Spin delocalization by triple-bonded functionalities in propargyl and heteropropargyl radicals, assessed from the EPR-spectral D parameter of 1,3-cyclopentanediyl triplet diradicals

Article abstract of DOI:10.1021/jo026579w

The cyclopentane-1,3-diyl triplet diradicals T and T′ with the triplet-bonded acetylene, cyano, and isocyano functionalities at one of the radical sites are readily prepared from the corresponding azoalkanes by photodenitrogenation in a 2-methyltetrahydro

Full text of DOI:10.1021/jo026579w

Sp in Deloca liza tion by Tr ip le-Bon d ed F u n ction a lities in P r op a r gyl
a n d Heter op r op a r gyl Ra d ica ls, Assessed fr om th e EP R-Sp ectr a l D
P a r a m eter of 1,3-Cyclop en ta n ed iyl Tr ip let Dir a d ica ls
Waldemar Adam* and Claudio M. Ortega-Schulte
Institute of Organic Chemistry, University of Wu¨rzburg, Am Hubland, D-97074 Wu¨rzburg, Germany
adam@chemie.uni-wuerzburg.de
Received October 17, 2002
The cyclopentane-1,3-diyl triplet diradicals T and T′ with the triplet-bonded acetylene, cyano, and
isocyano functionalities at one of the radical sites are readily prepared from the corresponding
azoalkanes by photodenitrogenation in a 2-methyltetrahydrofuran (2-MTHF) matrix at 77 K. The
EPR-spectral D values of these triplet diradicals show that the spin delocalizing ability of the triple-
bonded π substituent follows the order -CtCH > -NC ≈ -CN. Good correlations of the D values
have been obtained with the hyperfine coupling constants (aH) and with the calculated spin densities
(PM3/AUHF-CI method) of the corresponding monoradicals M. The propargyl-type mesomeric
structure is favored over the allenyl-type contributor for all three triple-bonded functionalities;
spin delocalization is less pronounced in the heteropropargyl derivatives due to the electronegativity
effect of the nitrogen atom.
In tr od u ction
calculated value is aH ) 20.0 G),5 which clearly indicates
a higher spin localization at the methylene carbon atom
than in the propargyl case. This is expected since the
electronegative nitrogen atom will resist spin delocaliza-
tion to populate the heteroallenic structure.5
The spin delocalization in allyl1 and heteroallyl2 radi-
cals has been extensively studied during the past few
years, whereas less is known about the electronic stabi-
lization by triple-bonded radical systems. Consequently,
the incentive of the present study was to assess the ex-
tent of spin delocalization in the propargyl radical sys-
tems and its nitrogen analogues, namely the cyanomethyl
The EPR-spectral work on the isocyanomethyl radical
was conducted by Williams and Wang,6 but strong
anisotropic effects precluded the determination of a
precise aH coupling constant, such that the extent of spin
delocalization by the isocyano group is not accurately
accessible. These EPR-spectral results manifest the dif-
ficulties in acquiring information on the spin delocaliza-
tion for unconventional groups such as the isocyano
functionality by means of hyperfine coupling constants.
Low-temperature persistent 1,3-cyclopentanediyl trip-
let diradicals may be readily generated through the
photochemical deazetation of diazabyciclo[2.2.1]heptene
(DBH) derivatives.7 These triplet diradicals have been
characterized by the zero-field-splitting (zfs) parameters
D and E. For such localized triplet 1,3-diradicals, the D
parameter depends on the interspin distance dab and the
spin densities Fa and Fb at the respective radical sites a
and b, as displayed in eq 1. The spin density as well as
•
•
(NC-CH2 ) and isocyanomethyl (CN-CH2 ) radicals.
