Phenyl isocyanate anion radicals and their cyclotrimerization to triphenyl isocyanurate anion radicals

Article abstract of DOI:10.1021/jo4003008

Room-temperature sodium metal reduction of phenyl isocyanate (PhNCO) in hexamethylphosphoramide yields the anion radical (PhNCO^?-) where the unpaired electron exhibits coupling to one nitrogen and five unique protons. The extent of coupling to

Full text of DOI:10.1021/jo4003008

Article
pubs.acs.org/joc
Phenyl Isocyanate Anion Radicals and Their Cyclotrimerization to
Triphenyl Isocyanurate Anion Radicals
Mark A. Servos, Nathaniel C. Smart, Mark E. Kassabaum, Cody A. Scholtens, and Steven J. Peters*
Department of Chemistry, Illinois State University, Normal, Illinois 61790-4160, United States
S
* Supporting Information
ABSTRACT: Room-temperature sodium metal reduction of
phenyl isocyanate (PhNCO) in hexamethylphosphoramide
yields the anion radical (PhNCO•−) where the unpaired
electron exhibits coupling to one nitrogen and five unique
protons. The extent of coupling to the carbon in the NCO
group was obtained via the reduction of 13C-labeled PhN13CO.
Remarkably, this coupling is over 2 orders of magnitude
smaller than that found for the alkyl-substituted analogue. This
large attenuation indicates that the electron is not localized
within the isocyanate group (as in the alkyl analogues) but is
distributed throughout the entire π system including the phenyl ring. The consequence of this delocalization is that the
isocyanate is expected to remain linear upon reduction of PhNCO. The anion radicals of p-tolyl- and p-methoxyphenyl
isocyanate have also been generated. We find that these electron-donating substituents on the phenyl ring have little effect on the
nitrogen coupling. Hence, the NCO group has the same geometry as the PhNCO•−. When PhNCO is reduced in
tetrahydrofuran, a solvent where ion association effects are common, PhNCO•− is not observed. Here, a cyclotrimerization
occurs (initiated by PhNCO•−), generating the triphenyl isocyanurate anion radical where the unpaired electron is predominately
localized in one of the carbonyl moieties.
mechanism describing polymer formation involves the rapid
radical−radical coupling of two isocyanate anion radicals with
formation of a diamagnetic oxanilide dianion (Scheme 1). The
propagation reaction continues with this dianion species, as a
bidirectional polymerization occurs generating two intra-
molecular polyisocyanate chains attached to the oxanilide
moiety, Scheme 1.16,20 The high reactivity of isocyanate anion
radicals, and their propensity to rapidly polymerize, makes
exploring their structure extremely challenging.
INTRODUCTION
■
The chemistry of isocyanates (R-NCO) has been extensively
explored since their discovery in 1848 by Wurtz,1 with their
greatest use being in the production of polyurethane
materials.2,3 Moreover, the cyclotrimerization of isocyanates
leads to the formation of heterocyclic aromatic isocyanurates
(1,3,5-triazinane-2,4,6-triones), which are also commonly used
in polyurethane blends giving these materials added resil-
ience.4−6 More recently, isocyanates have been widely used in
the production of new cyclic lactam precursors as potential
antibiotics and protein inhibitors,7−11 and for the synthesis of a
variety of natural products.2,12−14 Even polyisocyanate macro-
molecules (e.g., 1-nylon) formed through poly amide bond
linkages,2,15 have been widely studied as artificial helical
polymers that mimic some aspects of protein structure.16−20
Due to the unique structural properties of 1-nylon, these
polymers have also been exploited for possible use in liquid
crystalline materials and as linkers to optical and molecular
switches.17−22
Scheme 1. Formation of Polyisocyanate Initiated by the
Coupling of Two Isocyanate Anion Radicals
Formation of these polyisocyanates is commonly performed
using an anionic initiator that attacks the electrophilic carbon of
the isocyanate forming an intermediate anion that propagates
the polymerization reaction.15−18,20 Many of these methods
have utilized the formation of an isocyanate anion radical to
initiate the polymerization.15,16,20 The one-electron reduction
of an isocyanate and subsequent polymerization is done using
alkali metals or other electron donors (such as the naphthalene
anion radical, C10H8•−) at low temperatures.15,16,20 The
Received: February 8, 2013
Published: March 31, 2013
© 2013 American Chemical Society
3908
dx.doi.org/10.1021/jo4003008 | J. Org. Chem. 2013, 78, 3908−3917

The Journal of Organic Chemistry
Article
Our group recently investigated the one-electron reduction
of both ethyl and cyclohexyl isocyanate in dilute solution under
ambient conditions in the hope of observing the respective
anion radicals, but we were unsuccessful.23 Instead, the reduced
alkyl isocyanate participates in a rapid cyclotrimerization
generating a stable trialkyl isocyanurate anion radical, reaction
1.23 The unpaired electron was found to localize in the π
system of one of the carbonyl groups in the ring.
substituted phenyl isocyanates, and the first spectroscopic
observation of their respective anion radicals. Moreover, it will
be shown that the electron spin distribution in the NCO
moiety is substantially different from that found for
MeNCO•−,24 and that solvation effects do play an important
role in the chemistry of these newly observed anion radicals.
