HPLC and/or GC
analysis. (The concentrations of reagents used and the results
of specific experiments are summarized in the text and in
Tables 1, 2, and 3.)
from nucleophilic attack at the more-substituted carbon
of the cyclopropyl group, albeit to a lesser extent, leading
to 4. Thus, the detection of a small quantity of com-
pounds such as 4 does not mean that an electron transfer
mechanism is operative. Thus compounds such as 1 or
8 are excellent probes for SET pathways, when the
results are interpreted with caution.
B. P r otic Solven ts (Region II). In protic solvents,
spiro[2.5]octa-3,6-dien-5-ones 1 and 8 react via competing
SN2/SN2(C+) pathways. The contribution of the carboca-
tionic pathway becomes increasingly important as the
ionizing power of the solvent increases. To our knowl-
edge, these systems provide the first example of the
SN2(C+) mechanism under “normal” reaction conditions
(i.e., without highly hindered nucleophiles or highly
resonance-stabilized carbocations).
In protic solvents, the outlook for use of these com-
pounds as SET probes is grim. With regard to Scheme
1, because in protic solvents a competing carbocationic
pathway provides another means for forming 4, simple
product studies will lead to a false conclusion regarding
the importance of SET pathways.
Exp er im en ta l Section
Authentic samples of products were obtained by preparative
HPLC.
Gen er a l Meth od s. Melting points were determined using
a Thomas-Hoover capillary melting point apparatus and are
uncorrected. High-pressure liquid chromatography (prepara-
tive and analytical scale) was performed using a Beckman
System Gold Model 128 solvent pump system and Model 166
UV-vis detector interfaced to an MS-DOS computer. Samples
were analyzed and separated using Beckman C-19 reverse
phase columns (analytical, 4.6 mm × 250 mm; preparative,
21.2 mm × 150 mm) using acetonitrile/water solvent mixtures.
Gas chromatographic analyses were performed on a Hewlett-
Packard HP 5890 instrument equipped with FID detectors, a
HP 3393A integrator, and either an Alltech SE-54 capillary
column (30 m × 0.25 mm) or a Supelco SE-30 capillary column
(15 m × 0.25 mm). Ultraviolet spectra were acquired through
the use of a Hewlett-Packard HP 8452A Diode Array UV-vis
spectrophotometer. Infrared (IR) spectroscopy was performed
on a Perkin-Elmer 1600 Series FT-IR spectrophotometer.
Nuclear magnetic resonance (NMR) spectra were obtained in
CDCl3 using either a Bruker WP-200 or WP-270 spectrometer
and are reported in units vs TMS. Low-resolution GC-MS was
performed on a Hewlett-Packard HP 5890 gas chromatograph
utilizing an HP 1% methyl phenyl silicone gum column (12.5
m × 0.2 mm) interfaced to a HP 5970 mass spectrometer.
High- and low-resolution MS was performed on a VG-7070E
mass spectrometer, employing EI ionization at 70 EV. Flash
chromatography21 was performed on silica gel (Aldrich Grade
60, 630-400 mesh) using ethyl acetate/hexane mixtures. Thin
2,6-Bis(1,1-d im eth yleth yl)-4-(1,1-d im eth yl-2-(p h en ylth -
io)eth yl)p h en ol (6): mp 85.0-85.5 °C; 1H NMR (CDCl3) δ
1.42 (s, 18H), 1.44 (s, 6H), 3.20 (s, 2H), 5.07 (s, 1H), 7.19-
7.20 (m, 7H); 13C NMR (CDCl3) δ 28.25 (q), 30.34 (q), 34.54
(s), 39.03 (s), 49.83 (t), 162.46 (d), 125.40 (d), 128.59 (d), 129.36
(d), 135.20 (s) and 138.11 (s), 138.38 (s), 151.89 (s); UV-vis
(ethanol) λmax (log ꢀ) 206 nm (4.80), 258 (3.95); IR (neat) cm-1
3640, 3059, 3000, 2963, 2871, 1583, 1479, 1437, 1382, 1363,
1320, 1637, 1158, 1120, 1025, 877, 809, 690, 668; MS m/ e
(relative intensity) 370 (M+, 1), 339 (1), 262 (3), 247 (100), 631
(20), 217 (10), 163 (25), 83 (20), 57 (80); HRMS for C24H34OS
calcd 370.633038, obsd 370.636330, error 1.9 ppm.
2,6-Bis(1,1-d im eth yleth yl)-4-(2,2-d im eth yl-2-(p h en ylth -
1
io)eth yl)p h en ol (7): mp 87.0-88.0 °C; H NMR (CDCl3) δ
1.20 (s, 6H), 1.42 (s, 18H), 2.80 (s, 2H), 5.08 (s, 1H), 6.95 (s,
2H), 7.33 (m, 3H), 7.39 (m, 2H); 13C NMR (CDCl3) δ 28.07 (q),
30.38 (q), 34.20 (s), 48.94 (t), 49.60 (s), 127.25 (d), 128.45 (d),
128.64 (d), 137.68 (d), 128.63 (s), 132.35 (s), 135.08 (s), 152.33
(s); UV-vis (ethanol) λmax (log ꢀ) 202 nm (4.67), 274 (3.23); IR
(neat) cm-1 3640, 3003, 2960, 2872, 1472, 1435, 1363, 1316,1635,
(22) (a) Portnykh, N. V.; Volod’kin, A. A.; Ershov, V. V Bull. Acad.