Early EPR-spectral work on the propargylic radical
was carried out by Collin and Lossing,3 who generated
it by hydrogen abstraction from methylacetylene. Sub-
sequent more detailed EPR studies by Kochi and Krusic4
concluded from the hyperfine coupling constants (see
structures below) that the propargyl mesomeric structure
is favored over the allenylic one. The hyperfine coupling
constants (aH) calculated by Benson’s5 group (INDO
method) confirmed the preference for spin localization in
the propargylic (aH ) 19.2 G) versus the allenic (aH
10.9 G) structure.
)
For the cyanomethyl radical, the observed coupling
constant was measured to be aH ) 20.8 G (the INDO
(1) (a) Adam, W.; Emmert, O. J . Org. Chem. 2000, 65, 4127-4130.
(b) Adam, W.; Emmert, O. J . Org. Chem. 2000, 65, 5664-5667. (c)
Adam, W.; van Barnefeld. C.; Emmert, O. J . Chem. Soc., Perkin Trans.
2 2000, 2, 637-641.
the distance dependences have been confirmed experi-
(2) (a) Adam, W.; Ortega-Schulte, C. M. J . Org. Chem. 2002, 67,
4565-4573. (b) Creary, X.; Engel, P. S.; Kavasalukas, N.; Pan, L.; Wolf,
A. J . Org. Chem. 1999, 64, 5634-5643.
(5) Benson, H. G.; Bowles, A. J .; Hudson, A.; J ackson, R. A. Mol.
Phys. 1971, 4, 713-719.
(6) Wang, J .; Williams, F. J . Am. Chem. Soc. 1972, 94, 2930-2934.
(3) Collin, J .; Lossing, F. P. J . Am. Chem. Soc. 1957, 79, 5848-5853.
(4) Kochi, J. K.; Krusic, P. J. J . Am. Chem. Soc. 1970, 13, 4110-4114.
10.1021/jo026579w CCC: $25.00 © 2003 American Chemical Society
Published on Web 12/31/2002
J . Org. Chem. 2003, 68, 1007-1011
1007

Adam and Ortega-Schulte
mentally and theoretically,8 which provide valuable
structural and electronic information on spin delocaliza-
tion and radical stabilization in triplet diradicals. When
one radical site is kept electronically constant (e.g.,
phenyl substitution at the radical site b), and a set of
triplet diradicals is considered for which dab is the same
(1,3-cyclopentanediyl triplet diradicals), the D parameter
is a sensitive probe of the electronic effects exerted by
the substituents at the radical center a through the spin
density Fa. This provides an accurate measure of radical
stabilization in terms of the efficacy of spin delocalization
in the monoradical fragments M.
TABLE 1. D Va lu es of th e Tr ip et Dir a d ica ls T a n d T′
a n d th e P M3(AUHF /CI)-Ca lcu la ted Sp in Den sities of th e
Cor r esp on d in g Mon or a d ica l F r a gm en ts M a n d M′
In view of the conspicuous lack of information in most
radical scales9-13 on the ability of triple-bonded function-
alities to delocalize an unpaired electron, we have
measured the D values of the 1,3-cyclopentanediyl triplet
diradicals T1-3 and T′1-3, generated photochemically
from the respective azoalkanes A1-3 and A′1-3 (Table
1). From the spin densities (F) accessible through the D
values, we have determined the electronic stabilization
in the propargylic versus the heteropropargylic radical
fragments. The present results demonstrate unequivo-
cally that the D parameter provides valuable data on the
efficacy of radical stabilization by the triple-bonded
functionalities -CtCH, -CtN, and -NtC.
a
The triplet diradicals T and T′ generated from the respective
azoalkanes A and A′ by irradiation with the 364-nm laser line of
the argon-ion laser for 3 min in a 2-methyltetrahydrofuran (2-
b
MTHF) matrix at 77 K. The D values were measured by EPR
spectroscopy (see Supporting Information) and are given in cm-1
,
accuracy (0.0001 cm-1 c The spin densities were computed for
.
the monoradical fragments M and M′ with the semiempirical
PM3(AUHF/CI) method. D value taken from ref 21. e The D value
d
was extrapolated from the correlation in Figure 3.