RESULTS AND DISCUSSION
■
The sodium metal reduction of phenyl isocyanate (PhNCO, 1)
in hexamethylphosphoramide (HMPA) results in a solution
exhibiting a unique EPR spectrum, Figure 1. A nearly perfect
To the best of our knowledge, the only published finding of a
stable isocyanate anion radical is that of methyl isocyanate
(MeNCO) by Symons and Trousson; however, this was done
at exceedingly low temperatures.24 In their studies, the
MeNCO•− was generated in a dilute solid solvent matrix of
2-methylTHF or methanol at 77 K using γ-radiation to generate
the anion radical. Under these conditions, the anion radical is
trapped within the solvent matrix, unable to undergo
polymerization, and was readily observable using EPR spec-
troscopy. They concluded, from the large observed electron
coupling to the nitrogen and carbon atoms in the NCO moiety,
that the unpaired electron is localized on the carbon atom in
MeNCO•− resulting in a bent geometry for the isocyanate
group.24 Other than in this study, no spectral data have been
observed for any other isocyanate anion radical.
We are particularly interested in generating an isocyanate
anion radical for spectroscopic investigation to explore the
similarities, if any, with the structure and the electron-spin
distribution found in ketyls (e.g., RNCO•− vs R2CO•−), which
are possibly the most important and most investigated class of
anion radicals.25−40 EPR spectroscopy has been instrumental in
understanding the structure of ketyls as well as any perturbation
in electron spin density as a consequence of ion association in
solvents of varying polarity.26,27,29−36 As with MeNCO•−,
aliphatic ketyls have the majority of the electron density
residing in the carbonyl moiety resulting in a deviation from
planarity for the sp2-hybridized carbon.33 Alternatively, aryl-
substituted ketyls have the spin population throughout the
extended conjugated π-system resulting in a structure where the
ketyl moiety remains planar.40 Will the same be true for
reduced aryl isocyanates?
Figure 1. (Upper) X-band EPR spectrum recorded at 295 K after a
HMPA solution containing phenyl isocyanate (PhNCO) was exposed
to Na metal under vacuum. (Lower) Computer-generated EPR
spectrum using aN of 1.19 G and aH’s of 2.80, 2.40, 0.860, 0.710, and
0.085 G for five unique protons. The peak-to-peak line width (Δwpp)
is 0.09 G.
computer simulation reveals that the odd electron is coupled to
five nonequivalent protons and one nitrogen atom, which
strongly suggests that the signal is due to the anion radical of
phenyl isocyanate (1•−). No ion-pairing effects with the Na+
cation were observed due to strong solvation by the HMPA.41
Upon further exposure of this solution to the metal the signal
from 1•− eventually disappears and no other radical was
observed during the remainder of the reduction.
A complete picture of the electron-spin density in the
isocyanate moiety, especially at the carbon atom, is crucial to
predicting the geometry of this anion radical. This was
accomplished by reducing 13C labeled phenyl isocyanate
(PhN13CO) that was synthesized from 13C enriched benzoic
acid (Ph13COOH). The Na metal reduction was repeated using
a 1:1 mixture of PhNCO/PhN13CO in HMPA, and the
resulting EPR spectrum clearly shows the additional coupling to
the isocyanate 13C atom (a13C = 0.83 G), Figure 2. Notably, we
find that the measured nitrogen and carbon coupling constants
for 1•− are quite different from those found for MeNCO•−
The reactivity of ketyls is also well documented, and they
have been shown to dimerize under conditions that favor
strong ion association to the countercation.26−28,34,36,39
Remarkably, benzophenone ketyl (a commonly used anion
radical for the preparation of absolute THF) not only can
dimerize but also can undergo hydrogen isotope exchange with
THF-d8.38 This exchange appears to be dependent on the alkali
metal used (and therefore the degree of ion association) in the
one-electron reduction.38 We imagine that ion association is
critical to any chemistry involving isocyanate anion radicals
(such as polyisocyanate formation) and an understanding of
their structure is currently lacking.
determined by Symons and Trousson (e.g., aN = 7 G and a13C
=
110 G) where they concluded, based upon these large hyperfine
couplings, that the electron is localized on the isocyanate
carbon and that the NCO takes on a bent geometry upon
reduction. However, it is clear from our measured aN and a13C
values for 1•− that the electron is not localized in the isocyanate
Herein, we report the results describing the one-electron
reduction of phenyl isocyanate, as well as of three para-
3909
dx.doi.org/10.1021/jo4003008 | J. Org. Chem. 2013, 78, 3908−3917

The Journal of Organic Chemistry
Article
the neutral molecule where the ∠NCO bond angle remains
linear and the ∠CNC remains bent. Moreover, the rotation of
the isocyanate group with respect to the phenyl ring must be
slow on the EPR time scale causing all five protons to be
nonequivalent. Results obtained from density functional theory
(DFT) calculations on 1•− support these conclusions.
