3540 J . Org. Chem., Vol. 62, No. 11, 1997
Engel et al.
observation that r∞ ) 3ro implies that the recombination
rate kr of free 5 and 6 pairs is twice their disproportion-
ation rate kp. A much higher PhSH concentration (1.42
M) caused only a 14% additional rate increase, which we
attribute to the known effect of solvent polarity on TAA
thermolysis rates10 and which therefore indicates that
0.123 M PhSH was a completely effective scavenger. H•
transfer from PhSH to nitroxyl radicals is facile,21,25 but
no rate constant seems to be available. The rate constant
for 5 with PhSH is at least 106 M-1 s-1 at the reaction
temperature of 120 °C,26 a high enough value that all
free radicals should be scavenged at the PhSH concentra-
tions employed.
discrepancy might be attributed to the lower temperature
of the reaction involving 1, disproportionation to recom-
bination ratios do not vary that strongly with tempera-
ture.14,33 A partial explantion for this change in products
is that recombination is always the main fate of 5 and 6
pairs but at 170 °C, recombination is invisible and
disproportionation, an irreversible process, is the only
observed reaction.
In summary, alkoxyamine 3 is more stable thermally
than analogs derived from stable nitroxyl radicals.2,4,6,7,10
On heating to 170 °C, it disproportionates mainly to
styrene and diethylhydroxylamine 2 but, in a new reac-
tion, it also yields ethylbenzene and nitrone 12. The
latter then undergoes 1,3-dipolar cycloaddition to the
formed styrene to afford diastereomeric oxazolidines 8
and 9. Some of the styrene suffers attack by R-phenethyl
radicals, yielding dimers 10 and 11. The reaction of
diethylhydroxylamine with R-phenethyl radical is slow
(k ) 5 × 103 M-1 s-1) relative to many other hydrogen
transfers, despite its exothermicity of 13.6 kcal/mol.
The formation of 10, 11, 13, and 14 from 1, 2, and
styrene shows that styrene suffers attack by 5 despite
the presence of the hydrogen donor 2, and a quantitative
treatment allows calculation of an approximate H• trans-
fer rate. The rate constant for styrene polymerization
(kprop) at 120 °C is 2.2 × 103 M-1 s-1, based on our own
least squares fit of 39 data points.27 In view of the poor
correlation coefficient (0.886) and the extrapolation out-
side the experimental temperature range, kprop is more
realistically expressed as (2 ( 1) × 103 M-1 s-1. Using
Exp er im en ta l Section
Gen er a l Meth od s. Melting points (uncorrected) were
obtained on a Mel-Temp apparatus. Benzene, diethylhydroxyl-
amine, styrene, thiophenol, toluene, and tetrahydrofuran were
purified by distillation. The NMR solvents CDCl3 and C6D6
from Cambridge Isotope Laboratory were used without further
purification.
•
this rate as a radical clock, we derive the equation kH /
kprop ) ([styrene]/[2])(moles of PhEt/(moles of 10 + 11 +
13 + 14 + unknowns)). It follows from the product
composition (Table 2)28 and the initial concentration of
styrene and 2 that the hydrogen transfer rate from 2 to
5 is 5 × 103 M-1 s-1 at 120 °C. From the bond
dissociation energy of ethylbenzene29 and Tempo13 (85.4
and 71.8 kcal/mol, respectively), we calculate that this
slow hydrogen transfer is exothermic by 13.6 kcal/mol.
Previous literature on H• abstraction rates from hydroxyl-
amines is confined to the observations that t-Bu• ab-
stracts H• from 2 1.2 times faster than from isoprene30
and that the H• transfer rate from 2 to ethyl radicals in
NMR spectra were recorded on a Bruker AC-250 with
chemical shifts (δ, ppm) relative to internal TMS, hexa-
methyldisiloxane (1H δ ) 0.115), or solvent signal (CDCl3 1H
δ ) 7.26, 13C δ ) 77.0; C6D6 1H δ ) 7.15, 13C δ ) 128.5). ESR
spectra were run on a Varian E-12 spectrometer. Mass spectra
were taken on Finnigan MAT 95 mass spectrometer. UV
spectra were obtained on a Hewlett-Packard 8452A diode array
spectrometer. Analytical GC was carried out on a Hewlett-
Packard 5890A gas chromatograph equipped with a DB-5
capillary column (0.25 mm × 30 m) and FID. Analytical GC
data were collected and manipulated on an IBM PC compatible
computer. Preparative GC of 3 was attempted on an Antek
the gas phase at 25 °C is 7.7 × 105 M-1 s-1 31
.
