Dimerization and trapping of diazirinyl radicals

Article abstract of DOI:10.1039/b310148g

Computational and experimental methods have been utilized to examine the facile dimerization of diazirinyl radicals. Two potential dimers were investigated using density functional theory. Both were shown to have low-barrier reaction coordinates leading t

Full text of DOI:10.1039/b310148g

View Article Online / Journal Homepage / Table of Contents for this issue
R E S E A R C H P A P E R
Dimerization and trapping of diazirinyl radicals
Robert A. Thompson, Joseph S. Francisco and John B. Grutzner*
Department of Chemistry, Purdue University, 560 Oval Drive, West Lafayette,
Indiana 47907-2084, USA. E-mail: grutzner@purdue.edu
Received 22nd August 2003, Accepted1st December 2003
F|rst published as an AdvanceArticle on the web15th January 2004
Computational and experimental methods have been utilized to examine the facile dimerization of diazirinyl
radicals. Two potential dimers were investigated using density functional theory. Both were shown to have
low-barrier reaction coordinates leading to formation of HCN and N2 . A cross-over experiment was employed
to establish the relative contributions of C–N and N–N dimers towards product formation. The N–N
dimerization was found to be dominant under the current conditions, and the corresponding reaction
coordinate energetics were further explored using coupled-cluster methods. A detailed mechanism for dimer
decomposition is presented. The competition between unimolecular and bimolecular reactions of diazirinyl
radicals is explored under high dilution conditions. The high reactivity of the diazirinyl radical as a nitrogen
atom transfer agent suggests a possible bimolecular contribution for ‘‘prompt’’ NO formation in hydrocarbon
combustion with diazirinyl radicals as intermediates.
In the current work we have examined two possible diaziri-
Introduction
nyl radical dimers, the N–N and C–N dimers, both experimen-
tally and using density functional theory calculations. In
accord with simple bond energy concepts, the C–N dimer
was found to be of lower energy. A mixture of 3-aryldiazirinyl
radicals were thermally generated in solution, and the outcome
of the label distribution determined through this cross-over
experiment to determine the relative contributions of each
dimer towards product formation. The results are discussed
in terms of a single dimeric species responsible for nitrile for-
mation, and the corresponding reaction coordinate was further
Diazirinyl radicals were initially reported in the ground-
breaking work of Ingold and Maeda.1 Various 3-alkyl and
3-aryldiazirinyl radicals were generated by UV-photolysis of
the parent 3-halodiazirine in the presence of hexa-n-butylditin.
EPR characterization revealed that the observed radicals are
localized on the diazirine ring with essentially all of the
unpaired spin density located on the two nitrogens regardless
of the 3-substituent, including phenyl. These observations were
further supported by INDO molecular orbital calculations
performed on the 3-H and 3-methyl analogues. They also
reported the decay kinetics for 3-methyl, 3-ethyl and 3-phenyl-
diazirinyl radicals measured by kinetic EPR spectroscopy at
ꢀ75 ꢁC. In each case, decay occurred with second-order
kinetics at rates approaching the diffusion-controlled limit
(for 3-phenyl, kE2PR ¼ 2.0 ꢂ 108 Mꢀ1 sꢀ1) with the only products
being the corresponding nitrile and, presumably, nitrogen. They
attributed product formation to an initial head-to-head (N–N)
coupling of diazirinyl radicals (Scheme 1). It is also noteworthy
that in no case was the tetrazine observed.1
investigated using coupled-cluster methods.
A detailed
mechanism for dimer decomposition is presented, and the
competition between unimolecular and bimolecular decompo-
sition of diazirinyl radicals is explored with reactions at high
dilution. A possible bimolecular contribution is considered
for the formation of ‘‘prompt’’ NO in hydrocarbon combus-
tion from diazirinyl intermediates.2–4
Ingold and Maeda were also able to generate 3-substituted
diazirinyl radicals thermally by the following route:1
Results and discussion
Theoretical exploration of possible diazirinyl radical dimers and
their fate
D
ðCH3Þ3CONNOCðCH3Þ3
2ðCH Þ COꢃ þ N
ð1Þ
ƒƒƒ!
3
2
3
Three possible c-2HCN2 dimers exist: a nitrogen–nitrogen
dimer, 1, a nitrogen–carbon dimer, 2, and a carbon–carbon
dimer, 3 (Scheme 2). However, only 1 and 2 could directly
result in nitrile formation.
ðCH3Þ COꢃ þðCH3CH2Þ SiH!
3
3
ðCH3Þ COHþðCH3CH2Þ Siꢃ ð2Þ
3
3
ðCH3CH2Þ Siꢃ þRCðBrÞN2 !ðCH3CH2Þ SiBrþc-RCN2
3
3
Therefore, 1 and 2 were optimized at the B3LYP/6-
31G(d,p) level of theory. The geometry and relative energy
(with respect to 1) for both dimers are shown in Fig. 1. Each
was verified to be a rotational minimum (with respect to the
N2–N4 bond of 1 and the C5–N1 bond of 2) by constraining
the torsion angle in 30ꢁ increments through a total of 360ꢁ
while re-optimizing at each point. The corresponding opti-
mized coordinate values are shown in Table 1. 2 was calculated
to be 128 kJ molꢀ1 lower in energy than 1. Therefore, even
though formation of 1 is favored by high spin density residing
ð3Þ
on each nitrogen (0.61/N, ꢀ0.22 for Ca of A2 c-2HCN2), 2 is
2
significantly more stable thermodynamically. The very long
˚
nitrogen–nitrogen bonds of each dimer (1.67 A for N2–N3
Scheme 1
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
756
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 7 5 6 – 7 6 5