Two DFT geometry optimizations were performed for 1•−
both at the B3LYP/6-311++G** level of theory. The initial
geometry used for these anion radical calculations was that of
the B3LYP/6-31G* optimized structure for 1. For the first
calculation on 1•− the NCO moiety was kept at the same bond
angle as that found for the optimized geometry in 1 (∠NCO =
174°), while the second calculation had no constraints on the
isocyanate group. These results along with the B3LYP/6-311+
+G**/EPR-II computed Fermi contact electron−nuclear
coupling constants (shown in parentheses) are displayed in
Figure 3. Interestingly, the optimized geometry where no
constraints were used gives the more stable gas phase structure
for 1•− by 53 kJ/mol and contains a bent NCO geometry
(∠NCO = 132°). However, we find with this structure that the
calculated a13C and aN are in very poor agreement with the
experimentally determined values (Figure 3) and suggests that
the electron will localize in the isocyanate group of 1•− when
formed in the gas phase. Remarkably, by keeping the NCO
geometry the same as that obtained for the optimized neutral
molecule (e.g., nearly linear), we find that the electron spin is
more delocalized into the phenyl π system and that the
calculated carbon and nitrogen coupling constants are in good
agreement with our measured values, Figure 3. We favor this
calculated structure over the former to best resemble 1•− in
solution, which indicates that solvation effects are important to
the stability and structure of the phenyl isocyanate anion
radical.
Figure 2. (Upper) EPR spectrum recorded at 295 K of a HMPA
solution containing a 1:1 mixture of PhNCO and PhN13CO reduced
with Na metal. (Lower) Computer-generated EPR spectrum using a
1:1 mixture of two spectra representing the PhNCO•− and
PhN13CO•−. For both simulated spectra the same coupling constants
were used: aN of 1.19 G for one nitrogen and aH’s of 2.80, 2.40, 0.860,
0.710, and 0.085 G for five unique protons. An additional coupling,
a13C, of 0.83 G was included for one of the simulated spectra. The
Δwpp is 0.19 G.
group given that the 13C coupling is over 2 orders of magnitude
smaller than that found for MeNCO•−.24 This large attenuation
in spin density is caused by the extended π conjugation
between the isocyanate group and the phenyl ring which allows
for greater delocalization of the unpaired electron throughout
1•−. With substantially less total spin density now residing at
the carbon, the geometry for 1•− is likely to be similar to that of
The assignment of the measured aH’s to the five ring
hydrogens is also displayed in Figure 3 and is based upon the
magnitude of the calculated Fermi contact coupling values
where we find better agreement between the measured and
calculated aH’s for the linear isocyanate structure over that for
the bent geometry. Furthermore, these assignments are
consistent with the results obtained with other similarly
Figure 3. B3LYP/6-311++G**-predicted geometries for two structures of 1•−. The left structure was obtained by locking the NCO moiety in the
same configuration as that obtained for the B3LYP//6-31G* optimized neutral molecule (∠NCO = 174°), while in the other calculation no
restrictions were made to the NCO group (right structure). All measured coupling constants are shown with the left structure along with the 6-311+
+G**/EPR-II-computed Fermi contact coupling constants (in parentheses) for both calculated 1•− structures. The assignment of the measured aHs
to each proton was made based on the magnitude of the DFT calculated values.
3910
dx.doi.org/10.1021/jo4003008 | J. Org. Chem. 2013, 78, 3908−3917

The Journal of Organic Chemistry
Article
Figure 4. (Left) X-band EPR spectrum recorded at 295 K after a HMPA/p-tolyl isocyanate (2•−) solution was exposed to Na metal under vacuum. A
computer simulation of this spectrum (lower) was generated using aN of 1.16 G and aH’s of 2.74 G (1H), 2.45 G (1H), 0.842 G (1H), 0.075 G (1H),
and 0.73 G (3H). The Δwpp is 0.12 G. (Right) EPR spectrum recorded at 295 K after reducing p-methoxy phenyl isocyanate in HMPA with Na
metal. The computer simulation (lower) was generated using an aN of 1.14 G and aH’s of 2.78 G (1H), 2.51 G (1H), 0.851 G (1H). The Δwpp is
0.10 G. Structures shown include the B3LYP//6-311++G**/EPR-II-computed Fermi contact electron−nuclear coupling constants (shown in
parentheses).
substituted phenyl anion radicals where the substituent has
electron withdrawing characteristics.40a,42 The anion radicals of
nitrosobenzene and benzaldehyde exhibit larger spin densities
in the ortho and para positions relative to those found for the
meta positions. We see in the case of 1•− that a similar result is
obtained where one of the ortho and the para hydrogens have
the largest aHs. Furthermore, a significant difference in the
measured aHs for the two ortho and the two meta hydrogens has
been observed for the nitrosobenzene and benzaldehyde anion
radicals.40a,42 In the case of the latter this difference is 1.29 G
(ortho positions) and 0.56 G (meta positions).40a A similar
effect caused by the asymmetric isocyanate group would
account for the differences in the observed couplings found in
1•− (Figure 3). Interestingly, we conclude from the different
proton couplings that the amount of spin density at these
positions is significantly smaller than that found for the
benzaldehyde ketyl,40a and this would suggest that the NCO is
a much weaker electron-withdrawing group.