Our rate
constant is comparable in magnitude to primary alkyl
abstracting H• from dodecane, k ) 3.5 × 103 M-1 s-1 at
100 °C.32
1
300 with a /4 in. × 10 ft 10% FFAP on Chromosorb W column.
All samples for thermolysis were freeze-thaw degassed
three times and sealed on a vacuum line. The tubes were
immersed completely in a well-stirred DC-200 silicone oil bath
contained in a 1.5 gallon Dewar flask. The bath temperature
was regulated by a Bayley Model 123 temperature controller
and was measured with a Hewlett-Packard Model 3456A
digital voltmeter and a platinum thermometer. The least
squares program LINFIT was used to calculate the rate
constants and uncertainties for each first-order plot. The tubes
were cracked open and the contents analyzed by GC with a
0.12 mm × 30 m DB-5 capillary column. The following GC
conditions were used: head pressure, 16 psi; He flow at
detector, 53 mL/min; injector, 210 °C; detector, 250 °C; oven
initial temperature, 35 °C; initial time, 2.5 min; program rate,
10 °C/min; oven final temperature, 250 °C; final time, 10 min.
All hydrocarbons were assumed to have the same weight
response factor so that the GC peak areas of the products
(Table 2) correspond to an approximate molar ratio of PhEt:
(10 + 11 + 13 + 14 + unknowns) of 2.25:1.00. The 23.26 min
peak was assumed to be a styrene dimer.
Azo-r-p h en yleth a n e (1) was made by the method of
Cohen.8 The crude material was purified by recrystallization
from ethanol at -15 °C: mp 71 °C (lit.8 mp 72-73 °C); UV
(hexane) λmax ) 360 nm; ꢀ ) 46; 1H NMR (250 MHz, CDCl3) δ
1.53 (6H, d), 4.62 (2H, q), 7.38 (10H, m).
N-(1-p h en yleth oxy)d ieth yla m in e (3). A 250 mL three-
necked flask equipped with a condenser, nitrogen inlet, and
magnetic stirrer was flushed with nitrogen and charged with
di-tert-butyl peroxide (18.75 mL, 0.1 mol) and an excess of
The thermolysis of 3 at 170 °C, where disproportion-
ation is the major net reaction, contrasts with the
generation of 3 from 1 at 120 °C, where only a little
styrene is observed. The fact that 3 is a product of 1 +
2 shows that R-phenethyl and 6 are present at the same
time, yet they hardly disproportionate. Although this
(25) Forrester, A. R.; Hay, J . M.; Thomson, R. H. Organic Chemistry
of Stable Free Radicals; Academic Press: London, 1968; p 229.
(26) Kuwae, Y.; Kamachi, M. Bull. Chem. Soc. J pn. 1989, 62, 2474.
(27) Polymer Handbook, 3rd ed., Brandrup, J ., Immergut, E. H.,
Eds.; Wiley-Interscience: New York, 1989; p 76. Our least squares fit
of ln(kprop/T) versus 1/T gave a slope of -4357 ( 375, an intercept of
12.84 ( 1.23, and a correlation coefficient of 0.886. These values
correspond to Ea ) 9.27 kcal/mol and log A ) 8.50, similar to the given
Ea ) 8.96 kcal/mol, log A ) 8.38.
(28) The small unknown at 23.26 min was assumed to be an
unsaturated styrene trimer. Of greater concern for the calculation than
its actual structure is the possibility that higher styrene oligomers were
not seen by GC but should be included as attack on styrene. When a
solution of 0.02 M 1, 0.02 M 2, and 0.20 M styrene in benzene was
thermolyzed at 120 °C, many GC peaks of longer retention time were
observed; however, these were absent in the 0.02 M styrene experi-
ment.
(29) McMillen, D. F.; Golden, D. M. Ann. Rev. Phys. Chem. 1982,
33, 493.
(30) Encina, M. V.; Lissi, E. A.; Soto, H. J . Photochem. 1981, 16,
43. See also: Khan, J .; Casey, D.; Linschitz, H.; Cohen, S. G. J . Org.
Chem. 1991, 56, 6080.
(31) Abuin, E.; Encina, M. V.; Diaz, S.; Lissi, E. A. Int. J . Chem.
Kinet. 1978, 10, 677.
(32) Burton, A.; Ingold, K. U.; Walton, J . C. J . Org. Chem. 1996, 61,
3778.
(33) Gibian, M. J ; Corley, R. C. Chem. Rev. 1973, 73, 441.