Ketene reactions with the aminoxyl radical TEMPO: Preparative, kinetic, and theoretical studies

Article abstract of DOI:10.1021/jo0011336

Tetramethylpiperidinyloxy (TEMPO, TO·) reacts with ketenes RR¹C=C=O generated by either Wolff rearrangement or by dehydrochlorination of acyl chlorides to give products resulting from addition of one TEMPO radical to the carbonyl carbon and a second to the resulting radical. Reactions of phenylvinylketenes 4b and 4f, phenylalkynylketene 4c, and the dienylketene AcOCMe=CHCH=CHCMe=C=O (11) occur with allylic or propargylic rearrangement. Even quite reactive ketenes were generated as rather long-lived species by photochemical Wolff rearrangement in isooctane solution, characterized by IR and UV, and used for kinetic studies. The rate constants of TEMPO addition to eight different ketenes have been measured and give a qualitative correlation of log k₂(TEMPO) = 1.10 log k(H₂O) -3.79 with the rate constants for hydration of the same ketenes. Calculations at the B3LYP/6-311G**//B3LYP/6-311G** level are used to elucidate the ring opening of substituted cyclobutenones leading to vinylketenes and of 2,4-cyclohexadienone (17) forming 1,3,5-hexatrien-1-one (18).

Full text of DOI:10.1021/jo0011336

J . Org. Chem. 2001, 66, 2611-2617
2611
Keten e Rea ction s w ith th e Am in oxyl Ra d ica l TEMP O:
P r ep a r a tive, Kin etic, a n d Th eor etica l Stu d ies
Annette D. Allen, Bernice Cheng, Michael H. Fenwick, Babak Givehchi, Huda Henry-Riyad,
Valerij A. Nikolaev, Elena Aleksadrovna Shikhova, Daryoush Tahmassebi,
Thomas T. Tidwell,* and Silas Wang
Department of Chemistry, University of Toronto, Toronto, Ontario, M5S 3H6 Canada, and Department of
Chemistry, St. Petersburg State University, St. Petersburg, 198904, Russia
ttidwell@chem.utoronto.ca
Received J uly 25, 2000
Tetramethylpiperidinyloxy (TEMPO, TO•) reacts with ketenes RR1CdCdO generated by either
Wolff rearrangement or by dehydrochlorination of acyl chlorides to give products resulting from
addition of one TEMPO radical to the carbonyl carbon and a second to the resulting radical.
Reactions of phenylvinylketenes 4b and 4f, phenylalkynylketene 4c, and the dienylketene AcOCMed
CHCHdCHCMedCdO (11) occur with allylic or propargylic rearrangement. Even quite reactive
ketenes were generated as rather long-lived species by photochemical Wolff rearrangement in
isooctane solution, characterized by IR and UV, and used for kinetic studies. The rate constants of
TEMPO addition to eight different ketenes have been measured and give a qualitative correlation
of log k2(TEMPO) ) 1.10 log k(H2O) -3.79 with the rate constants for hydration of the same ketenes.
Calculations at the B3LYP/6-311G**//B3LYP/6-311G** level are used to elucidate the ring opening
of substituted cyclobutenones leading to vinylketenes and of 2,4-cyclohexadienone (17) forming
1,3,5-hexatrien-1-one (18).
Free radical additions of H•, CH3•, HO•, F•, Cl•, and
SiH3• to ketene are predicted by molecular orbital
calculations to be highly exothermic.1a Calculations for
the stabilized aminoxyl radical H2NO• at the B3LYP/6-
31G*//B3LYP/6-31G* level indicate that attack at CH2
is endothermic by 7.5 kcal/mol, while addition to the
carbonyl carbon is exothermic by 18.7 kcal/mol (eq 1).1b
This prediction was tested experimentally using tetra-
methylpiperidinyloxy (TEMPO, TO•) with diphenylketene
(1), bisketene 2, and allenylketene 3,1b and each of these
reacted at the carbonyl carbon with one molecule of
TEMPO to give a radical which in the case of 1 gave the
peroxide (OCPh2CO2T)2 on exposure to O2 (eq 2), and
cyclized with expulsion of tetramethylpiperidinyl radical
in the case of 2 and 3.1b Further calculations of reactions
of Cl1c and OH1d radicals with ketenes have appeared,
and there have been experimental studies of reactions
g
of (CF3)2NO•,1e NO•,1f NO2•,1f and NO3•1 with ketenes.
In a preliminary communication1h we reported the
reaction with TEMPO of ketenes RCHdCdO (4a -c) and
5 generated from Wolff rearrangements (eq 3, 4), and of
another TEMPO at Câ (eq 3) or addition of TEMPO with
4e (R ) t-Bu) and 4h (R ) PhO) formed by dehydrochlo-
rination of acyl chlorides.1h In each case facile reaction
occurred, and on the basis of the theoretical1a,b as well
as kinetic and product studies1h the reactions were
interpreted as proceeding as in the example of eq 4
through initial attack of one TEMPO at the carbonyl car-
bon forming an R-acyl radical intermediate. The fate of
the intermediate radical was either simple reaction with
allylic or propargylic rearrangement in the case of 4b,c.
(1) (a) Sung, K.; Tidwell, T. T. J . Org. Chem. 1998, 63, 9690-9697.
(b) Huang, H.; Henry-Riyad, H.; Tidwell, T. T. J . Am. Chem. Soc. 1999,
121, 3939-3943. (c) Hou, H.; Wang, B.; Gu, Y. J . Phys. Chem. A 2000,
104, 320-328. (d) Hou, H.; Wang, B.; Gu, Y. Phys. Chem.. Phys. 2000,
2, 2329-2334. (e) Banks, R. E.; Haszeldine, R. H.; Murray, M. B. J .
