Clocking Tertiary Cyclopropylcarbinyl Radical Rearrangements

Article abstract of DOI:10.1021/jo961685h

Three independent methods have been employed to estimate the rate constant, k₁, for ring-opening of the 2-cyclopropyl-2-propyl radical, 1, at room temperature. These three estimates are based on chemical trapping of 1 and the ring-opened 4-methylpent-3-ene-1-yl radical by thiophenol (k₁ = (1.6₅ ± 0.4₁) × 10⁷ M^-1 s^-1), 9-azabicyclo[3.3.1]nonane-N-oxyl (k₁ = (1.76 ± 0.34) × 107 M^-1 s^-1) and 2,2,6,6-tetramethylpiperidine-N-oxyl (k₁ = (2.1 ± 0.4) × 10⁷ M^-1 s^-1) and absolute rate constants for nonrearranging radicals structurally related to 1. The mean value for k₁) ((1.84 ± 0.4) × 10⁷ M^-1 s^-1) should be used when 1 is employed as a tertiary alkyl free radical clock at ambient temperatures.

Full text of DOI:10.1021/jo961685h

1210
J . Org. Chem. 1997, 62, 1210-1214
Clock in g Ter tia r y Cyclop r op ylca r bin yl Ra d ica l Rea r r a n gem en ts1
Paul S. Engel,*,2a Shu-Lin He,2a J . T. Banks,2b,c K. U. Ingold,*,2b and J . Lusztyk*,2b
Department of Chemistry, Rice University, 6100 Main, Houston, Texas 77005-1892 and Steacie Institute
for Molecular Sciences, National Research Council of Canada, Ottawa, Ontario, Canada K1A 0R6
Received September 3, 1996X
Three independent methods have been employed to estimate the rate constant, k1, for ring-opening
of the 2-cyclopropyl-2-propyl radical, 1, at room temperature. These three estimates are based on
chemical trapping of 1 and the ring-opened 4-methylpent-3-ene-1-yl radical by thiophenol (k1 )
(1.65 ( 0.41) × 107 M-1 s-1), 9-azabicyclo[3.3.1]nonane-N-oxyl (k1 ) (1.76 ( 0.34) × 107 M-1 s-1) and
2,2,6,6-tetramethylpiperidine-N-oxyl (k1 ) (2.1 ( 0.4) × 107 M-1 s-1) and absolute rate constants
for nonrearranging radicals structurally related to 1. The mean value for k1 ((1.84 ( 0.4) × 107
M-1 s-1) should be used when 1 is employed as a tertiary alkyl free radical clock at ambient
temperatures.
Radicals which undergo unimolecular rearrangements
The only “direct” measurement of k1 involved the
addition of muonium (Mu ) µ+e-), a light isotope of
hydrogen, to 2-cyclopropylpropene.7 The line width of the
have been described as “free radical clocks”.3 Provided
the rate constant, kr, for conversion of the unrearranged
radical, U•, to the rearranged radical, R•, is known (i.e.,
provided the clock has been calibrated)3,4 the rate con-
stant, kT, for any competing reaction involving U• can be
calculated by determining the relative yields of the U•
and R• derived products, e.g.,
muon spin rotation (µSR) spectrum increased with in-
creasing temperature, and from measurements between
233 K and 282 K the following Arrhenius equation for
the ring-opening was obtained:
log (k1*/s-1) ) 14.5 - 10.4/θ
Similarly, if kT is known for a radical which is structur-
ally similar to U• but which does not undergo rearrange-
ment, then a measurement of the [UT]/[RT] ratio will
yield kr and a new radical clock will have been calibrated.
Most cyclopropylcarbinyl radicals rearrange relatively
rapidly to give 3-butenyl radicals, e.g.,3-6
where θ ) 2.3RT kcal/mol. This equation yields
k1*298 K ) 7.5 × 106 s-1, but the Arrhenius parameters are
questionable since the preexponential factor should be
1013.15 s-1 5,6
If one assumes that values of k1* measured
.
in the middle of the experimental temperature range are
reliable, a revised Arrhenius equation can be calculated:
log (k1*/s-1) ) 13.15 - 8.8/θ
which yields k*298 K ) 5.0 × 106 s-1
.
1
Two “indirect” estimates of k1 are 1 order of magnitude
larger than this direct, µSR-determined rate constant.
Both indirect estimates relied on the competitive trapping
of 1 with stable nitroxide radicals, Tempo6 and TMIO,8
e.g., for T ) Tempo:
Cyclopropylcarbinyl radicals make excellent clocks and
many, including the parent radical (see above), have been
calibrated with high precision.3-6 However, there is
considerable uncertainty regarding the rate of ring-
opening of 2-cyclopropyl-2-propyl, 1, the simplest tertiary
cyclopropylcarbinyl radical.
X Abstract published in Advance ACS Abstracts, February 1, 1997.