Placement of a methyl group or methoxy group at the para
position on phenyl isocyanate appears to have a negligible effect
on the nitrogen hyperfine coupling upon reduction; however,
significant redistribution of spin density in the ring is evident by
the change in coupling to the remaining four protons. We find
with the anion radicals of p-tolyl isocyanate (2•−) and p-
methoxyphenyl isocyanate (3•−) that the presence of these
electron-pushing substituents attenuates the nitrogen coupling
only slightly from that of the parent 1•−. The hyperfine
couplings to the four ring protons are very similar for both
anion radicals. The assignment of each aH to the different
protons was made using the same computational procedure
described for 1•− (see Figure 4) even though the calculated aHs
are in poor agreement (especially for 3•−) with the
experimental values suggesting that solvation effects are
important to the distribution of electron spin (and structure)
within these anion radicals. In both systems we see a significant
augmentation in the aH’s for the hydrogens that are meta to the
isocyanate while coupling to one of the ortho protons has
lessened considerably.
One-electron reduction experiments were also performed
with p-nitrophenyl isocyanate, and it was found that the
presence of the strongly electron-withdrawing nitro substituent
draws much of the electron density away from the isocyanate
group resulting in an EPR signal that is similar to those of other
para-substituted nitrobenzene anion radicals.43 The coupling
constants obtained for the p-nitrophenyl isocyanate anion
radical are: aN = 7.78 G (1 N), aN = 0.10 G (1 N), aH = 3.40 G
(2 H), and aH = 1.12 G (2 Hs). The large observed nitrogen
coupling, as well as the triplet of triplets due to the proton
couplings, is consistent with these anion radicals; see the
Supporting Information for EPR spectrum and simulation.
Symons and Trousson also investigated the structure of the
methyl isothiocyanate anion radical (MeNCS•−) under similar
experimental conditions as that for MeNCO•− (e.g., low
temperature in a solid solvent matrix) and obtained similar
spectroscopic results indicating that the electron is localized in
the p orbital of the isothiocyanate carbon atom.24 Our attempts
to generate PhNCS•− in HMPA were unsuccessful and would
3911
dx.doi.org/10.1021/jo4003008 | J. Org. Chem. 2013, 78, 3908−3917

The Journal of Organic Chemistry
Article
suggest that this anion radical is more susceptible to
polymerization than the isocyanate analogue.
Reduction of Phenyl Isocyanate in Tetrahydrofuran.
When the reduction of PhNCO is performed in THF, a solvent
where ion association effects are common with anion radicals,44
the results are quite different than those obtained in HMPA. As
described above, only 1•− was detected when reduction
experiments were carried out in HMPA. This was somewhat
surprising since the one electron reduction of alkyl isocyanates
in HMPA resulted in the formation of isocyanurate anion
radicals (see reaction 1).23 However, when the one-electron
reduction of 1 is carried out in tetrahydrofuran (THF) the
resulting EPR spectrum reveals that a stable isocyanurate anion
radical is formed.
The potassium metal reduction of 1 in THF gives a solution
that exhibits a strong EPR signal with a 1:2:3:2:1 pentet from
the unpaired electron coupling to two equivalent nitrogen
atoms (aN = 1.37 G), Figure 5. No signal from 1•− was detected
Figure 6. (Upper) EPR spectrum recorded at 295 K after a THF
solution containing Ph(2H5)NCO was exposed to K metal. (Lower)
Computer simulation using aN = 1.37 G (2 Ns) and aN = 0.095 G (1
N) and Δwpp = 0.12 G.
Upon closer inspection of the spectrum shown in Figure 6
we find that the pentet exhibits further splitting due to a third
nitrogen (aN = 0.095 G). This splitting is not from any
2
deuterium atoms since a H ≅ 0.012 G (aH x 0.154), and
2
therefore any coupling (even to six Hs) would be lost in the
spectral line width. As further evidence that an isocyanurate
anion radical has formed in solution, the reduction of PhN13CO
was also performed and the resulting spectrum clearly exhibits
additional couplings to only three carbons with a13C = 1.52 G
(1 × 13C) and a13C = 0.835 G (2 × 13C), Figure 7. The
observed coupling to 3 N’s and 3 13C’s, as well as the
Figure 5. (Upper) EPR spectrum recorded at 295 K after a THF
solution containing phenyl isocyanate was exposed to K metal under
vacuum. (Lower) Computer-generated EPR spectrum using aN of 1.37
G (2 Ns) and aH of 0.080 G (6 Hs) with Δwpp = 0.14 G.
under these conditions, which suggests that this anion radical is
much more reactive in solvents of lower polarity.45 In
generating a computer simulation of this spectrum it was
necessary to include a small unresolved coupling from
hydrogens in two of the phenyl rings (aH = 0.080 G, 6 Hs),
Figure 5. Six hydrogens were used in the simulation since the
majority of spin density in the rings attached to the two Ns is
expected to reside at the ortho and para positions.40a,42 Further
proof of this unresolved proton splitting was obtained via the
reduction of penta-deuteriophenyl isocyanate (Ph(2H5)NCO,
1
2
1-D5). By replacing the Hs on the phenyl rings with Hs, a
noticeable decrease in the apparent peak width should occur
2
since the ratio of gyromagnetic ratios is, γ H/γH = 0.154, and
2
therefore, any coupling to Hs will be lost in the spectral line
width.46 Figure 6 contains the EPR spectrum obtained from the
one-electron reduction of 1-D5 using experimental conditions
identical to those used with 1; as expected, the spectral peak
width is much narrower when compared to the EPR spectrum
in Figure 5. Clearly these results show that very little spin
density is distributed into the phenyl rings of the anion radical
formed.