Resu lts a n d Discu ssion
The known14 azoaldehyde A was prepared in an overall
yield of 24% from the commercially available benzoyl-
acetone according to the Hu¨nig route.15 The azoaldehyde
A was employed for the synthesis of the various func-
tionalized azoalkanes A1, A3, A′2, and A′3 as sum-
marized in Scheme 1.
trichloromethyl chloroformate16 in acetonitrile. The isoni-
trile- and acetylene-functionalized azoalkanes A′2 and
A′3 were synthesized by Horner-Emmons olefination of
the azoaldehyde A with the respective known phospho-
nates in THF.17,18 The acetylene-substituted azoalkane
A3 was prepared by Peterson olefination of the azoalde-
hyde A with trimethylsilyldiazomethane at -78 °C,
followed by thermal denitrogenation on warm-up of the
Peterson product through the intermediary methylidene
carbene.
The cyano-substituted A1 azoalkane was obtained from
the corresponding oxime2a through dehydration with
(7) (a) Dougherty, D. A. In Kinetics and Spectroscopy of Carbenes
and Diradicals; Platz, M. S., Ed.; Plenum Press: New York, 1990; p
77. (b) Weil J . A.; Bolton, J . R.; Wertz, J . E. In Electronic Paramgnetic
Resonance: Elementary Theory and Practical Applications; Wiley &
Sons: New York, 1994.
(8) (a) Adam, W.; Kita, F.; Harrer, H. M.; Nau, W. M.; Zipf, R. J .
Org. Chem. 1996, 61, 7056-7065. (b) Adam, W.; Kita, F.; Nau, W. M.
Pure Appl. Chem. 1997, 69, 91-96. (c) Adam, W.; Harrer, H. M.;
Heidenfelder, T.; Kammel, T.; Kita, F.; Nau, W. M.; Sahin, C. J . Chem.
Soc., Perkin Trans. 2 1996, 2085-2090.
(9) (a) Arnold, D. R. In Diradicals; Borden, W. T., Ed.; J ohn Wiley
& Sons: New York, 1982; Chapter 11. (b) Arnold, D. R.; Nicholas, A.
M. de P.; Snow, M. S. Can. J . Chem. 1985, 63, 1150-1155.
(10) (a) J ackson, R. A. In Diradicals; Borden, W. T., Ed.; J ohn Wiley
& Sons: New York, 1982; Chapter 11. (b) Dinc¸tu¨rk, S.; J ackson, R. A.
J . Chem. Soc., Perkin Trans. 2 1981, 1127-1131. (c) Agirbas, H:,
J ackson, R. A. J . Chem. Soc., Perkin Trans. 2 1983, 739-742. (d)
J ackson, R. A.; Sharifi, M. J . Chem. Soc., Perkin Trans. 2 1996, 775-
778.
The triplet diradicals T and T′ were generated in a
2-methyltetrahydrofuran (MTHF) glass matrix at 77 K
by means of irradiation with the 364-nm line of an argon-
ion laser. In all cases, the characteristic half-field signal
(∆ms) (2) for the triplet state is located at 1650-1680
G; the relevant diradical z signals (∆ms) (1) are Bmin
)
1650 ( 60 and Bmax) 3924 ( 60 G at a microwave
frequency of 9.43 GHz. The symmetry parameter (E) of
the triplet diradical is very small and, thus, the upper
limit was estimated to be (0.0001 cm-1. All triplet
diradicals were persistent for hours at this temperature,
as evidenced by the constant EPR signals.
(11) (a) Fisher, T. H.; Meierhoefer, A. W. J . Org. Chem. 1978, 43,
220-224. (b) Fisher, T. H.; Meierhoefer, A. W. J . Org. Chem. 1978,
43, 224-228. (c) Fisher, T. H.; Derschem, S. M.; Prewitt, M. L. J . Org.
Chem. 1990, 55, 5, 1040-1043.
(12) (a) Creary, X. J . Org. Chem. 1980, 45, 280-284.(b) Creary, X.;
Mehrsheikh-Mohammadi, M. E.; McDonald, S. J . Org. Chem. 1987,
52, 3254-3263. (c) Creary, X. In Substituents Effects in Radical
Chemistry; Reidel: Dordrecht, Niederlande; NATO ASI Ser., Ser. C:
Mathematical and Physical Sciences, 1986; Vol. 189, pp 245-262.