(1) (a) Presented at the International Chemical Congress of Pacific
Basin Societies, Honolulu, HI, Dec 17-22, 1995, paper No. 944. (b)
Issued as NRCC No. 39136.
(2) (a) Rice University. (b) NRCC. (c) NRCC Research Associate,
1994-1996.
(3) Griller, D.; Ingold, K. U. Acc. Chem. Res. 1980, 13, 317-323.
(4) For the most recent compilation of calibrated clock reactions,
see Newcomb, M. Tetrahedron 1993, 49, 1151-1176.
(5) Newcomb, M.; Glenn, A. G. J . Am. Chem. Soc. 1989, 111, 275-
277.
(6) Bowry, V. W.; Lusztyk, J .; Ingold, K. U. J . Am. Chem. Soc. 1991,
113, 5687-5698.
Analyses for the relative yields of the two hydroxylamine
products, UT and RT, and application of eq I yielded rate
constant ratios, k1T(Tempo)/k1.6 Absolute rate constants for
(7) Burkhard, P.; Roduner, E.; Hochmann, J .; Fischer, H. J . Phys.
Chem. 1984, 88, 773-777.
(8) (a) TMIO is 1,1,3,3-tetramethylisoindoline-2-oxyl. (b) Beckwith,
A. L. J .; Bowry, V. W. J . Am. Chem. Soc. 1994, 116, 2710-2716.
S0022-3263(96)01685-4 CCC: $14.00 Published 1997 by the American Chemical Society

Tertiary Cyclopropylcarbinyl Radical Rearrangements
J . Org. Chem., Vol. 62, No. 5, 1997 1211
Sch em e 1
the trapping of a variety of carbon-centered radicals by
nitroxides have been measured by laser flash photolysis
(LFP).6,9-11 These reactions are somewhat slower than
diffusion-controlled and are subject to considerable sol-
vent10 and steric11 effects. In the Tempo/1 trapping
experiment6 it was assumed that the tert-butyl radical
was a suitable model for 1, i.e., it was assumed that
Resu lts
To obtain k1 we have employed three procedures: (i)
trapping
1 with thiophenol, (ii) trapping 1 with
9-azabicyclo[3.3.1]nonane-N-oxyl (ABNO), (iii) measure-
ment by LFP of the rate constant for trapping by Tempo
of the 2,3-dimethylbut-2-yl radical (chosen because it
would appear to be a much better model for 1 than the
tert-butyl radical whenever steric effects might be im-
portant).
k1T(Tempo) was equal to the LFP measured rate constant
Me3C
T(Tempo)
for tert-butyl trapping by Tempo, k
. With this
assumption k1298 K was estimated to be 5.3 × 107 s-1.6 A
(i) Th iop h en ol Tr a p p in g. The azo compound, 4,16
was photolyzed in fluorobenzene at 22 °C in the presence
of known concentrations of thiophenol (as described in
earlier work with 2 and 3).12 The chemistry (Scheme 1)
was found to be complicated both by the formation of
some isopropylcyclopropane, 5, by the in-cage dispropor-
tionation of photogenerated 1 pairs and by various
addition reactions involving the phenylthiyl radical and
the alkenes formed in the overall reaction. A kinetic
analysis of Scheme 1 yields the following relationship:17
somewhat analogous series of assumptions using TMIO
as the trap gave k2198 K ) 4.2 × 107 s-1 8b
.
The ring openings of the two bicyclic radicals, 2 and 3,
have also been studied using thiophenol to trap the
unrearranged and rearranged radicals.12 With the as-
sumption that Franz et al.’s13 LFP measured rate con-
stant for the tert-butyl/thiophenol reaction, kMe C
T(PhSH), is
3
equal to kT1 (PhSH) values of k2298 K and k2398 K were calcu-
lated to be 1.5 × 107 s-1 and 1.1 × 107 s-1, respectively.12
k1T(PhSH)
5 - (7 + 8 + 9 + 10 + 11)
1
)
×
k1
6
[PhSH]avg
where the structure numbers refer to the measured yields
of the indicated compounds (see Table 1). The conversion
of the azo compound, 4, was about 40%, and product
One order of magnitude disagreement regarding the
correct value of k1 is unacceptable if 1 is to be useful as
a calibrated radical clock. A more reliable value for k1
would permit the calculation of more reliable triplet
lifetimes for several biradicals14 and the calculation of
other kinetic parameters.15 We present, herein, our
attempts to refine k1.
(14) See e.g., Engel, P. S.; Keys, D. E. J . Am. Chem. Soc. 1982, 104,
6860-6861. Engel, P. S.; Keys, D. E.; Kitamura, A. J . Am. Chem. Soc.
1985, 107, 4964-4975. Adam, W.; Grabowski, S.; Scherhag, F.
Tetrahedron Lett. 1988, 29, 5637-5640. Engel, P. S.; Lowe, K. L.
Tetrahedron Lett. 1994, 35, 2267-2270.
(15) See e.g., Engel, P. S.; Wu, A.-Y. J . Org. Chem. 1994, 59, 3969-
3974.