Fluorine Chem. 1981, 17, 561-564. (f) Adam, W.; Bottle, S. E.; Grice,
I. D.; Pfeiler, D.; Wentrup, C. Tetrahedron Lett. 1996, 37, 2113-2116.
(g) Seland, J . G.; Noremsaune, I. M. W.; Nielsen, C. J . J . Chem. Soc.,
Faraday Trans. 1996, 92, 3456-3450. (h) Allen, A. D.; Cheng, B.;
Fenwick, M. H.; Huang, W.; Missiha, S.; Tahmassebi, D.; Tidwell, T.
T. Org. Lett. 1999, 1, 693-696.
The adducts of nitroxyl radicals with organic free
radicals have attracted considerable attention because
10.1021/jo0011336 CCC: $20.00 © 2001 American Chemical Society
Published on Web 03/22/2001

2612 J . Org. Chem., Vol. 66, No. 8, 2001
Allen et al.
Ta ble 1. Ca p tu r e of Keten es by TEMP O (TO•)
of their potential utility as free radical initiators,2 and
because of the important role of reversible dissociation
of these TEMPO adducts in living free radical polymer-
ization.3 The kinetic measurements of TEMPO reactions
with ketenes also revealed an unexpected correlation1h
of the rate constants for radical additions with rate
constants for reaction with H2O.4a-c The current inves-
tigation reports full details of our preliminary results1h
and the extension of these studies of nitroxyl additions
to ketenes of other representative types in order to
elucidate the scope of this process. We have also made
further kinetic studies to test the generality of the
previously observed correlation of reactivities and have
carried out molecular orbital calculations to elucidate
routes to ketene formation.
Resu lts
Upon photolysis of diazo ketones 6a -d and diazocy-
clohexanone (6e) in isooctane strong ketenyl IR absorp-
tions of 4a -d and 5 were observed at room temperature
using conventional techniques with frequencies in excel-
lent agreement with literature values.4d,e These ketenes
had previously only been observed by matrix isolation
or fast reaction techniques. Addition of TEMPO to these
solutions led to formation of the products of attachment
of two TEMPO molecules, with rearrangement in the case
of 4b and 4c (Table 1). Diazoacetophenone (6a ) and the
diazo ketone 6b were found not to react with TEMPO at
room temperature, but upon thermolysis of 6a and
photolysis of 6b in toluene, the adducts 7a and E/Z-7b
were isolated, as given in Table 1.
a
E/ Z 3/1 and 2/1 from Wolff rearrangement and acyl chloride,
respectively. Yield given is for Wolff rearrangement. bZ/ E 93/7.
cReference 1d.
Reaction of the acyl chlorides t-BuCH2COCl, 85a (eq 5),
and PhOCH2COCl with Et3N in the presence of TEMPO
in THF at room temperature leads to capture of the
ketenes 4e,f,h , respectively, by addition of two TEMPO
radicals forming 7e,f,h (Table 1). Photolysis of cy-
clobutenone 9a 5b,c in the presence of TEMPO gave the
product 7g of 1,2-addition of TEMPO to ketene 4g (eq
6). The products were characterized by their spectral
properties, including HSQC, NOSY, and COSY NMR
spectra of 7b and 7f. The identities of the Z/E stereo-
isomers of 7f was confirmed by the higher field absorption
of the CH2OT signal in Z-7f, as found for the correspond-
ing CH3 in Z-8 and the acid precursor.5a
(2) (a) Nakamura, T.; Busfield, W. K.; J enkins, I. D.; Rizzardo, E.;
Thang, S. H.; Suyama, S. J . Org. Chem. 2000, 65, 16-23. (b) Benoit,
D.; Chaplinski, V.; Braslau, R.; Hawker, C. J . J . Am. Chem. Soc. 1999,
121, 3904-3920. (c) Braslau, R.; Burrill, L. C., II.; Siano, M.; Naik,
N.; Howden, R. K.; Mahal, L. K. Macromolecules 1997, 30, 6445-6550.
(d) Kothe, T.; Marque, S.; Martschke, R.; Popov, M.; Fischer, H. J .
Chem. Soc., Perkin Trans. 2 1998, 1553-1559.
(3) (a) Georges, M. K.; Veregin, R. P N.; Kazmaier, P. M.; Hamer,
G. K. Macromolecules 1993, 26, 2987-2988. (b) Hawker, C. J . Acc.
Chem. Res. 1997, 30, 373-382. (c) Nakamura, T.; Busfield, W. K.;
J enkins, I. D.; Rizzardo, E.; Thang S. H.; Suyama, S. J . Am. Chem.
Soc. 1997, 119, 10987-10991. (d) Patten, T. E.; Matyjaszewski, K. Acc.
Chem. Res. 1999, 32, 895-903.
(4) (a) Allen, A. D.; Andraos, J .; Kresge, A. J .; McAllister, M. A.;
Tidwell, T. T. J . Am. Chem. Soc. 1992, 114, 1878-1879. (b) Andraos,
J .; Kresge, A. J . J . Photochem. Photobiol. A: Chem. 1991, 57, 165-
173. (c) Andraos, J .; Kresge, A. J . Can. J . Chem. 2000, 78, 508-515.
(d) Wagner, B. D.; Arnold, B. R.; Brown, G. S.; Lusztyk, J . J . Am. Chem.
Soc. 1998, 120, 1827-1834. (e) Schultz, R.; Schweig, A. Z. Naturforsch.
1984, 39B, 1536-1540.
(5) (a) Maccarone, E.; Mamo, A.; Perrini, G.; Torre, M. J . Heterocycl.
Chem. 1981, 18, 395-398. (b) Danheiser, R. L.; Savariar, S.; Cha, D.