(13) (a) J iang, X. K.; J i, G. Z.; Yu, C. X. Acta Chim. Sin. (Engl. Ed.)
1984, 82-85. (b) J iang, X. K.; J i, G. Z. J . Org. Chem. 1992, 57, 6051-
6056.
In Table 1 are listed the experimental D values of the
triplet diradicals T1-3 and T′1-3, calculated from the
EPR-spectral data, together with the theoretical spin
densities (F) of the monoradicals M and M′. The theoreti-
cal spin densities (F) were computed for the model radical
fragments M and M′ in the corresponding triplet diradi-
cals T versus T′. The geometry optimization of the model
monoradical fragments M and M′ was carried out by
using the semiempirical PM3 method with annihilated
(14) Adam, W.; Heidenfelder, T.; Sahin, C. Synthesis 1995, 1163-
1170.
(15) (a) Beck, K.; Ho¨hn, A.; Hu¨nig, S.; Prokschy, F. Chem. Ber. 1984,
117, 517-583. (b) Beck, K.; Hu¨nig, S. Chem. Ber. 1987, 120, 477-
483. (c) Beck, K.; Hu¨nig, S.; Ka¨rner, F. G.; Kraft, P.; Artschwager-
Perl., U. Chem. Ber. 1987, 120, 2041-2051.
(16) Saendya, A. Synthesis 1983, 9, 748-749.
(17) Scho¨llkopf, W. et al. Liebigs Ann. 1974, 44-53.
(18) Gibson, A. W.; Humphrey, G. R.; Kennedy, D. J .; Wright, S. H.
B. Synthesis 1991, 5, 414-416.
1008 J . Org. Chem., Vol. 68, No. 3, 2003

Spin Delocalization by Triple-Bonded Functionalities
SCHEME 1. Syn th esis of th e F u n ction a lized Azoa lk a n es A1, A′2, A3, a n d A′3 fr om th e Azoa ld eh yd e A
T′ against the aH values for the parent M and M′ radicals.
Although the EPR spectrum of the isocyanomethyl radi-
cal has been measured, its aH value could not be
determined precisely due to severe anisotropy effects,6
but from Figure 1 it may be estimated to be about 20 G.
Be this as it may, the correlation in Figure 1 conveys
clearly that the efficacy of spin delocalization is higher
for the propargyl than for the heteropropargyl radicals.
To determine quantitatively the efficiency of spin
delocalization by these triple-bonded substituents, we
assign a planar geometry to the diradicals, on the basis
of previous MO calculations (PM3 method),20 which
yielded an energy minimum for the planar 1,3-cyclopen-
tanediyl ring with coplanar substituents at the radical
sites. The planarity of the 1,3-cyclopentanediyl ring is
in line with earlier ab initio calculations for the 1,3-
cyclobutanediyl27 and the parent 1,3-cyclopentanediyl23
F IGURE 1. Correlation of the D parameter for the triplet
diradicals T and T′ with the hyperfine coupling contants (aH)
for the M and M′ radicals.
triplet diradicals.
According to semiempirical computations (AM1 method),
the rotation of the substituents about the radical centers
in the triplet diradicals T and T′ requires an appreciable
(7 kcal/mol) activation barrier on account of deconjuga-
tion.24 Therefore, the observed changes in the D values
reflect the electronic effects exerted by the substituents
on the spin delocalization in the planar triplet diradicals
T and T′. Since in all of these triplet diradicals one
radical center is kept electronically constant (phenyl
substitution), the experimentally assessed changes in the
D parameter provide a measure of the efficacy of delo-
calization by the substituent at the other radical site.