(9) Chateauneuf, J .; Lusztyk, J .; Ingold, K. U. J . Org. Chem. 1988,
(16) Timberlake, J . W.; Martin, J . C. J . Org. Chem. 1968, 33, 4054-
53, 1629-1632.
4060.
(10) Beckwith, A. L. J .; Bowry, V. W.; Ingold, K. U. J . Am. Chem.
(17) The in-cage disproportionation/combination ratio for 1 pairs,
which is given by the product ratio, (7 + 8 + 9 + 10 + 11)/[12], is 6.3
( 0.3. This value is similar to the disproportionation/combination ratio
for pairs of tert-butyl radicals at ambient temperatures in solvents of
similar viscosity to fluorobenzene.18 The cage effect calculated from
the amount of 4 which reacted and from the product ratio, (7 + 8 + 9
+ 10 + 11 + 12)/{[12] + (5 + 6 + 7 + 8 + 9 + 10 + 11)/2} were in
good agreement, viz., 0.54 ( 0.03 and 0.59 ( 0.03, respectively.
Soc. 1992, 114, 4983-4992.
(11) Bowry, V. W.; Ingold, K. U. J . Am. Chem. Soc. 1992, 114, 4992-
4996.
(12) Engel, P. S.; Culotta, A. M. J . Am. Chem. Soc. 1991, 113, 2686-
2696.
(13) Franz, J . A.; Bushaw, B. A.; Alnajjar, M. S. J . Am. Chem. Soc.
1989, 111, 268-275.

1212 J . Org. Chem., Vol. 62, No. 5, 1997
Engel et al.
Ta ble 1. P r od u ct Yield s a n d Der ived Kin etic Qu a n tities
fr om th e P h otolysis of 4 in C6H5F a t 22 °C in th e
P r esen ce of P h SHa
The rate constant for trapping of the tert-butyl radical
by ABNO was measured at 20 °C by 308 nm LFP. The
solvent was again fluorobenzene so as to eliminate any
solvent effects on this rate constant. The tert-butyl
radicals were generated from di-tert-butyl ketone, 15, and
their decay was monitored directly at 320 nm via their
weak absorption at this wavelength.19 The experimental
pseudo-first-order rate constants for tert-butyl decay,
b
[PhSH]o
[PhSH]avg
% 4 leftd
0.039
0.028
41.2
0.059
0.049
42.2
87.6
3.64
2.61
0.13
0.14
0.53
1.25
0.17
0.38
0.539
5.84
0.52
0.56
0.098
0.089
41.3
94.4
4.36
0.147
0.142
39.2
82.0
4.43
0.196
0.194
41.1
89.4
4.86
0.294
0.292
38.9
91.0
5.48
c
prod. bal %e 87.7
5f
3.17
6f
3.01
0.66
ndg
2.21
1.60
1.38
1.08
7f
ndg
ndg
ndg
ndg
8f
0.27
0.63
1.48
0.20
0.40
0.799
6.47
0.59
0.60
0.34
0.50
1.20
0.15
0.36
1.393
6.07
0.49
0.57
0.49
0.54
1.26
0.16
0.38
1.731
6.55
0.46
0.60
0.71
0.52
1.22
0.16
0.39
2.636
6.67
0.57
0.60
9f
0.45
1.09
0.14
0.37
0.275
6.36
0.53
0.59
10f
11f
12f
trap/rearh
disp/comb.i
cagej
cagek
kexptl
, were measured at various ABNO concen-
a
b
Initial 4 is 8.64 µmol in 0.5 mL of C6H5F. Initial PhSH
concentration, M. c Average PhSH concentration, M, during the
trations. A plot of kexptl vs [ABNO], i.e., kexptl ) ko
+
Me3C
T(ABNO)
d
run based on GC. Remaining 4 at the end of the run, by GC.
k
[ABNO], gave a straight line the slope of which
e Product balance. f Product yield, µmol. For structures, see
Me3C
T(ABNO)
yielded k
) (5.7 ( 0.9) × 108 M-1 s-1
.
g
h
Scheme 1. Not detected. Trapped 1/rearranged 1 ) {5 - (7 +
(iii) Tem p o Tr a p p in g of th e 2,3-Dim eth ylbu t-2-yl
Ra d ica l. The tert-butyl radical and 2,3-dimethylbut-2-
yl radical, 17, were generated by 308 nm LFP of ketones
15 and 16, respectively, in benzene20 at 20 °C in the
presence of various known concentrations of Tempo. The
8 + 9 + 10 + 11)}/6. i In cage disproportionation/combination ratio
j
) (7 + 8 + 9 + 10 + 11)/(12), see footnote 17. Total yield of cage
products (∑ 7 f 12) divided by the amount of 4 which was
k
photolyzed, see footnote 17. Cage effect calculated from the
product ratio: (∑ 7 f 12)/{12 + ∑ (5 f 11)/2}; see footnote 17.