View Article Online
Scheme 2
and N4–N5 of 1, 1.74 A N1–N2 of 2) in comparison to 2A2
theory (see Table 1), and subsequent single-point calculations
˚
2
c- HCN2 (1.58 A) are noteworthy.
˚
were performed on the optimized geometries using the 6-
311þþG(d,p), 6-311þþG(2d,2p), and 6-311þþG(2df,2p)
basis sets. The results are shown in Table 2. At the highest level
of theory used in the present work, the activation energy for
decomposition of 1 to 2 HCN and N2 via 1-TS was calculated
The conversion to HCN and N2 was explored as well as the
location of any transition states along the reaction coordinate
for decomposition of 1 and 2 that could result in product for-
mation. Transition states were located, and each was verified
to be a rotational minimum (see above) and proven to arise
from the corresponding dimer by performing intrinsic reaction
coordinate (IRC) calculations in the reverse direction. The
geometries and relative energies (with respect to 1) for both
transition states (1-TS and 2-TS) are shown in Fig. 1. The cor-
responding optimized coordinate values are shown in Table 1.
1-TS is only 12.1 kJ molꢀ1 higher in energy than 1. The coor-
dinate values of 1-TS reveal the following: further stretching of
to be only þ10.9 kJ molꢀ1
.
In summary, calculations performed at the B3LYP/6-
31G(d,p) level of theory have shown that dimerization of
c-2HCN2 can lead to 1 or 2, both of which have low barrier
transition states consistent with the observed products. How-
ever, the conversion of 2 to products was found to be indirect.
In order to experimentally determine the relative contributions
of each dimer towards product formation, a cross-over experi-
ment was designed based on the thermal generation of isotopi-
cally and substitutionally mixed pairs of diazirinyl radicals in
solution.
˚
the N2–N3 and N4–N5 bonds (1.67 to 1.83 A), lengthening of
˚
the C1–N2 and C6–N4 bonds (1.40 to 1.42 A) and contraction
˚
of the C1–N3 , C6–N5 (1.25 to 1.24 A) and N2–N4 bonds (1.44
˚
to 1.33 A), all of which are consistent with formation of HCN
and N2 .
Furthermore, 2-TS is only 36.8 kJ molꢀ1 higher in energy
than 2. The coordinate values of 2-TS reveal the following:
Synthesis of 3-bromo-3-aryldiazirines
15N-Benzamidine and p-toluamidine hydrochloride hydrates
were generated from the corresponding nitriles by the method
of Schaefer and Peters.5 3-Bromo-3-aryldiazirines 4, 5 and 6
were prepared from the corresponding amidine hydrochlorides
according to the standard Graham oxidation (Scheme 3).6
The extent of 15N isotopic incorporation into the diazirine
was determined by 13C NMR spectroscopy. An expansion of
the 125 MHz 13C NMR spectrum around the 3-C position
of 5 is shown in Fig. 3(B). The isotopically shifted doublet
˚
further stretching of the N1–N2 bond (1.74 to 2.12 A), an
˚
asymmetric lengthening of the C5–N4 bond (1.47 to 1.78 A),
˚
lengthening of the N1–C6 bond (1.40 to 1.46 A) and contrac-
tion of the N3–N4 , C5–N1 and N2–N6 bonds (1.23 to 1.18,
˚
1.42 to 1.29 and 1.24 to 1.22 A, respectively), all of which
are, again, consistent with formation of HCN and N2 .
The products were optimized at the B3LYP/6-31G(d,p)
level of theory and IRC calculations were performed in the for-
ward direction for both 1-TS and 2-TS. Only 1-TS was shown
to directly result in product formation. For 2-TS, the IRC
pathway indicated the loss of a single HCN and what we
presume to be a nitrene intermediate prior to final product
formation. The results are summarized in Fig. 2.
(with respect to 4, JC–15 ¼ 10.9 Hz) resulting from coupling
N
to a single 15N is clearly visible. However, also evident is a tri-
plet (JC–15 ¼ 10.9 Hz) resulting from coupling to two equiva-
N
lent 15N atoms. This is the result of a double exchange at the
imidium ion intermediate exchange. Fig. 3(A) shows a simu-
lated spectrum of Fig. 3(B) including the integrals for each
individual peak with the area of the central peak of the triplet
taken as 2.0. The corresponding difference spectrum is shown
in Fig. 3(C). Based on integration, 5 was calculated to be
72.7% singly and 27.3% doubly 15N-labeled with less than
1.0% unlabeled. The broader lines of the doublet are attributed
to residual coupling to the 14N quadrupole.
To more accurately define the reaction energetics, 1 and
1-TS were re-optimized at the CCSD(T)/6-31G(d,p) level of
Optimization of thermally generated diazirinyl radicals
Two complications were encountered in the thermal genera-
tion of diazirinyl radicals, the well known formation of car-
benes,7 and an additional competing side reaction. In their
original work, Ingold and Maeda reported the generation of
various 3-substituted diazirinyl radicals in solution via thermo-
lysis of the parent 3-bromo-3-aryldiazirines in the presence of
di-tert-butylhyponitrite (TBHN) and triethylsilane (TES, see
eqns. (1)–(3)).1 The only products from the diazirine were b-
bromostyrene (arising from the halocarbene) and phenylaceto-
nitrile (arising from the diazirinyl radical). Initially, reactions
were run with increased equivalents of TBHN and TES to
favor reactions (2) and (3) relative to generation of the
unwanted halocarbene. A series of such reactions were run
Fig. 1 B3LYP/6-31G(d,p) geometries and relative energies for 1,
1-TS, 2 and 2-TS.
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 7 5 6 – 7 6 5
757