According to eq 1, the D value depends on the spin
densities (F) at this radical site and relates the experi-
mental D values of the triplet diradicals T and T′ to the
UHF wave functions, and the spin densities (Table 1)
were obtained by means of a CI calculation.19,20
The D values in Table 1 have been arranged in
decending order, i.e., in increasing degree of spin delo-
calization. Both sets of T and T′ triplet diradicals reveal
that the efficiency of radical stabilization, as displayed
by the D values and the spin densities of the M and M′
monoradicals fragments (see eq 1), follows qualitatively
the order -CtCH > -NtC ≈ -CtN. This sequence is
also displayed by the aH coupling constants of the parent
M and M′ radicals, determined by EPR spectroscopy.4,22
This is confirmed in Figure 1 by the good (r2 ) 0.950)
linear plot of the D values for the triplet diradicals T and
(19) Rauhut, G.; Chandrasekhar, J .; Alex, A.; Steinke, T.; Sauer,
W.; Beck, B.; Hutter, M.: Gedeck, P. VAMP 6.1; University of
Erlangen-Nu¨rnberg, 1996.
(23) (a) Conrad, M. P.; Pfitzer, R. M.; Schaefer, H. F. J . Am. Chem.
Soc. 1979, 101, 2245-2246. (b) Sherill, C. D.; Seidl, E. T.; Schaefer,
H. F. J . Phys. Chem. 1992, 96, 3712-3716.
(20) Stewart, J . J . P. J . Comput. Chem. 1989, 10, 209-212.
(21) Adam, W.; Emmert, O. J . Org. Chem. 2000, 65, 5664-5667.
(22) (a) Grossi, L.; Strazzi, S. J . Org. Chem. 2000, 9, 2748-2754.
(b) Beckwith, A. L. J .; Brumby, S. J . Chem. Soc., Perkin Trans. 2 1987,
1801-1808. (c) Butcher, E.; Rhodes, C. J .; Standing, M.; Davison, R.
S. J . Chem. Soc., Perkin Trans. 2 1992, 1469-1474. (d) Fujisawa, J .;
Sato, S.; Shimokoshi, K.; Shida, T. J . Phys. Chem. 1985, 89, 5481-
5486. (e) Sustmann, R.; Trill, H.; Varenholt, F.; Brandes, D. Chem.
Ber. 1977, 110, 255-263.
(24) Adam, W.; Harrer, H. M.; Kita, F.; Korth, H. G.; Nau W. M. J .
Org. Chem. 1997, 62, 1419-1426.
(25) Adam, W.; Emmert, O.; Heidenfelder, T. J . Org. Chem. 1999,
64, 3417-3421.
(26) MacInnes, I.; Walton, J . C. J . Chem. Soc., Perkin Trans. 2 1987,
1077-1082.
(27) (a) Goldberg, A. H.; Dougherty, D. A. J . Am. Chem. Soc. 1983,
105, 284-290. (b) Pranata, J .; Dougherty, D. A. J . Phys. Org. Chem.
1989, 2, 161-168.
J . Org. Chem, Vol. 68, No. 3, 2003 1009

Adam and Ortega-Schulte
F IGURE 2. Correlation between the semiempirical PM3
(AUHF/CI) spin densities of the monoradicals M/M′ and the
D values of the triplet diradicals T/T′; the value of the
isocyanide M/T set has been assessed by extrapolation (see
Table 1, value in parantheses).
F IGURE 3. Correlation of the D values for the T′ versus the
T triplet diradicals.
edly bonded isonitrile-substituted triplet diradical T′2 is
available (Table 1), with the help of the excellent cor-
relation in Figure 3 we may extrapolate a reliable D value
for the unknown directly bonded isonitrile-substituted
triplet diradical T2; this D value is given in parentheses
in Table 1. Clearly, the advantage of the correlation in
Figure 3 should be evident, since once the D value for a
particular substituent in the T′ set has been experimen-
tally determined, the unknown D value in the T set may
be obtained by extrapolation, and vice versa.