Ta ble 2. P r od u ct Ra tios fr om th e P h otolysis of 4 in
C6H5F a t 22 °C in th e P r esen ce of ABNO
[ABNO]init (M)
[ABNO]avg (M)
13/14
6.0 × 10-3
8.0 × 10-3
1.2 × 10-2
1.6 × 10-2
2.0 × 10-2
2.5 × 10-2
3.0 × 10-3
4.0 × 10-3
7.0 × 10-3
1.16 × 10-2
1.6 × 10-2
2.16 × 10-2
0.112
0.158
0.284
0.427
0.567
0.730
balances ranged from 82 to 91%. Despite the fact that
PhSH conversions were as high as 58%, the plot of {5 -
(7 + 8 + 9 + 10 + 11)}/6 vs the average PhSH
concentration in each experiment, [PhSH]avg, was linear
with a correlation coefficient of 0.998. The slope of this
plot yields k1T(PhSH)/k1 ) 8.46 ( 0.29 M-1 at 22 °C.
(ii) ABNO Tr a p p in g. The Bredt’s rule-protected
bicyclic nitroxide, ABNO, is known to be an excellent trap
for alkyl radicals.10,11 Of more importance in the present
context, the rate constants for alkyl radical trapping by
ABNO are much less sensitive to the steric demands of
the alkyl radical than is the case for Tempo and TMIO.11
This means that the tert-butyl radical will be a better
model for 1 in ABNO trapping than it will be in Tempo
trapping experiments.
decays of these two tertiary carbon-centered radicals
were monitored directly via their weak absorptions at 320
nm. The advantage of directly monitoring the kinetics
of decay of 17 is that any reactions of the simultaneously
produced acetyl radical are “invisible”. Plots of kexptl vs
Me3C
T(Tempo)
[Tempo] yielded k
) (2.5 ( 0.4) × 108 M-1 s-1 and
17
k
) (1.0 ( 0.2) × 108 M-1 s-1. Thus, 17 reacts ca.
T(Tempo)
2.5 times more slowly with Tempo than the tert-butyl
radical.
Discu ssion
Photolysis of ca. 50% of the azo compound, 4, (0.02M
initial) in fluorobenzene at 22 °C in the presence of ABNO
gave the two expected hydroxylamines, 13 and 14. The
13/14 ratios at various ABNO concentrations are given
in Table 2. A plot of the 13/14 ratio vs average ABNO
Our work allows three different estimates to be made
of the rate constant, k1, for the ring-opening of the
cyclopropyldimethylcarbinyl radical, 1, at ambient tem-
peratures. These three estimates are based on “indirect”
measurements since they all rely on the chemical trap-
ping of 1 (by PhSH, ABNO, and Tempo) and assumptions
about which nonrearranging radical is an appropriate
model for estimating the absolute rate constants for the
(18) See, e.g., Gibian, M. J .; Corley, R. C. Chem. Rev. 1973, 73, 441-
464. Shuh, H.; Fischer, H. Int. J . Chem. Kin. 1976, 8, 341-356. Shuh,
H. H.; Fischer, H. Helv. Chim. Acta 1978, 61, 2463-2481. Tanner, D.
D.; Rahimi, P. M. J . Am. Chem. Soc. 1982, 104, 225-229.
(19) See, e.g., Huggenberger, C.; Fischer, H. Helv. Chim. Acta 1981,
64, 338-353. Chatgilialoglu, C.; Ingold, K. U.; Lusztyk, J .; Nazran, A.
S. Scaiano, J . C. Organometallics 1983, 2, 1232-1235.
(20) The absorption of 16 was blue-shifted in C6H5F relative to C6H6
and, with only a very limited supply of 16, it proved impossible to do
the LFP experiments in C6H5F.
concentration gave a straight line with a correlation
coefficient of 0.998 and a slope equal to k1T(ABNO)/k1
)
32.4 ( 1.1 M-1
.