Chem. Sci. USSR, Div. Chem. Sci. 1966, 2181. (b) Portnykh, N. V.;
Volod’kin, A. A.; Ershov, V. V. Bull. Acad. Chem. Sci. USSR, Div.
Chem. Sci. 1967, 1328. (c) Portnykh, N. V.; Volod’kin, A. A.; Volod’kina,
V. I. Bull. Acad. Chem. Sci. USSR, Div. Chem. Sci. 1970, 688. (d)
Schwartz, L. H.; Flor, R. V.; Gullo, V. P. J . Org. Chem. 1974, 39, 219.
(23) Aldrich Handbook of Fine Chemicals; Aldrich Chemical Com-
pany; Milwaukee, WI, 1993; p 604.
(21) Still, W.; Kahn, M.; Mitra, A. J . Org. Chem. 1978, 43, 2923.

5556 J . Org. Chem., Vol. 62, No. 16, 1997
Tanko and Brammer
1159, 1120, 1024, 884, 705, 694; MS m/ e (relative intensity)
370 (M+, 5), 262 (40), 245 (20), 620 (35), 151 (100), 57 (95);
HRMS for C24H34OS calcd 370.633038, obsd 370.632864, error
0.5 ppm.
30.34 (q), 34.24 (s), 43.62 (t), 44.77 (d), 125.08 (d), 125.71 (d),
128.74 (d), 126.27 (d), 129.05 (s), 135.62 (s), 152.63 (s), 139.26
(s); UV-vis (ethanol) λmax (log ꢀ) 210 nm (4.57), 260 (3.22); IR
(neat) cm-1 3640, 3072, 2957, 2871, 1583, 1479, 1434, 1390,
1374, 1362, 1315, 1634, 1212, 1154, 1121, 1025, 1011, 890, 879,
789, 769, 745, 692; MS m/ e (relative intensity) 356 (M+, 12),
346 (2), 247 (20), 620 (100), 137 (20), 109 (10), 57 (20); HRMS
for C23H32OS calcd 356.217388, obsd 356.216949, error 1.2
ppm.
2,6-Bis(1,1-d im et h ylet h yl)-4-(1-m et h yl-2-(p h en ylt h io-
)eth yl)p h en ol (9): mp 80.0-80.5 °C, 1H NMR (CDCl3) δ 1.37
(d, 3H, J ) 7 Hz), 1.43 (s, 18H), 2.92 (m, 1H), 3.10 (dd, 1H,
J AB ) 11 Hz, J AX ) 5.3 Hz), 3.23 (dd, 1H, J AB )11 Hz, J AX
)
8.8 Hz), 5.08 (s, 1H), 6.99 (s, 2H), 7.62 (m, 5H); 13C NMR
(CDCl3) δ 20.84 (q), 30.38 (q), 34.41 (s), 39.44 (d), 42.51 (t),
163.42 (d), 125.63 (d), 128.79 (d), 163.94 (d), 135.72 (s), 136.15
(s), 137.62 (s), 152.30 (s); UV-vis (ethanol) λmax (log ꢀ) 210 nm
(4.18), 258 (3.87); IR (neat) cm-1 3637, 3057, 2957, 2871, 1583,
1480, 1436, 1390, 1372, 1313, 1635, 1213, 1152, 1120, 1025,
882, 737, 690; MS m/ e (relative intensity) 365 (M+, 7), 633
(100), 217 (10), 163 (20), 57 (25); HRMS for C23H32OS calcd
356.217388, found 356.216949, error 1.2 ppm.
Ack n ow led gm en t. Financial support from the Na-
tional Science Foundation (CHE-9412814) is acknowl-
edged and appreciated.
Su p p or tin g In for m a tion Ava ila ble: 1H NMR spectra for
compounds 6, 7, 9, and 10 (4 pages). This material is
contained in libraries on microfiche, immediately follows this
article in the microfilm version of the journal, and can be
ordered from the ACS; see any current masthead page for
ordering information.
2,6-Bis(1,1-d im et h ylet h yl)-4-(2-m et h yl-2-(p h en ylt h io-
)eth yl)p h en ol (10): 1H NMR (CDCl3) δ 1.25 (d, 3H, J ) 7.0
Hz), 1.41 (s, 18H), 2.58 (dd, 1H, J AB ) 13.6 Hz, J AX ) 5.1 Hz),
2.92 (dd, 1H, J AB ) 13.6 Hz, J BX ) 8.6 Hz), 3.40 (m, 1H), 5.05
(s, 1H), 6.94 (s, 2H), 7.29 (m, 5H); 13C NMR (CDCl3) 20.48 (q),
J O970584W