D. Org. Synth. 1989, 68, 32-49. (c) Danheiser, R. L.; Savariar, S.
Tetrahedron Lett. 1987, 28, 3299-3302. (d) Danheiser, R. L.; Miller,
R. F.; Brisbois, R. G.; Park, S. Z. J . Org. Chem. 1990, 55, 1959-1964.
(e) Waring, A. J .; Morris, M. R.; Islam, M. M. J . Chem. Soc. C 1971,
3274-3280. (f) Burger A.; Ong, H. H. J . Org. Chem. 1964, 29, 2588-
2592.
Photolysis of the cyclohexadienone 105e to form the
unobserved dienylketene 115e,11b in the presence of TEMPO
led to the isolation of adduct 12 in 7% yield (eq 7). The
stereochemistry shown for 11 is known5e,11b to be strongly
favored based on the structures of products from capture
by nucleophiles. The structure of 12 was established by
spectral means, as well as by an X-ray determination.
The low yield of product isolated in this case and the
reactions below is due in part to the difficulty of removing
excess TEMPO by chromatography, as has been noted
by others.2c
Alkenyl and dienyl ketenes as in eqs 5-7 have long
been of interest, especially as the ring opening of 4,4-
diphenylcyclobutenone to a ketene followed by thermal
cyclization (eq 8) discovered in 1939 (Smith-Hoehn reac-

Ketene Reactions with the Aminoxyl Radical TEMPO
J . Org. Chem., Vol. 66, No. 8, 2001 2613
Ta ble 2. Ra te Con sta n ts for Rea ction of Keten es w ith
TEMP O a n d w ith H2O a t 25 °C
a
Reference 4b. bReference 1e. cReference 4a. dReference 4c.
eReference 8b.
tion) has been developed into a useful synthetic reaction.6
Alkenylketenes have also found application in intermo-
lecular cycloaddition reactions.7
Ta ble 3. Rela tive En er gies (k ca l/m ol) Ca lcu la ted for
Rin g Op en in g of Cyclobu ten on es a n d
2,4-Cycloh exa d ien on e (B3LYP /6-31G*//B3LYP /6-31G)9
Thermolysis of the diazodione E-13 in benzene to
generate ketene 148a in the presence of TEMPO gave a
13% isolated yield of the reduced ester 15 (eq 9). The
structure and stereochemistry of 15 were established by
comparison of the spectra with those reported for the
corresponding methyl ester.8a For the proton CHCO2R
the respective values of δ and J are 3.08 and 10.3 Hz for
R ) CH3,8a and 2.98 and 10.4 Hz for R ) T. The
stereochemistry of this proton cis to the adjacent t-Bu
group in the isolated product was proposed previously8a
to result from isomerization by enolization of an initially
formed product with the proton trans.
a
MP2/6-31G*//HF/6-31G*, ref 10c. bNot done.
6-31G* level were carried out9 for the ketenes 4f,g and
for the cyclobutenones 9a ,b as summarized in Table 3.
The cyclobutenone 9b has been described5f as an unstable
compound, perhaps indicating instability toward ring
opening.
Rate constants for the reactions with TEMPO at 25
°C of the stable ketenes 1 and 2 and of ketenes 4a -d , 4i
(R ) 1-naphthyl),8b and 5 generated by Wolff rearrange-
ments were determined by measurements of the rates of
disappearance of the ketene absorption in the presence
of excess TEMPO. Plots of kobs versus [TEMPO] were
linear, and the slopes gave second-order rate constants
k2(TEMPO) as summarized in Table 2. The positive
intercepts observed in a few cases are assigned to slow
reaction of the ketenes with traces of H2O, as has been
demonstrated in a different case.8c
Previous theoretical studies of the ring openings of
cyclobutenones have been reported,10 including studies
of the effect of 4-substituents on the stereochemistry of
this ring opening.10b Comparison between the current
B3LYP/6-31G* results and previous results10c at the MP2/
6-31G*//HF/6-31G* level (Table 3) indicate that for the
parent the barrier for ring opening is 2.4 kcal/mol less
than B3LYP than for the MP2 level. Similarly the favored
anti-conformation of the alkenylketene is 3.3 kcal/mol
lower in energy at B3LYP than for the MP2 level. It is
expected that the relative differences in energy calculated
For the understanding of the ring opening of phenyl-
cyclobutenones calculations at the B3LYP/6-31G*//B3LYP/

2614 J . Org. Chem., Vol. 66, No. 8, 2001
Allen et al.
Ta ble 4. En er gies (Ha r tr ees) a n d Rela tive En er gies
(k ca l/m ol, p a r en th eses) Ca lcu la ted for Rin g Op en in g of
2,4-Cycloh exa d ien on e (17) (B3LYP /6-31G*+ ZP VE//B3LYP /
6-31G)9 a n d (in br a ck ets) [B3LYP /6-311G**//B3LYP /
6-311G**]
consistently at the B3LYP level (Table 3) provide a
reliable comparison of substituent effects on the ring
opening.
The photochemical ring opening of cyclohexadienones
to dienylketenes (Barton-Quinkert reaction) has been
widely studied experimentally11a-f and occurs when
benzene oxide (16) is converted photochemically to the
parent 2,4-cyclohexadienone 17 which in turn forms
dienylketene 18 upon further photolysis (eq 10).11e Re-
cently, structures and relative energies calculated at the
B3LYP/6-31G**//B3LYP/6-31//G** level for parent 2,4-
cyclohexadienone 17, and four conformations each of E-
and Z-1,3-butadienylketene 18, have been reported.11g To
determine the barrier for interconversion of 17 and 18
we have calculated the energy of the transition structure
17a as well as those of 16, 17, and selected conformations
of 18 at the B3LYP/6-311G**//B3LY/6-311G** level (Table
4).9 Where comparisons are possible these relative ener-
gies are in good agreement with those calculated
previously.11g Details of the calculated structures are
given in the Supporting Information.