Tertiary Cyclopropylcarbinyl Radical Rearrangements
J . Org. Chem., Vol. 62, No. 5, 1997 1213
2,2′-Dicyclop r op yl-2,2′-a zop r op a n e, 4, first synthesized
by Timberlake and Martin16 was prepared in much improved
yield using the method developed by Duismann et al.23 for bis-
tert-alkyl azo compounds. To cyclopropylmethyl ketone (2.4
mL, 25 mmol) in 10 mL of hexane at room temperature was
added hydrazine hydrate (0.7 mL, 14 mmol), and the mixture
was refluxed for 20 h. Additional hexane (20 mL) was added,
and the solution was washed with water and brine. The
organic layer was dried over K2CO3, filtered, and concentrated
to give 1.6 g of cyclopropylmethyl ketone azine (84%). 1H-NMR
(250 MHz, CDCl3): 0.88 (m, 4H), 1.01 (m, 4H), 1.93 (m, 2H),
2.23 (s, 6H). Gaseous Cl2 was bubbled through a solution of
this azine (1.0 g, 6.1 mmol) in 25 mL of hexane at -60 °C until
a yellow color appeared. The 2,2′-dicyclopropyl-2,2′-dichloro-
2,2'-azoethane precipitated as a white solid, and the excess
Cl2 was removed rapidly under aspirator vacuum. After
evaporation of the hexane, the residue was recrystallized from
hexane at -30 °C, yielding 0.85 g (52%) of product. 1H-NMR
(250 MHz, CDCl3) 0.54-0.67 (m, 8H), 1.55 (m, 2H), 1.81 (s,
6H); MS (70 eV), m/ e (%) 234 (1), 199 (18), 103 (100), 67 (80),
41 (50); UV (hexane) λmax 366 nm, ꢀ ) 22. This R,R′-
dichloroazoalkane was converted to 4 as follows. A 1 mL
portion of 2 M Me3Al in hexane was diluted with 5 mL of
hexane and cooled to -70 °C. A solution of R,R′-dichloro-
azoalkane (120 mg, 0.5 mmol) in 1 mL of hexane was added
under Ar. The reaction mixture was stirred for 30 min at -70
°C and was then allowed to warm to room temperature and
was stirred for a further 2.5 h, after which time it was recooled
to -10 °C and quenched by carefully adding a mixture of
hexane/ethanol (4:1) and aqueous H2SO4 (10%). Additional
hexane (60 mL) was added and the organic layer was sepa-
rated. After drying over MgSO4 and filtering, evaporation of
the solvent gave crude 4. Further purification of 4 was carried
out by column chromatography (EtOAc/hexane 0.05:1); yield
81 mg (81%). Liquid, 1H-NMR (250 MHz, C6D6) 0.31 (m, 4H),
0.41 (m, 4H), 1.00 (m, 2H), 1.11 (s, 12H); in reasonable
agreement with the literature;16 13C-NMR (62.5 MHz, CDCl3)
0.29, 20.57, 24.11, 66.88; MS (70 ev), m/ e (%) 153 (1%), 83
(100), 55 (95), 41 (28); IR (neat) 2973 (s), 1463 (m), 1358 (m),
1168 (m, br) cm-1; UV (hexane) λmax 366 nm, ꢀ ) 16.5 (lit.16
λmax 372 nm, ꢀ ) 22).
trapping of 1. These three estimates of k1 are in
remarkably close agreement.
(i) Steric effects on the rates of hydrogen atom abstrac-
tion from thiophenol by carbon-centered radicals would
appear to be small.13 It is therefore reasonably safe to
assume that 1 can be modeled by the tert-butyl radical
for which the thiophenol trapping rate constant has been
measured by LFP.13 From the reported Arrhenius pa-
Me3C
T(PhSH)
rameters13 we calculate that k
) (1.4 ( 0.3) × 108
M-1 s-1 at 22 °C. If this rate constant is combined with
the result of our thiophenol trapping experiment, viz.,
k1T(PhSH)/k1 ) 8.46 ( 0.29 M-1, we obtain k1 ) (1.65
(
0.41) × 107 s-1
.
(ii) Steric effects on the rates of carbon-centered radical
additions to ABNO are also small11 and so we will again
assume that the tert-butyl radical is a suitable model for
1. Combining our LFP measured rate constant,
Me3C
T(ABNO)
k
) (5.7 ( 0.9) × 108 M-1 s-1 with the result of
our ABNO trapping experiments, viz., k1T(ABNO)/k1 ) 32.4
( 1.1 M-1 yields k1 ) (1.76 ( 0.34) × 107 s-1
.
(iii) The trapping by Tempo of carbon-centered radicals
is known to be rather strongly influenced by the steric
requirements of the carbon radical.11 The 2.5 fold slower
Tempo trapping of the 2,3-dimethylbut-2-yl radical, 17,
relative to the tert-butyl radical is, therefore, not too
surprising. If we assume that 17 is an appropriate model
for 1, our original estimate6 that k1 was equal to 5.3 ×
107 s-1 (with a probable error of ( 20%) should be reduced
by a factor of 2.5. This yields a corrected k1 ) (2.1 ( 0.4)
× 107 s-1 21
.
Thus, our three separate estimates of the rate constant
for the ring-opening of the cyclopropyldimethylcarbinyl
radical, 1, are in rather satisfactory agreement. We
therefore believe that their mean value at ambient
temperatures, viz., k1 ) (1.84 ( 0.4) × 107 s-1 should be
fairly reliable.22 We are at a loss to explain why our
present estimate of k1 is three times (or more probably
four times, see Introduction) larger than would be implied
on the basis of the “direct” µSR measurements.