The excellent correlation in Figure 3 between the two
sets of triplet diradicals T and T′ signifies that the
electronic effects of the corresponding substituents in
delocalizing spin and thereby the radical species are
stabilized identically, except that in the T set, with the
substituent directly bonded to the radical center, the
electronic effects are more pronounced. Consequently, to
analyze the efficacy of spin delocalization by the triple-
bonded functionalities examined herein, we shall con-
sider the directly bonded T set, especially since now the
D value for the isocyanide functionality has become
available through extrapolation (Table 1, value in pa-
rentheses).
theoretical spin densities of the monoradicals M and M′,
which allows the quantitative evaluation of the observed
electronic substituent effects. Indeed, as displayed in
Figure 2, the D values of the triplet diradicals T and T′
correlate impressively well (r2 ) 0.981, n ) 15) with the
spin densities (F) of the monoradical fragments M and
M′. This confirms once again that the PM3(AUHF/CI)
semiempirical method is reliable to assess the spin
distribution in radicals. Included in this correlation are
the data for the previously reported substituents, to-
gether with the CO2Me functionality. The latter had
hitherto not been examined and was presently prepared
for comparison purposes and its D value measured (see
Supporting Information).
Closer inspection of Figure 2 reveals that the directly
bonded substituents in the triplet diradicals T (n ) 0)
and the extendedly bonded (alkenyl spacer) ones in T′
(n ) 1) fall into two well-defined sets without overlap:
The directly bonded substituents (codified by the solid
square) occupy the upper end and the extended ones
(codified by the solid circle) the lower end of the correla-
tion; note that in Table 1 the D values of the triplet
diradicals, as well as the spin densities (F) of the
monoradicals, are about 50% larger for the directly
bonded T/M compared to the extendedly bonded T′/M′
species.13 This is due to the spin dilution by the allylic
resonance in the alkenyl spacer of the T′/M′ entities.
Nonetheless, the correlation in Figure 2 covers perfectly
the directly (T) as well as the extendedly (T′) bonded sets
of triplet diradicals, which implies that the electronic
effects of these substituents in the two series are the
same and the extent of spin delocalization identical. That
this is the case is most convincingly demonstrated by the
essentially perfect straight line (r2 ) 0.995), when the D
values of the directly bonded T set are plotted against
those of the extendedly bonded T′ set (Figure 3). Again,
in this plot the previously reported derivatives of T and
T′ have been included, as well as the newly prepared
CO2Me functionality (see Supporting Information). Con-
spicuous in this plot is the omission of the isonitrile
substituent in the directly bonded T2 derivative, due to
the fact that the corresponding azoalkane A2 could not
be prepared. However, since the D value for the extend-
Figure 2 reveals that the unsaturated functionalities
follow the sequence -CO2Me < -CN ≈ -NC < -CtCH
< -Ph < -CdCH2 in increasing order of spin delocal-
ization. Evidently, the triplet-bonded substituents fall
between the least stabilizing ester group and the most
stabilizing phenyl and vinyl groups. We have shown
previously that the spin delocalization in allyl and
•
heteroallyl radicals follows the order CH2dCH-CH2
>
RNdCH-CH2• > OdCH-CH2 , for which the reluctance
of spin localization on the more electronegative heteroa-
tom is responsible. Thus, comparison with the above
sequence, which contains the triple-bonded substituents,
shows the acetylenic group and its heteroatom congeners
are all more effective in stabilizing the unpaired elec-
tron than the ester functionality, but not as effective
as the vinyl group. That in the CH2dCH-CH2 radical
the unpaired electron is better stabilized than in
HCtC-CH2 is readily explained in terms of the more
•
•
•
efficient allylic versus propargylic resonance. In the allyl
radical, the odd electron is equally distributed between
the equivalent methylene terminals through conjugation.
1010 J . Org. Chem., Vol. 68, No. 3, 2003

Spin Delocalization by Triple-Bonded Functionalities
F IGURE 4. Mesomeric structures of the heteropropargyl (M2-3) and the propargyl M3 monoradical fragments and their calculated
spin densities (F) and EPR-spectral hyperfine coupling constants (aH, in parentheses).