The calculated pathway for ring opening of 17 forming
ketene 18 involves a highly twisted transition state 17a
with a barrier of 27.3 kcal/mol. Four minimum energy
conformations Z-18a -d of the product dienylketene were
found (Table 4) of which 18a is favored, but is 18.2 kcal/
mol less stable than 17, with a 9.1 kcal/mol barrier
calculated for ring closure back to 17. The experimental
barrier in solution for ring closure of 11 has been
measured as 13.5 kcal/mol,11b and the lower barrier
calculated for 18 may reflect a steric barrier due to the
6,6-substituents in 11. Photolysis of 17 leads to genera-
(6) (a) Smith, L. I.; Hoehn, H. H. J . Am. Chem. Soc. 1939, 61, 2619-
2624. (b) Tidwell, T. T. Ketenes; Wiley: New York, 1995; Chapter 3.4.1.
(c) Danheiser, R. L.; Casebir, D. S.; Firooznia, F. J . Org. Chem. 1995,
60, 8341-8350. (d) Wang, K. K. Chem. Rev. 1996, 96, 207-222. (e)
Moore, H. W.; Yerxa, B. R. Adv. Strain Org. Chem. 1995, 4, 81-162.
(f) Ohno, M.; Yamamoto, Y. Eguchi, S. Synlett 1998, 1167-1174. (g)
Zhang, S.; Liebeskind, L. S. J . Org. Chem. 1999, 64, 4042-4049.
(7) (a) Loebach, J . L.; Bennett, D. M.; Danheiser, R. L. J . Org. Chem.
1998, 63, 8380-8389. (b) Loebach, J . L.; Bennett, D. M.; Danheiser,
R. L. J . Am. Chem. Soc. 1998, 120, 9690-9691. (c) Bennett, D. M.;
Okomoto, I.; Danheiser, R. L. Org. Lett. 1999, 1, 641-644.
(8) (a) Nikolaev, V. A.; Korneev, S. M.; Terent′eva, I. V.; Korobitsyna,
I. K. Zh. Org. Khim. 1991, 27, 2085-2100. (b) Carter, J .; Fenwick, M.
H.; Huang, W.; Popic, V. V.; Tidwell, T. T. Can. J . Chem. 1999, 77,
806-809. (c) Liu, R.; Lusztyk, J .; McAllister, M. A.; Tidwell, T. T.;
Wagner, B. D. J . Am. Chem. Soc. 1998, 120, 6247-6251.
(9) Gaussian 94W, Revision E.1: Frisch, M. J .; Trucks, G. W.;
Schlegel, H. B.; Gill, P. M. W.; J ohnson, B. G.; Robb, M. A.; Cheeseman,
J . R.; Keith, T.; Petersson, G. A.; Montgomery, J . A.; Raghavachari;
K.; Al-Laham, M. A.; Zakrzewski, U. G.; Ortiz, J . V.; Foresman, J . B.;
Cioslowski, J .; Stefanov, B. B.; Nanayakkara, A.; Challacombe, M.;
Peng, C. Y.; Ayala, P. W.; Chen, W.; Wong, M. W.; Andres, J . L.;
Replogle, E. S.; Gomperts, R.; Martin, R. L.; Fox, D. J .; Binkley, J . S.;
Defrees, D. J .; Baker, J .; Stewart, J . P.; Head-Gordon, J .; Gonzales,
C.; Pople, J . A. Gaussian Inc., Pittsburgh, PA, 1995.
(10) (a) Nguyen, M. T.; Ha, T.; More O’Ferrall, R. A. J . Org. Chem.
1990, 55, 3251-3256. (b) Niwayama, S.; Kallel, E. A.; Sheu, C.; Houk,
K. N. J . Org. Chem. 1996, 61, 2517-2522. (c) McAllister, M. A.; Tidwell,
T. T. J . Am. Chem. Soc. 1994, 116, 7233-7238. (d) Gong, L.; McAllister,
M. A.; Tidwell, T. T. J . Am. Chem. Soc. 1991, 113, 6021-6028. (e)
McAllister, M. A.; Tidwell, T. T. J . Org. Chem. 1994, 59, 4506-4515.
(11) (a) Barton, D. H. R.; Quinkert, G. J . Chem. Soc. 1960, 1-9. (b)
Quinkert, G.; Kleiner, E.; Freitag, B.-J .; Glenneberg, J .; Billhardt, U.;
Cech, K. R; Schmeider, K. R.; Schudok, C.; Steinmetzer, H.-C.; Bats,
J . W.; Zimmermann, G.; Du¨rner, G.; Rehm, D.; Paulus, E. F. Helv.
Chim. Acta 1986, 69, 469-535. (c) Schultz, A. G. CRC Handbook of
Organic Photochemistry and Photobiology; Horspool, W. M.; Song, P.-
S., Eds.; CRC Press: Boca Raton, FL, 1995; Chapter 58, pp 728-736.
(d) Barton, D. H. R.; Chung, S. K.; Kwon, T. W. Tetrahedron Lett. 1996,
37, 3631-3634. (e) J erina, D. M.; Witkop, B.; McIntosh, C. L.;
Chapman, O. L. J . Am. Chem. Soc. 1974, 96, 5578-5580. (f) Capponi,
M.; Gut, I.; Wirz, J . Angew Chem. Int. Ed. Engl. 1986, 25, 344-345.