Figure 7. (Upper) EPR spectrum recorded at 295 K after a THF
solution containing PhN13CO was exposed to K metal. (Lower)
Computer-generated EPR simulation using aN = 1.38 G (2 Ns), a13C
=
1.52 G (1 13C), a13C = 0.835 G (2 13Cs) and aH = 0.10 G (6 Hs), Δwpp
= 0.15 G.
3912
dx.doi.org/10.1021/jo4003008 | J. Org. Chem. 2013, 78, 3908−3917

The Journal of Organic Chemistry
Article
Scheme 2. Potassium Metal Reduction of a THF/PhNCO Solution Resulting in a Rapid Cyclotrimerization and the Formation
of the Triphenyl Isocyanurate Anion Radical
symmetrical nature of these couplings, leaves little doubt that
the anion radical was generated from a cyclotrimerization of
three PhN13CO molecules. The electron-spin distribution
within this trimer is nearly identical to that observed in the
reduction of trialkyl isocyanurates,23 which also strongly
supports the formation of the triphenyl isocyanurate anion
radical (4•−) upon reduction of 1 in THF. Moreover, we
envision that a cyclotrimerization mechanism, initiated by 1•−,
drives the rapid formation of the trimer anion radical, Scheme
2. Finally, upon oxidation of the reduced THF solution with
1
iodine, H and 13C{1H} NMR spectroscopy were used to
determine that 4 was the major species (∼ 93%) found in
Figure 8. B3LYP/6-31G*-predicted geometries for 4 and 4•−
.
solution. As was observed with the trialkyl isocyanurate anion
radicals,23 the electron spin in 4•− is predominately localized
within the π system of one of the three carbonyls.
Structure 4 has D3h symmetry where the plane of the phenyl rings is
orthogonal to the planar isocyanurate ring. In the case of 4•− the
isocyanurate is no longer flat and the carbonyl with high electron
density is slightly pyramidalized. Note that the phenyl rings adjacent to
this carbonyl have rotated to become somewhat coplanar with the
isocyanurate moiety. Also shown is the shape of the π-SOMO surface
for 4•−, which supports the experimental result that the electron is
localized in one carbonyl.
B3LYP/6-31G* calculations on both 4 and 4•− suggest that
the neutral molecule will undergo a significant structural
rearrangement once an electron is added to the π system. The
structure of 4 has D3h symmetry where the phenyl rings are
orthogonal to the planar isocyanurate ring, Figure 8. These
calculations also indicate that the isocyanurate ring in 4 must
undergo a first order Jahn−Teller (J−T) distortion and lose of
planarity upon electron attachment. This distortion reduces the
symmetry of the ring in 4•− which now has a nonplanar
geometry (Cs symmetry) where the carbonyl with high spin
density is now slightly pyramidalized. The decrease in
symmetry upon addition of an electron means that two or
more deformed structures (or J−T species) must be present
which can undergo rapid exhange.47 In the EPR data, we
observe only one J−T species in solution, even at low
temperatures, but this may be the result of a dynamic J−T
effect between the different conformers that interchange rapidly
on the EPR time scale.47
rate ring could be readily found. The lowest energy conformer
of the trimethyl isocyanurate anion radical has nearly the same
geometry, π-SOMO and spin distribution for the isocyanurate
ring as that found for 4•−. Two other possible conformers were
found and their structures are shown in Supporting
Information.
We see from the optimized structure for 4•− (Figure 8) that
two of the phenyl rings have become almost coplanar with that
of the isocyanurate ring. This structure allows for increased π
orbital overlap between the rings as well as increased spin
density into the two phenyl rings. On the whole, the shape of
the π-SOMO surface for 4•− conforms remarkable well to the
magnitude of the coupling constants (and therefore the
electron spin density) observed for the carbons and nitrogens
in the isocyanurate ring.
We performed a conformational search of other possible J−T
species involving the isocyanurate ring. These calculations were
performed using the less structurally complex trimethyl
isocyanurate anion radical where conformers of the isocyanu-
3913
dx.doi.org/10.1021/jo4003008 | J. Org. Chem. 2013, 78, 3908−3917

The Journal of Organic Chemistry
Article
Attempts were made to reduce 4 directly in THF; however,
no anion radical was detected via EPR, and NMR analysis
revealed that 4 was still present in solution. It was apparent in
these experiments that the metal became covered by a thin-film
residue and that this film likely inactivated the metal surface.
Even exposure to solvated electron in THF (generated with 18-
crown-6 ether) did not form 4•−; instead, complete
decomposition of 4 occurs as evident in the NMR data.
Further experimental techniques are currently being explored
(e.g., cyclic voltammetry) to attempt to reduce 4, as well as
other aryl-substituted isocyanurate compounds. However, we
surmise that the reaction pathway described in Scheme 2
circumvents this difficulty, which has made it possible to obtain
4•− in solution starting with 1•−.
In the case of 4•−, an increase in electron distribution
throughout the ring allows for more spin density to reside on
the remaining nitrogen and enhances our ability to observe the
small hyperfine coupling (aN = 0.095 G), Figure 6.