In contrast, in the propargyl-allenyl mesomerism (Figure
4), the unpaired electron prefers to populate the prop-
argyl rather than the allenic structure, because in the
latter the spin would have to reside at the less favored
vinyl-type radical site.5 This is also manifested by the
hyperfine coupling constants (aH) in Figure 1, which are
14.2 G for the allyl22c and 18.9 G for the propargyl4
radicals. Clearly, in the unsymmetrical propargyl radi-
cals, delocalization is less pronounced than in the sym-
metrical allylic radical, as displayed by the correlation
of Figure 2.
When in the propargyl radical the CH terminal is
replaced by a N atom, the resulting isoelectronic cya-
nomethyl radical is further destabilized, as evidenced by
the higher spin density of the M1 versus the M3
monoradical at the methylene site and the higher D
values of the corresponding triplet diradical T1 versus
T3 (Table 1). Also the aH values in Figure 1 expose this
trend, since they are 18.9 G for the propargyl4 but 20.8
G for the cyanomethyl18 radical and, hence, the spin is
not as well delocalized in the heteroatom analogues. Here
operates again the electronegativity effect, since in the
corresponding heteroatom-substituted propargyl-allenyl
resonance (Figure 4), the unpaired electron would occupy
the nitrogen terminal, which is unfavorable.2a,4,26 Thus,
the spin-delocalizing ability of the cyano group in the M1
monoradical is markedly lower than that of the acetylene
group in the M3 species (Table 1).
A remarkable case constitutes the heteropropargyl
radical M2, which results on replacement of the central
carbon atom in the propargyl radical M3 by a nitrogen
atom (Figure 4). Although the EPR spectroscopy on the
parent isocyanomethyl radical M2 has been documented,
only the nitrogen hyperfine coupling constant (aN 6.5 G)
is available, because anisotropy effects precluded the
precise determination of the hydrogen coupling constant
(aH).6 Also it should be emphasized that such heteropro-
pargyl radicals are difficult to generate by the usual
chemical means. Consequently, the photodenitrogenation
of the azoalkane A1 and A2 to afford the corresponding
triplet diradical T1 and T2, which contains the desired
heteropropargyl radical fragment M, constitutes a defi-
nite advantage of the present methodology to assess the
electronic radical-stabilizing ability of such complex
functionalities as the isocyano substituent. In fact, from
the D versus aH plot (Figure 1) we have extrapolated an
aH value of about 20 G, which is close to the estimated
value of ca. 21 G from the EPR spectrum of the parent
isocyanomethyl radical.6 Clearly, it should be apparent
from the semiempirically calculated [PM3(AUHF/CI)
method] spin densities of the radical fragments M
(numbers above the structures at the two radical termi-
nals), as well as the aH values (numbers in parentheses
below the structures) determined from the parent radi-
cals M by EPR spectroscopy (Figure 4), that the isocy-
anomethyl radical is about as effective as the cyanom-
ethyl radical in stabilizing the unpaired electron, but both
are definitely worse than the propargyl radical.
On the basis of these spin-density data, we speculate
on the radical-stabilizing ability of the isocyanomethyl
radical (Figure 4) in terms of the various mesomeric
structures. Although in all resonance structures the
unpaired electron resides on the carbon-atom terminals,
evidently neither the heteroallenic nor the iminic forms
stabilize the spin effectively. Thus, the strong electron-
withdrawing effect of the positively charged nitrogen
atom oppossess such delocalization in the heteroallenic
structure, whereas a rather electron-deficient carbon
atom is generated in the iminic form, such that stabiliza-
tion of the heteropropargyl species M2 is ineffective
compared to that of the propargyl species M3 (Table 1).
It hardly could have been anticipated that the isocyano
group in the M2 radical is about as efficient in stabilizing
an unpaired electron as a cyano group in M1.
Ack n ow led gm en t. We thank the Deutsche Fors-
chungsgemeinschaft, the Volkswagen-Stiftung, and the
Fonds der Chemischen Industrie for generous financial
support.
Su p p or tin g In for m a tion Ava ila ble: Synthetic details
and characteristic spectral data of the azoalkanes. This
material is available free of charge via the Internet at
http://pubs.acs.org.
J O026579W
J . Org. Chem, Vol. 68, No. 3, 2003 1011