(g) Parker, J . K.; Davis, S. R. J . Am. Chem. Soc. 1999, 121, 4271-
4277.
tion and direct observation of 18 in matrixes at 77 K,11e,f
but this rapidly decays at -90 °C.11e The photooxidation
of benzene and ozone in an argon matrix at 12 K leads
to the formation of 18 as observed by IR spectroscopy,
and this is proposed to occur by photochemical ring
opening of 17.11g Vinylketenes and allenylketenes are
known to prefer the anti-periplanar geometry,12 consis-
tent with the calculated greater stability of 18a . The less
stable conformations Z-18c,d are both nonplanar, with
dihedral angles of the butadienyl moieties of 42° and 35°,
respectively. The isomer E-18 is calculated to be 1.4 kcal/
mol more stable than Z-18a , but is not accessible in this
reaction.
(12) (a) Brown, R. D.; Godfrey, P. D.; Woodruff, M. Austr. J . Chem.
1979, 32, 2103-2109. (b) Trahanovsky, W. S.; Surber, B. W.; Wilkes,
M. C.; Preckel, M. M. J . Am Chem. Soc. 1982, 104, 6779-6781. (c)
Huang, W.; Fang, D.; Temple, K.; Tidwell, T. T. J . Am. Chem. Soc.
1997, 119, 2832-2838.

Ketene Reactions with the Aminoxyl Radical TEMPO
J . Org. Chem., Vol. 66, No. 8, 2001 2615
Discu ssion
Our previous theoretical1a,b and experimental1b,h,8b
studies have established the mechanism of TEMPO
reaction with ketenes as involving initial in-plane radical
attack on the carbonyl carbon forming acyl-substituted
radical intermediates. The large accelerations on the rate
of TEMPO attack by radical stabilizing aryl, vinyl, and
alkynyl substituents (Table 2) confirm that this first step
is rate limiting. Simple acyl-substituted radicals are
known to have spin density predominantly on the
R-carbon,1a and addition of TEMPO to this carbon may
occur (eq 4). The alkenylketene 4f initially forms the
allylic radical 19a which gives a 93/7 preference for the
Z stereochemistry in the product 7f of attack at Cγ (eq
5). Alkenylketene E-4b gives the radical 19b and TEMPO
addition to Cγ forms an E/ Z mixtures of 7b, and TEMPO
addition to 19c from ketene 4g occurs at CR. The position
of attack at TEMPO on the allylic radicals 19a -c
presumably reflects a combination of steric and electronic
factors.
The reaction of the dienylketene 11 (eq 7) with TEMPO
is expected to form a dienyl radical 20, and formation of
the observed product 12 could occur by TEMPO addition
forming an unobserved unstable acylal 21 which loses
acetic acid (eq 11). The favored radical attack at the
terminus of the dienyl radical parallels that found for
polar additions.11
F igu r e 1. Correlation of rate constants k2 for ketene reactions
with TEMPO with rate constants for hydration.
(Table 2), as illustrated in Figure 1. A qualitative
correlation is observed, r ) 0.93, and such a correlation
between a free radical and a polar reaction is perhaps
surprising. However, in both cases in-plane attack by
oxygen on the carbonyl carbon of the ketene is involved,
and the oxygen of nitroxyl radicals is often considered
as being nucleophilic and possessing some negative
charge.13a The pKa of TEMPO has been estimated as -5.5
( 1,13b indicating the nitroxyl is about 3 pKa units less
basic than CH3OH. Steric effects of the bulky tetra-
methylpiperidinyl moiety may play a role, but this is not
apparent from the result. Investigation of structural
effects on the reactivity of ketenes with TEMPO and
other nitroxyl radicals is continuing in our laboratory.
The calculations of substituent effects on the ring
opening of 2-substituted cyclobutenones (Table 3) show
that while SiH3 favors ring opening, relative to H, by 1.8
kcal/mol, that CH3, F, Cl, and Ph disfavor ketene forma-
tion by 5.2, 13.1, 9.3, and 8.4 kcal/mol, respectively. These
reflect the effects of substituents on ketene stabilities,10d,e
and similar trends were found for substituent effects of
SiH3 and F on ring opening of cyclobutenediones.10c
Differences in the magnitude of the effects between the
two systems may reflect substituent effects on the
electron-deficient cyclobutendiones.
For ring opening of the 3-substituted cyclobutenones,
the alkenylketene is favored for all substituents except
for phenyl, but by lesser amounts than for R ) H. For
both series the anti-conformation of the alkenylketene
is favored, as found previously.10a,d,12 For R ) F and Cl
the ring-opened forms for the 3-substituted series are
predicted to be favored by 12.4 and 6.5 kcal/mol relative
to the 2-series. These large net effects may be attributed
to destabilization of the 2-substituted ketenes and the
3-substituted cyclobutenones by halogen substituents.
The photochemical ring opening of cyclohexadienones
to dienylketenes has been extensively examined and is
known to be a facile process, but the product ketene
Addition of TEMPO to acylketene 14 (eq 9) would give
a crowded but stabilized diacyl radical 22 that could form
15 by hydrogen abstraction, perhaps by disproportion-
ation with another radical 22 (eq 12). A referee has
suggested the possibility that 15 could be formed by
addition of the hydroxylamine TOH to 14. This possibility
cannot be excluded, although VPC analysis of the TEMPO
used revealed the presence of no detectable impurities.
(13) (a) Volodarsky, L. B.; Reznikov, V. A.; Ovcharenko, V. I.
Synthetic Chemistry of Stable Nitroxides; CRC Press: Boca Raton, FL,
1994. (b) Malatesta, V.; Ingold, K. U. J . Am. Chem. Soc. 1973, 95,
6404-6407.
The second-order rate constants for the reactions of
ketenes with TEMPO give the correlation of eq 13 with
the rate constants for hydration of the same ketenes

2616 J . Org. Chem., Vol. 66, No. 8, 2001
Allen et al.