Tight ion pairing is clearly evident with some isocyanurate
anion radicals which exhibit coupling to the metal cation.
Notably, the degree of ion pairing has a noticeable influence on
the hyperfine coupling to the nitrogen and carbon atoms in
these systems. In the case of the triethyl isocyanurate anion
radical, 5•− (Figure 9), a tight ion pair exists in solution even
CONCLUSION
■
The reduction of phenyl isocyanate, under conditions where
ion association is typically absent (e.g., in HMPA), leads to the
formation of the respective anion radical where the electron is
not localized on the carbon in the isocyanate group but is
distributed throughout the π system. The magnitude of the
carbon- and nitrogen-coupling constants as well as DFT
calculations reveal that the isocyanate group remains linear in
these anion radicals (as in the neutral molecule). Alternatively,
when the reduction is carried out under conditions where ion
association is likely to occur, no isocyanate anion radical is
observed. Instead, we find that a rapid cyclotrimerization,
initiated by the isocyanate anion radical, takes place resulting in
the formation of an isocyanurate anion radical. Finally, ion
pairing effects result in a noticeable perturbation in the
electron-spin distribution throughout the isocyanurate ring.
This methodology for reducing isocyanates represents a
convenient way to study the chemistry of these anion radicals
and to see how different substituents attached to the isocyanate
moiety influence the reduction chemistry.
Figure 9. (Upper) EPR spectrum of a THF solution containing 1.5:1
mixture of EtNCO/18-crown-6 ether after exposure to K metal at 295
K. (Lower) Computer-generated EPR spectrum using aN = 1.54 G (2
N) and aK = 0.10 G (1 K), Δwpp = 0.10 G.
EXPERIMENTAL SECTION
■
Materials. 13C-Benzoic acid (C6H513COOH) and perdeuteriated
benzoic acid (Ph(2H5)COOH) were purchased from Cambridge
Isotopes, Inc., and used in the synthesis of the respective isocyanates.
Synthesis of 13C-Phenyl Isocyanate (PhN13CO) and Perdeu-
teriated Phenyl Isocyanate (Ph(2H5)NCO). A mixture of 8.2 mmol
of Ph13COOH (or Ph(2H5)COOH) and 10.4 mmol of SOCl2 was
dissolved in 50 mL of CH2Cl2. The solution was refluxed for 8 h under
an argon atmosphere after which the CH2Cl2 was distilled off, leaving
the crude benzoyl chloride (Ph13COCl) behind. To the flask
containing Ph13COCl were added 30 mL of benzene and 50 mL of
an aqueous solution containing a large excess of NaN3 (34 mmol).
This solution was stirred vigorously to maximize the mixing of both
layers. Aliquots of the benzene solution were removed and
interrogated using GCMS to monitor the complete conversion of
Ph13COCl to the acyl azide (Ph13CON3). The conversion took
approximately 24 h. After this time, the benzene layer was separated
from the aqueous layer and dried with anhydrous MgSO4. The filtered
benzene solution containing Ph13CON3 was refluxed for 6 h under
argon to drive the conversion of the acyl azide to the isocyanate via a
Curtius rearrangement. The benzene was then removed and the crude
product (PhN13CO) was distilled under reduced pressure (18 Torr) at
71 °C. 13C{1H} NMR spectral analysis confirm the formation of the
isocyanate, Figure 10. The same procedure was used to synthesize
when the K+ is encapsulated in an 18-crown-6 ether. (The EPR
spectrum exhibits hyperfine coupling to the K+ nucleus (I = 3/
2), aK = 0.10 G.) This tight (or contact) ion pair likely involves
the oxygen that has the majority of charge density within the
isocyanurate ring, see Figure 9. Alternatively, no metal splitting
is apparent for the anion radical of 4•− (see Figures 3 and 4),
and therefore a loose or solvent separated ion pair must exist
where it is likely the phenyl rings are sterically inhibiting the K+
from forming a contact ion pair with the negatively charged
carbonyl group. These differences in ion association result in a
measurable difference in the nitrogen and carbon hyperfine
coupling constants in 4•− and 5•−. Although the electron is
predominately localized in one carbonyl for both anion radicals,
we find that the contact ion pair formed with 5•− has more
electron density in the carbonyl, which is apparent when
comparing the nitrogen and carbon hyperfine couplings with
those found for the loose ion pair (4•−), see Structures 4•− and
5•−. (Supporting Information contains the EPR spectrum of
13C labeled 5•− along with the computer generated simulation).
3914
dx.doi.org/10.1021/jo4003008 | J. Org. Chem. 2013, 78, 3908−3917

The Journal of Organic Chemistry
Article
Figure 10. 125 MHz 13C{1H} NMR spectrum of PhN13CO in CDCl3. Note that the ipso carbon resonance is split by coupling to the isocyanate
carbon (2JCC = 6.3 Hz).
Figure 11. 125 MHz 13C{1H} NMR spectrum of synthesized Ph(2H5)NCO in CDCl3. Note the carbon resonances that are split into a triplet by the
deuterium (I = 1) attached.