2-Cyclop r op ylp r op en e, 7, was prepared in 98% yield from
cyclopropyl methyl ketone according to a literature proce-
dure.24 Liquid, 1H-NMR (250 MHz, CDCl3) 0.35 (m, 2H), 0.55
(m, 2H), 1.29 (m, 1H), 1.53 (s, 3H), 4.55 (d, 1H, J ) 0.2 Hz),
4.62 (d, 1H, J ) 0.2 Hz) in good agreement with the litera-
ture.24
Exp er im en ta l Section
2-Cyclop r op yl-1-(p h en ylth io)p r op a n e,25 8. A solution of
PhSSPh (79 mg, 0.36 mmol) in 3 mL of PhSH in a Pyrex tube
was purged with nitrogen for 20 min, after which 2-cyclopro-
pylpropene (60 mg, 0.73 mmol) was added and the tube was
irradiated (313 nm) for 2.5 h. The mixture was diluted with
hexane (30 mL) and washed twice with aqueous NaOH (5%),
twice with water, and with brine. The organic layer was dried
over MgSO4. Evaporation of solvents and column chromatog-
raphy (hexane) gave, according to GC, a mixture of four
compounds (120 mg, 86%). The retention time of the smallest
peak (0.5%) coincided with that of 11 while the other two peaks
(4.6% and 1.9%) coincided with 9 and 10 (see below). The 9:10
ratio found in this preparation (2.42) was the same as was
found in the photolysis of 4 in the presence of PhSH. The
major product 8, which constituted 93% of the mixture, was
isolated by preparative TLC (91 mg), and its NMR spectrum
showed the AMX pattern expected for the PhSCH2CH group.
Liquid, 1H-NMR (300 MHz, CDCl3) 0.09 (m, 2H), 0.46 (m, 2H),
0.62 (m, 1H), 0.99 (m, 1H), 1.10 (d, 3H, J ) 6.4 Hz), 2.83 (dd,
In str u m en ta tion . NMR spectroscopy: IBM AF-300 or
Bruker AC-250 with δ relative to either CDCl3 (δ 7.26) or C6D6
(δ 7.15). IR spectroscopy: Nicolet 205 FT-IR. Mass
spectrometry: Finnigan MAT 95, all HRMS were obtained in
CI mode. UV spectroscopy: HP8452A diode array spectrom-
eter. Analytical GC: HP5890A with a DB-5 capillary column
(0.25 mm × 30 m) and HP3365 software. Preparative GC:
Antek 300 with a 0.25 in × 10 ft 10% FFAP on Chromosorb
W column.
Ma ter ia ls. Ether and THF were distilled from Na/Ph2CO,
hexane was distilled from sodium, and benzene and CH2Cl2
were distilled from CaH2. Decane, used as a GC internal
standard, was distilled. Thiophenol was distilled immediately
before use. Tempo was sublimed, and di-tert-butyl ketone was
passed through basic alumina prior to use. 2-Cyclopropylpro-
pane, 2-methyl-2-pentene, and (E)-2-methyl-2-pentenoic acid
(all from Aldrich) were used as received.
1H, J 1 ) 8.1 Hz, J 2 ) 12.5 Hz), 3.17 (dd, 1H, J 1 ) 4.8 Hz, J 2
)
12.5 Hz), 7.14-7.33 (m, 5H), in satisfactory agreement with
(21) Since the absolute rate constants for trapping of a wide variety
of carbon-centered radicals by Tempo and by TMIO8a under the same
experimental conditions are very similar,11 it would appear reasonable
the literature25 provided 1.82 in this reference is a misprint
to use the same factor of 2.5 to correct the TMIO-derived value of k1
8b
equal to 4.2 × 107 s-1
,
which yields a corrected k1 ) 1.68 × 107 s-1
.
(23) Duismann, W.; Beckhaus, H.-D.; Ru¨chardt, C. Tetrahedron Lett.
(22) For comparison, the 25 °C rate constant for ring-opening of the
1974, 265-268.
primary cyclopropylmethyl radical is well established as 9.4 × 107 s-1 4,5
,
(24) Teraji, T.; Moritani, I.; Tsuda, E.; Nishida, S. J . Chem. Soc.
and for the ring-opening of the secondary 1-cyclopropylethyl radical
(C) 1971, 3252-3257.
the reported Arrhenius parameters yield 25 °C rate constants of 1.3 ×
(25) Huyser, E. S.; Taliaferro, J . D. J . Org. Chem. 1963, 28, 3442-
7
8b
6
107 s-1
,
2.9 × 107 s-1
,
and 4.5 × 107 s-1
.
3445

1214 J . Org. Chem., Vol. 62, No. 5, 1997
Engel et al.
for 2.82; 13C-NMR (62.5 MHz, CDCl3) 3.50, 4.44, 17.54, 19.13,
38.64, 40.95, 125.39, 128.51, 128.72, 137.58; HRMS, Cald. for
C12H16S: 192.09727. Found: 192.09721.