The product was chromatographed four times (1% Et3N in CH2-
Cl2) to give 7d as an oil (404 mg, 0.99 mmol, 76%): 1H NMR
(CDCl3) δ 0.90 (t, 3, J ) 7.2 Hz), 1.0-1.8 (m, 40), 1.96 (m, 22),
4.46 (t, 1, J ) 3.2 Hz). 13C NMR (CDCl3) δ 13.9, 17.0, 17.1,
20.3 20.5, 20.6 20.7, 22.8, 26.6, 32.1, 32.15, 32.2, 33.7, 34.5,
39.2, 39.3, 40.5, 59.6, 59.9, 60.1, 60.6, 83.5, 171.7. IR (CDCl3)
1736 cm-1. EIMS m/z 411 (MH+, 0.05), 254 (17), 156 (28), 140
(100). HREIMS m/z calcd for C24H47N2O3 (MH+) 411.3587,
found 411.3573. Related procedures for the generation and
reaction with TEMPO of 4b,c,e and 5 are given in the
Supporting Information.
rapidly reverts to the reactant, although it can be trapped
with nucleophiles.11 Even incandescent light is satisfac-
tory for ring opening in some cases.11d The ring opening
of cyclohexadienone 17 is calculated to be endothermic
by 18.5 kcal/mol (Table 4) and is 26.3 kcal/mol less
favorable than the comparable process of cyclobutenone,
and this can be attributed to the great release of strain
in the latter process. The great ease of the endothermic
photochemical cyclohexadienone ring opening may be
attributed to the enhancement of light absorption by the
extended chromophore. The barrier for ring opening of
17 is 27.3 kcal/mol, and this is easily accessible upon
photolysis.
Experimentally it has been found for ring-opening of
cyclobutenones 23 (eq 14) with R1 ) Ph that rate ratio
of 80 °C (R ) H)/(R ) CH3) of 8.4,14 in agreement with
the predicted rate-retarding effect of H relative CH3 of
2.0 kcal/mol (Table 3). For R ) CH3 the rate ratio (R1 )
Ph)/(R1 ) CH3) is 2400,14 consistent with enhanced
stabilization of the forming diphenylvinyl group.
Gen er a tion of 2-P h en ylbu ta -1,3-d ien -1-on e (4f) a n d
Tr a p p in g w ith TEMP O. Acyl chloride 8 (0.55 g, 3.05 mmol,
1
96/4 Z/E by H NMR) from the acid5a was added dropwise to
a solution of Et3N (0.73 g, 7.2 mmol) and TEMPO (2.4 g, 15
mmol) in 5 mL of dry benzene with 0.73 g of 4A molecular
sieves and the mixture was stirred overnight. The solution was
filtered and the solvent evaporated, and excess TEMPO was
partly removed by sublimation at 50 °C and 30 Torr. The
residue was chromatographed on silica gel first with 10/90
EtOAc/benzene and then with 20/80 EtOAc/hexanes to give
7f (0.62 g, 1.35 mmol, 44%) as a 93/7 Z/E mixture. Z-7f 1H
NMR (CDCl3) δ 0.86 (s, 6, 2CH3), 1.03 (s, 6, 2CH3), 1.05 (s, 6,
2CH3), 1.07 (s, 6, 2CH3), 1.2-1.7 (m, 12CH2), 4.28 (d, 2, J )
6.4 Hz, CH2OT), 7.04 (t, 1, J ) 6.3 Hz, CHdC), 7.0-7.2 (m,
5), (COSY spectrum used to confirm assignments). 13C NMR
(CDCl3) δ 16.8 (CH2), 16.9 (CH2), 19.9 (CH3), 32.7 (CH3), 33.5
(CH3), 38.7 (CH2), 39.4 (CH2), 59.6 (C), 60.0 (C), 74.3 (CH2-
OT), 127.4, 127.8, 128.8, 133.9, 135.1, 139.8, 166.7 (HSQC
spectrum used to confirm assignments). IR (CDCl3) 1736 cm-1
.
EIMS m/z 456 (MH+, 0.04), 300 (1), 156 (100). HREIMS m/z
calcd for C28H45N2O3 (MH+) 457.3430, found 457.3446. An 1H
NMR signal assigned to the CH2OT of E-7f was observed at δ
4.84 (d, J ) 5.4 Hz).
In summary, the reaction of ketenes with the nitroxyl
radical TEMPO has been shown to be a facile process
that provides a new method for generating acyl substi-
tuted radicals, and is also useful for mechanistic inves-
tigations. It is also shown that even highly reactive
ketenes can be generated and characterized in solution
at room temperature using conventional methodology.
Theoretical calculations provide useful insights into the
factors governing the formation of alkenyl- and di-
enylketenes by ring-opening reactions. Further applica-
tions of these reactions are under investigation.
Gen er a tion of 3-P h en ylbu ta -1,3-d ien -1-on e (4g) a n d
Tr a p p in g w ith TEMP O. A solution of phenylcyclobutenone
9a 5b,c (23 mg, 0.16 mmol) and TEMPO (68 mg, 0.43 mmol) in
5 mL of CDCl3 was degassed by bubbling in Ar and was then
irradiated with 350 nm light 2.5 h at 25 °C. The solvent was
evaporated, and the mixture was chromatographed on silica
gel first with 1/20 EtOAc/toluene and then with 1/10 EtOAc/
hexanes to give 7g (26 mg, 0.058 mmol, 36%) as a yellow gum.