Ph(2H5)NCO and the 13C{1H} NMR spectrum is displayed in Figure
11.
vacuum system directly into bulb E, and the evacuated apparatus was
sealed from the vacuum line at point F. The glass tube containing the
isocyanate was broken by shaking, and the solution was well mixed
before brief exposure to the metal mirror.
The EPR spectrum was immediately recorded after pouring a
sample of the solution into the 3 mm EPR tube. The apparatus was
removed from the EPR instrument and the solution exposed to more
metal and the spectrum recorded again. This process was continued
until the best signal-to-noise ratio for the anion radical was obtained.
All EPR data was collected on a Bruker EMX-080 spectrometer
equipped with a variable-temperature unit.
Reduction of Phenyl Isocyanate Compounds in HMPA and
THF. A sealed glass tube (with fragile ends) was charged with 0.15
mmol of phenyl isocyanate and placed into bulb E of the Pyrex glass
apparatus shown in Figure 12. A small amount of either sodium or
potassium metal was placed into bulb B, which was then sealed at
point A, and the entire apparatus was evacuated. The alkali metal was
distilled into bulb D to form a pristine metal mirror, after this, bulb B
was sealed at point C. Three milliliters of HMPA (dried with
potassium metal) or THF (dried over NaK) were distilled from the
3915
dx.doi.org/10.1021/jo4003008 | J. Org. Chem. 2013, 78, 3908−3917

The Journal of Organic Chemistry
Article
(9) Rigby, J. H.; Brouet, J.; Burke, P. J.; Rohach, S.; Sidique, S.; Heeg,
M. J. Org. Lett. 2006, 8, 3121−3123.
(10) Roberson, C. W.; Woerpel, K. A. J. Am. Chem. Soc. 2002, 124,
11342−11348.
(11) Barrett, A. G. M.; Betts, M. J.; Fenwick, A. J. Org. Chem. 1985,
50, 169−175.
(12) Rigby, J. H.; Shyama, S. Org. Lett. 2007, 9, 1219−1221.
(13) Lei, A.; Lu, X. Org. Lett. 2000, 2, 2357−2360.
(14) Brandi, A.; Cicchi, S.; Cordero, F. M. Chem. Rev. 2008, 108,
3988−4035.
(15) Shashoua, V. E.; Sweeny, W.; Tietz, R. F. J. Am. Chem. Soc.
1960, 82, 866−873.
(16) Bur, A. J.; Fetters, L. J. Chem. Rev. 1976, 76, 727−746.
(17) Green, M. M.; Peterson, N. C.; Sato, T.; Teramoto, A.; Cook,
R.; Lifson, S. Science 1995, 268, 1860−1866.
(18) Green, M. M.; Park, J.-W.; Sato, T.; Teramoto, A.; Lifson, S.;
Selinger, R. L. B.; Selinger, J. V. Angew. Chem., Int. Ed. 1999, 38, 3138−
3154.
(19) Satoh, T.; Ihara, R.; Kawato, D.; Nishikawa, N.; Suemasa, D.;
Kondo, Y.; Fuchise, K.; Sakai, R.; Kakuchi, T. Macromolecules 2012, 45,
3677−3686.
(20) Shah, P. N.; Min, J.; Chae, C.; Nishikawa, N.; Suemasa, D.;
Kakuchi, T.; Satoh, T.; Lee, J. Macromolecules 2012, 45, 8961−8969.
(21) Zhao, W.; Kloczkowski, A.; Mark, J. E.; Erman, B.; Bahar, I.
Macromolecules 1996, 29, 2796−2804.
(22) Pijper, D.; Jongejan, M. G. M.; Meetsma, A.; Feringa, B. L. J.
Am. Chem. Soc. 2008, 130, 4541−4552.
(23) Peters, S. J.; Klen, J. R.; Smart, N. S. Org. Lett. 2008, 10, 4521−
4524.
(24) Symons, M. C.; Trousson, P. M. Radiat. Phys. Chem. 1984, 23,
Figure 12. Apparatus used for the generation of the isocyanate and
isocyanurate anion radicals in HMPA or THF.
127−135.
(25) Lown, J. W. Can. J. Chem. 1965, 43, 2571−2575.
(26) Evans, A. G.; Evans, J. C.; Hodden, E. H. J. Chem. Soc. B 1969,
546−547.
ASSOCIATED CONTENT
* Supporting Information
Additional EPR spectral data and DFT calculation results. This
material is available free of charge via the Internet at http://
pubs.acs.org.
■
(27) Chen, K. S.; Mao, S. W.; Nakamura, K.; Hirota, N. J. Am. Chem.
Soc. 1971, 93, 6004−6013.
S
(28) Staples, T. L.; Szwarc, M. J. Am. Chem. Soc. 1970, 92, 5022−
5027.
(29) Russell, G. A.; Wallraff, G.; Gerlock, J. L. J. Phys. Chem. 1978,
82, 1161−1168.
AUTHOR INFORMATION
Corresponding Author
*E-mail: sjpeter@illinoisstate.edu.
Notes
(30) Nakamura, K. J. Am. Chem. Soc. 1981, 103, 6973−6974.
(31) Echegoyen, L.; Neives, I.; Stevenson, C. D. J. Phys. Chem. 1982,
82, 1611−1614.
■
(32) Hirayama, M.; Ohhata, H. Bull. Chem. Soc. Jpn. 1987, 60, 2751−
2756.