9-Aza bicyclo[3.3.1]n on a n e-N-oxyl, ABNO, was synthe-
sized by the method employed previously10,28,29 except that
phenylselenol was used to demethylate N-methyl-9-azabicyclo-
[3.3.1]nonane.30
2-Meth yl-1-(p h en ylth io)p en t-2-en e, 9 a n d 10. (E)-2-
Methyl-2-pentenol26 was prepared by LiAlH4 reduction of (E)-
2-methyl-2-pentenoic acid. 1H-NMR (250 MHz, CDCl3) 0.96
(t, 3H, J ) 7.5 Hz), 1.66 (s, 3H), 2.10 (m, 2H), 3.99 (s, 2H), 5.4
(t, 1H, J ) 6.5 Hz). This alcohol was converted to 9 and 10 as
follows: (E)-2-methyl-2-pentenol (600 mg, 6 mmol) in 9 mL of
ether and 4 mL of HMPA was treated at room temperature
with 4.6 mL of 1.3 M MeLi in ether, followed by TsCl (1.2 g,
6.1 mmol) in 5 mL of ether and 2 mL of HMPA. A solution of
PhSLi (1.32 g, 12 mmol) in 5 mL of ether and 2 mL of HMPA
was added, and the reactants were stirred overnight. A white
solid precipitated. Ether (50 mL) was added, the mixture was
washed with water four times, and the organic layer was dried
over MgSO4. After filtration and evaporation of solvents, the
mixture of E,Z products was isolated by column chromatog-
raphy (hexane). The yield was 820 mg (72%, E:Z ) 6:1).
Liquid, HRMS, Calcd for C12H16S: 192.09727. Found:
192.09752. Although the E,Z mixture was never separated,
the individual NMR spectra could be deduced by inspection:
9, 1H-NMR (250 MHz, CDCl3) 0.86 (t, 3H, J ) 7.5 Hz), 1.74 (s,
3H), 2.00 (m, 2H), 3.49 (s, 3H), 5.20 (m, 1H), 7.25-7.45 (m,
5H); 13C-NMR (62.5 MHz, CDCl3) 13.88, 15.01, 21.31, 44.32,
126.2, 127.5, 128.6, 130.5, 131.4, 136.5; 10, 1H-NMR (250 MHz,
CDCl3) 0.84 (t, 3H, J ) 7.5 Hz), 1.87 (s, 3H), 1.95 (m, 2H),
3.55 (s, 3H), 5.29 (m, 1H), 7.15-7.45 (m, 5H); 13C-NMR (62.5
MHz, CDCl3) 14.23, 22.37, 29.70, 36.36, 126.4, 128.7, 129.5,
130.6, 136.8, 137.2.
3,3,4-Tr im eth yl-2-p en ta n on e,31-33 16, was prepared by
methylation of isovaleronitrile with lithium isopropylcyclo-
hexylamide and CH3I followed by treatment with CH3Li. The
overall yield was 65%. Liquid, 1H-NMR (250 MHz, CDCl3) 0.83
(d, 6H, J ) 7.0 Hz), 1.01(s, 6H), 2.00 (m, 1H), 2.13 (s, 3H); 13
C
NMR (62.5 MHz, CDCl3) 17.33, 20.20, 24.99, 33.63, 50.79; IR
(neat) 1700 cm-1
.
P h otolyses. Unless otherwise noted, all samples for pho-
tolysis were degassed and sealed in 5-mm standard wall Pyrex
tubes. Irradiations were carried out with a 500-W Oriel high-
pressure mercury lamp employing either a 313 nm or a 366
nm filter solution. The tubes were cracked open and the
contents analyzed by GC using the following conditions:
injector, 200 °C; detector, 250 °C; oven, 35 °C for 2.5 min and
then ramp at 10 °C/min. The weight response factor of 8
relative to decane was established as 0.647, and all of the other
sulfides were assumed to have the same response factor.
Authentic samples of 2-methyl-2-(phenylthio)pentane, 2-meth-
yl-3-(phenylthio)pentane, 2-cyclopropyl-2-(phenylthio)propane,
and 4-methyl-1-(phenylthio)-3-pentene were prepared and
found to be absent from the products formed by the irradiation
of 4 with thiophenol. ABNO was shown to be photostable by
irradiating it in fluorobenzene for 60 min, twice the time of
the azoalkane runs. This control experiment showed no
decrease in the visible absorbance, consistent with the litera-
ture on photolysis of nitroxyl radicals.34,35 To check the
stability of the hydroxylamines to GC, a sample of 4 was
irradiated with ABNO in C6D6 in an NMR tube until 4 was
gone. The ratio of 13 to 14 found by GC was the same as that
determined by NMR. This observation and the symmetrical
shape of their GC peaks showed the stability of these hydroxyl-
amines to GC analysis; in fact, we were also able to quantify
4 and ABNO itself in the same run.
2-Meth yl-3-(p h en ylth io)-1-p en ten e, 11, was prepared
from 3-chloro-2-methylpropene following a literature proce-
dure.27 To a suspension of sodium methoxide (1.8 g, 30 mmol)
in methanol (40 mL) was added PhSH (3.3 g, 30 mmol) at room
temperature. After 30 min, 3-chloro-2-methylpropene (3.6 g,
40 mmol) was added, and the reaction mixture was stirred
for 3 h. Following removal of a white precipitate, the filtrate
was diluted with ether (100 mL) and washed 3× with water.