1H NMR (CDCl3) δ 0.86 (s, 6), 1.08 (s, 6), 1.10 (s, 6), 1.17 (s,
6), 1.2-1.7 (m, 12), 5.35 (s, 2), 6.04 (s, 1), 7.3-7.5 (m, 5). 13C
NMR (CDCl3) δ 16.9, 17.0, 19.9, 20.5, 31.9, 33.0, 39.0, 39.5,
59.5, 59.9, 73.6, 116.5, 127.5, 127.9, 128.5, 139.8, 156.9, 166.0.
IR (CDCl3) 1740 cm-1. EIMS m/z 457 (MH+, 0.04), 300 (9), 156
(100). HREIMS m/z calcd for C28H45N2O3 (MH+) 457.3430,
found 457.3417.
Exp er im en ta l Section
Reactions were conducted in dried glassware under an
atmosphere of N2 or Ar. Chromatography was carried out on
silica gel. Triethylamine was dried by distillation from CaH2,
and argon was bubbled through dehydrochlorination solutions
for 1 h at the beginning of these reactions.
Rea ction of P h en oxyk eten e (4h ) w ith TEMP O. A
solution of TEMPO (0.84 g, 5.38 mmol), Et3N (2.15 mmol), and
phenoxyacetyl chloride (0.309 g, 1.81 mmol) in 9.0 mL of CH2-
Cl2 was stirred 60 h with 0.86 g of 4A molecular sieves at room
temperature. The mixture was filtered through Celite, 10 mL
of CH2Cl2 and 10 mL of NaHCO3 solution were added, and
the layers were separated. The aqueous layer was extracted
three times with 10 mL aliquots of CH2Cl2, the combined
organic layers were dried with MgSO4, and the solvent was
evaporated to give a viscous liquid which was chromato-
graphed (7% EtOAc/hexane) and rechromatographed (3%
EtOAc/benzene) to give 7h as a gum (470 mg, 1.054 mmol,
Rea ction of P h en ylk eten e (4a ) fr om Wolff Rea r r a n ge-
m en t w ith TEMP O. A solution of diazoacetophenone (6a ,
21.4 mg, 0.15 mmol) and TEMPO (58.6 mg, 0.38 mmol) in 1
mL of toluene was refluxed for 48 h. Chromatography (10%
EtOAc/hexanes) gave the bis(TEMPO) adduct 7a (28.7 mg,
45%) as a white solid which was recrystallized from MeOH/
H2O, mp 90-92 °C; IR (CDCl3) 1772 cm-1; 1H NMR (CDCl3) δ
0.41 (s, 3, CH3), 0.63 (s, 3), 0.80 (s, 3), 1.02 (s, 3), 1.06 (s, 3),
1.12 (s, 3), 1.24 (s, 3), 1.27 (s, 3), 1.24-1.62 (m, 12), 5.23 (s, 1),
7.20-7.50 (m, 5); 13C NMR (CDCl3) δ 16.84, 17.02, 19.90, 20.09,
20.13, 20.52, 30.73, 31.85, 33.30, 34.07, 39.10, 39.30, 40.22,
59.48, 59.82, 60.30, 60.41, 87.59, 127.47, 127.80, 128.12,
139.32, 170.84; CIMS (isobutane) m/z 431 (MH+, 0.2), 276 (5),
156 (100), 142 (46), 126 (23); EIMS m/z 430 (M+, 0.2), 156 (100).
Gen er a tion of n -Bu tylk eten e (4d ) a n d Tr a p p in g w ith
TEMP O. A solution 1-diazo-2-hexanone (6d , 165 mg, 1.31
mmol) in 75 mL of pentane was irradiated 10 min with 350
nm light, and then TEMPO (2.89 g, 18.5 mmol) was added at
room temperature and the solution stirred 20 h. The pentane
was evaporated and excess TEMPO removed by sublimation.
1
58%); IR (CH2Cl2) 1779 cm-1; H NMR (CDCl3) δ 0.96 (s, 3),
1.01 (s, 3), 1.07 (s, 3), 1.14 (s, 3), 1.16 (s, 3), 1.23 (s, 3), 1.24 (s,
3,), 1.27 (s, 3), 1.31-1.67 (m, 12, 6), 5.84 (s, 1), 6.96-7.29 (m,
5). 13C NMR (CDCl3) δ 16.71, 16.88, 19.93, 20.24, 20.36, 31.40,
31.51, 32.56, 33.27, 38.94, 38.97, 39.84, 40.12, 59.68, 60.13,
60.16, 61.07, 103.22, 116.89, 121.90, 129.22, 156.55, 166.07.
EIMS m/z 447 (M+, 0.5), 290 (12), 140 (100). HREIMS m/z calcd
C
26H43N2O4 447.3223, found 447.3203.
Gen er a t ion of Dien ylk et en e 11 a n d R ea ct ion w it h
TEMP O. A solution of cyclohexadienone 105e (0.266 g, 1.48
mmol) and TEMPO (0.44 g, 2.80 mmol) in 30 mL of hexanes
in a Pyrex tube was degassed by bubbling argon for 15 min
and was then irradiated 4 h with 350 nm light. The product
was chromatographed several times with EtOAc/hexanes to
(14) Mayr, H.; Huisgen, R. J . Chem. Soc., Chem. Commun. 1976,
57-58.

Ketene Reactions with the Aminoxyl Radical TEMPO
J . Org. Chem., Vol. 66, No. 8, 2001 2617
give 12 (0.043 g, 0.10 mmol, 7%), as a white solid, color change
to orange brown 120 °C, mp 160 °C H NMR (CD2Cl2) δ 1.01
absorptions due to 4b (2116.5 cm-1, lit.4d 2116 cm-1) and 4c
(2130.9 cm-1, lit.4d 2131.9 cm-1) so that after irradiation for 1
min the absorptions due to 6b and 6c had decreased by 90%,
and the bands due to 4b and 4c, respectively, had grown with
equal intensity. Three minutes after ceasing photolysis the
band due to 4b had disappeared.