The authors declare no competing financial interest.
(33) Davies, A. G.; Neville, A. G. J. Chem. Soc., Perkin Trans. 2 1992,
163−169.
ACKNOWLEDGMENTS
(34) Candida, M.; Shohoji, B. L. Tetrahedron Lett. 1995, 36, 6167−
6170.
■
We thank the American Chemical Society-Petroleum Research
Fund (ACS-PRF 51677-UR4) and Illinois State University for
funding of this work.
(35) Stevenson, C. D.; Reiter, R. C.; Burton, R. D.; Halvorsen, T. D.
Inorg. Chem. 1995, 34, 1368−1372.
(36) Scholz, M.; Gescheidt, G.; Daub, J. J. Chem. Soc., Chem.
Commun. 1995, 803−804.
REFERENCES
■
(37) Micha-Screttas, M.; Heropoulos, G. A.; Steele, B. R. J. Chem.
Soc., Perkin Trans. 2 1999, 2685−2690.
(1) Wurtz, A. Compt. Rend. 1848, 27, 241−243.
(2) Ulrich, H. Chemistry and Technology of Isocyanates; John Wiley
and Sons: New York, 1996.
(3) Delebecq, E.; Pascault, J.; Boutevin, B.; Ganachaud, F. Chem. Rev.
2013, 113, 80−118.
(4) Raffel, B.; Loevenich, C. J. J. Cell. Plast. 2006, 42, 17−47.
(5) Loevenich, C. J.; Raffel, B. J. Cell. Plast. 2006, 42, 289−305.
(6) Zitinkina, A. K.; Sibanova, N. A.; Tarakonov, O. G. Russ. Chem.
Rev. 1985, 54, 1866−1898.
(7) Huard, K.; Bagley, S. W.; Menhaji-Klotz, E.; Preville, C.;
Southers, J. A.; Smith, A. C.; Edmonds, D. J.; Lucas, J. C.; Dunn, M. F.;
Allanson, N. M.; Blaney, E. L.; Garcia-Irizarry, C. N.; Kohrt, J. T.;
Griffith, D. A.; Dow, R. L. J. Org. Chem. 2012, 77, 10050−10057.
(8) Fisher, J. F.; Meroueh, S. O.; Mobashery, S. Chem. Rev. 2005,
105, 395−424.
(38) Kamaura, M.; Hanamoto, T.; Kuwatani, Y.; Inanaga, J. J. Am.
Chem. Soc. 1999, 121, 6320−6321.
(39) Macias-Ruvalcaba, N. A.; Felton, G. A. N.; Evans, D. H. J. Phys.
Chem. C 2009, 113, 338−345.
(40) (a) Steinberger, N.; Fraenkel, G. K. J. Chem. Phys. 1964, 40,
723−729. (b) Rieger, P. H.; Fraenkel, G. K. J. Chem. Phys. 1962, 37,
2811−2832.
(41) Levin, G.; Jagur-Grodzinski, J.; Szwarc, M. J. Am. Chem. Soc.
1970, 92, 2268−2275. We did find ion pairing effects with the alkyl
isocyanurate anion radicals and the Na+ cation in our previous studies
involving the reduction of alkyl isocyanates; see ref 23.
(42) Ayscough, P. B.; Sargent, F. B.; Wilson, R. J. Chem. Soc. B 1966,
903−906.
3916
dx.doi.org/10.1021/jo4003008 | J. Org. Chem. 2013, 78, 3908−3917

The Journal of Organic Chemistry
Article
(43) Stevenson, C. D.; Echegoyen, L. J. Phys. Chem. 1973, 77, 2339−
2342.
(44) (a) Ayscough, P. B.; Wilson, R. Proc. Chem. Soc. 1962, 229−230.
(b) Canters, G. W.; De Boer, E. Mol. Phys. 1973, 26, 1185−1198.
(c) Peters, S. J.; Turk, M. R.; Kiesewetter, M. K.; Reiter, R. C.;
Stevenson, C. D. J. Am. Chem. Soc. 2003, 125, 11212−11213.
(45) The same EPR results were obtained when p-tolyl isocyanate
and p-methoxy isocyanate were reduced under these conditions.
(46) Wertz, J. E.; Bolton, J. R. Electron Spin Resonance Elementary
Theory and Practical Applications; Chapman and Hall: New York, 1986.
(47) For a discussion and examples of systems that undergo J−-T
distortion upon electron addition, see: (a) Jahn, H. A.; Teller, E. Proc.
R. Soc. London Ser. A 1937, 161, 220. (b) Horvat, D. A.; Hammons, J.
H.; Stevenson, C. D.; Borden, W. T. J. Am. Chem. Soc. 1997, 119,
9523−9526. (c) Kurth, T. L.; Brown, E. C.; Hattan, C. M.; Reiter, R.
C.; Stevenson, C. D. J. Phys. Chem. A 2002, 106, 478−481. (d) Gerson,
F.; Huber, W. Electron Spin Resonance Spectroscopy of Organic Radicals;
Wiley-VCH: Weinheim, 2003; p 161.
3917
dx.doi.org/10.1021/jo4003008 | J. Org. Chem. 2013, 78, 3908−3917