After drying over MgSO4 and filtering, the solvent was rotary
evaporated, and the residue was distilled under vacuum to
yield 4.2 g of phenyl methallyl sulfide (82%, 78-80 °C/1.8
mmHg). 1H-NMR (250 MHz, CDCl3) 1.87 (s, 3H), 3.54 (s, 2H),
4.83 (s, 2H), 7.21-7.38 (m, 5H). BuLi (1.6 mL, 2.5 M in
hexane) was added at -78 °C to this sulfide (656 mg, 4 mmol)
dissolved in THF (10 mL), causing a yellow coloration. After
20 min EtI (858 mg, 5.5 mmol) was added and the yellow color
disappeared. The reaction mixture was warmed to room
temperature for 2 h. Water (10 mL) was added, and the
solution was extracted with ether (3 × 50 mL). The ether
solution was dried over MgSO4, filtered and concentrated.
La ser F la sh P h otolysis (LF P ). The equipment and
experimental procedures have been described previously.36 The
tert-butyl radical and 2,3-dimethylbut-2-yl radical, 17, were
generated by 308 nm LFP of solutions containing their ketone
precursors (di-tert-butyl ketone and 3,3,4-trimethyl-2-pen-
tanone, respectively) at a concentration giving an optical
density of ∼0.3 at 308 nm. The decays of these tertiary alkyl
radicals were directly monitored at 320 nm.19 Pseudo-first-
order rate constants (kexptl) were determined by fitting digitally
averaged decay curves from 10 laser flashes. Absolute second-
order rate constants were calculated by least-squares fitting
of kexptl vs [TEMPO] (or [ABNO]) for five different nitroxide
concentrations in the range (0.6-5.0) × 10-3 M.
1
Flash chromatography gave pure 11 (690 mg, 91%). Oil, H-
NMR (250 MHz, CDCl3) 0.98 (t, 3H, J ) 7.3 Hz), 1.73 (m, 2H),
1.77 (s, 1H) 3.55 (m, 1H), 4.64 (s, 1H), 4.75 (m, 1H), 7.23-
7.39 (m, 5H); 13C-NMR (62.5 MHz, CDCl3) 12.17, 17.61, 25.78,
58.09, 113.6, 126.8, 128.5, 132.4, 135.5, 143.5; IR (neat) 2966
(s), 1480 (s), 1438 (s), 1087 (m), 894 (s) cm-1; HRMS, Calcd
for C12H16S: 192.09727. Found: 197.09709.
Isola tion of 2,3-Dicyclop r op yl-2,3-d im eth ylbu ta n e, 12.
A 39 mg portion of 4 in a Pyrex tube was irradiated (366 nm)
under nitrogen for 2 h after which 0.5 mL of hexane was added,
and 12 was isolated by preparative GC (9 mg, 25%). Liquid,
1H-NMR (250 MHz, CDCl3) 0.15-0.30 (m, 8H), 0.71 (s, 12H),
0.98 (m, 2H); 13C-NMR (62.5 MHz, CDCl3) 0.93, 17.51, 20.43,
37.87; HRMS, Calcd for C12H22 + H+: 167.17998. Found:
167.18043.
J O961685H
(28) Mendenhall, G. D.; Ingold, K. U. J . Am. Chem. Soc. 1973, 95,
6395-6400.
(29) Nelsen, S. F.; Weisman, G. R.; Clennan, E. L.; Peacock, V. E.
J . Am. Chem. Soc. 1976, 98, 6893-6896.
(30) Reich, H. J .; Cohen, M. L. J . Org. Chem. 1979, 44, 3148-3151.
(31) Dubois, J . E.; Cosse´-Barbi, A. Spectrochim. Acta, Part A 1972,
28, 523-537.
(32) Conia, J .-M. Bull. Soc. Chim. Fr. 1956, 1392-1398.
(33) Zook, H. D.; Smith, W. E.; Greene, J . L. J . Am. Chem. Soc. 1957,
79, 4436-4439.
(34) Buchachenko, A. L.; Khloplyankina, M. S. Kinet. Catal. (Transl.
of Kinet. Katal.) 1967, 8, 624-628.
(26) Burrell, J . W. K.; Garwood, R. F.; J ackman, L. M.; Oskay, E.;
Weedon, B. C. L. J . Chem. Soc., C 1966, 2144-2154.
(27) Grieco, P. A.; Boxler, D.; Hiroi, K. J . Org. Chem. 1973, 38,
2572-2573.
(35) Nelsen, S. F.; Bartlett, P. D. J . Am. Chem. Soc. 1967, 88, 143-
149.
(36) Kazanis, S.; Azarani, A.; J ohnston, L. J . J . Phys. Chem. 1991,
95, 4430-4435.