Photolysis of 6b and 6c in isooctane with UV detection led
to almost complete disappearance of the absorption due to the
diazoketones and the appearance of new strong maxima at 223
and 305 nm for 4b, and 269 and 284 nm due to 4c, and bands
at 237 and 248 nm attributed to 1-phenylpropyne from 6c.
1
(s, 6, 2CH3), 1.04 (s, 6, 2CH3), 1.20 (s, 6, 2CH3), 1.23 (s, 6,
2CH3), 1.3-1.7 (m, 12, 6CH2), 2.04 (s, 3, CH3), 4.36 (s, 1, Cd
CHH), 5.09 (s, 1, CdCHH), 6.28 (d, J ) 11 Hz, CdCH), 6.88
(dd, 1, J ) 11, 12 Hz), 7.21 (d, 1, J ) 12 Hz) (assignments
confirmed by g COSY and TOCSY). 13C NMR (CDCl3) δ 13.1,
17.0, 20.7, 20.8, 31.9, 32.3, 39.0, 39.7, 60.2, 60.5, 94.8, 122.6,
127.0, 133.6, 137.8, 160.3, 168.1 (assignments confirmed by g
hexane
HSQC and g HMBC). IR (CDCl3) 1726 cm-1. UV λmax
302
nm (ꢀ ) 15 000). CIMS 433 (MH+, 10), 157 (TOH+, 35).
Gen er a t ion of Acylk et en e 14 a n d R ea ct ion w it h
TEMP O. A solution of trans-diazodione 138a (52 mg, 0.21
mmol) and TEMPO (65 mg, 0.42 mmol) in 4 mL of benzene
was heated 4 h at 80 °C. The solvent was evaporated and the
residue chromatographed with CH2Cl2 to give 15 as an oil (11
Kin etic Mea su r em en ts. Kinetics of the reaction of the
ketenes 4a -d and 4i1d and tetramethyleneketene 5 with
TEMPO were measured by injecting 20 µL of (3.6 to 8.5) ×
10-4 M of 4a -d and 0.0115 M 5 in isooctane prepared as above
into 1.2 mL of a TEMPO solution in isooctane and observing
the decrease in absorption at 249 nm (4a ), 287 nm (4b), 267
nm (4c), 355 nm (4d ), 309 nm (4i), and 390 nm (5). Good first-
order plots with stable infinity values were observed, and the
derived rate constants are reported in Table 2. In the absence
of TEMPO, slow decay of the ketene absorption is observed,
and this is attributed to reaction with traces of moisture, as
demonstrated previously for a different example.8c
1
mg, 0.031, mmol, 15%). H NMR (CDCl3) δ 0.94 (s, 9), 1.01 (s,
9), 1.08 (s, 3), 1.10 (s, 3), 1.17 (s, 3), 1.26 (s, 3), 1.4-1.8 (m, 7),
2.09 (ddd, 1, J ) 15.0, 12.3, 8.4 Hz), 2.28 (dd, 1, J ) 13.8, 8.4
Hz), 2.52 (ddd, 1, J ) 17.6, 10.6, 7.0 Hz), 2.98 (d, 1, J ) 10.4
Hz). 13C NMR (CDCl3) δ 16.5, 20.2, 20.3, 25.2, 27.1, 27.2, 31.3,
31.5, 31.7, 32.0, 39.1, 46.4, 56.8 59.2, 169.4, 210.3. IR (CDCl3)
1757, 1732 cm-1
.
Dir ect Ob ser va t ion of Ket en es Gen er a t ed b y Wolff
Rea r r a n gem en t. Photolysis of 1.0 × 10-4 M diazoacetophe-
none (6a ) in isooctane purged with argon with 350 nm light
for 5 min resulted in the disappearance of the UV spectrum
of 6a and the formation of a new spectrum due to 4a , λmax 248.5
nm, and ꢀ ) 1.2 × 104, assuming complete conversion to 4a .
Similar photolysis of 6a (8.5 × 10-4 M in isooctane) showed
another band at 378.0 nm, ꢀ ) 47, ascribed to the n f p*
transition of 4a . Photolysis of 6a (8.5 × 10-4 M in isooctane)
for 5 min with 300 and 350 nm light gave complete disap-
pearance of the IR bands of 6a at 2108 and 1639 cm-1, and
formation of a new band at 2117 cm-1 for 4a (lit.4d 2118 cm-1
in CH3CN). These absorption bands were fairly long-lived, but
upon standing overnight at 5 °C the IR band at 2117 cm-1
disappeared and was replaced by one at 1713 cm-1, assigned
as due to PhCH2CO2H (an authentic sample gave 1716 cm-1
in isooctane).
The kinetics of 1 were measured by injecting 30 µL of 0.0215
M 1 in isooctane into 2 mL of TEMPO solution in isooctane
which had been degassed with a stream of Ar and monitoring
the decrease in absorption at 404 nm. For 2, kinetics were
measured by injecting 12 µL of 0.662 M 7 into 1.2 mL of
TEMPO solution in mesitylene and measuring the decrease
in absorption at 385 nm. All rate constants (Table 2) are
averages of at least two runs.
Ack n ow led gm en t. Financial support by the Natu-
ral Sciences and Engineering Research Council of
Canada is gratefully acknowledged.
Su p p or tin g In for m a tion Ava ila ble: Experimental de-
tails, spectra, calculated structures, and an X-ray crystal-
lographic file of 12 in CIF format. This material is available
free of charge via the Internet at http://pubs.acs.org.
Photolysis of the diazo ketones 6b and 6c in isooctane in
an IR cell using 300 and 350 nm lamps led to the disappear-
ance of the IR absorptions due to 6b and 6c and the growth of
J O0011336