View Article Online
Table 1 B3LYP/6-31G(d,p) and CCSD(T)/6-31G(d,p) Coordinate Values for 1, 1-TS, 2 and 2-TS
B3LYP
CCSD(T)
B3LYP
Coordinate
1
1-TS
1.421
1
1-TS
1.425
Coordinate
2
2-TS
r(C1N2)
r(C1N3)
r(N2N4)
r(N4N5)
r(N5C6)
r(C1H7)
r(C6H8)
1.396
1.250
1.441
1.672
1.250
1.085
1.085
1.404
1.265
1.464
1.706
1.265
1.080
1.080
r(N1N2)
r(N1N3)
r(N3N4)
r(N1C5)
r(N2C6)
r(C5H7)
r(C6H8)
1.740
2.438
1.228
1.416
1.245
1.087
1.088
2.118
2.431
1.177
1.289
1.230
1.082
1.094
1.238
1.334
1.834
1.238
1.089
1.089
1.257
1.359
1.889
1.257
1.084
1.084
y(N3C1N2)
y(N4N2C1)
y(N5N4N2)
y(C6N5N4)
y(H7C1N2)
y(H8C6N5)
78.1
112.9
115.2
54.8
86.9
79.3
111.5
113.4
54.0
89.4
y(N3N1N2)
y(N4N3N1)
y(C5N1N2)
y(C6N2N1)
y(H7C5N1)
y(H8C6N2)
139.4
76.5
134.8
85.9
120.7
117.5
50.7
118.6
116.0
49.0
110.8
52.9
113.4
42.3
139.0
142.7
131.1
142.0
138.5
142.2
130.6
140.1
121.5
141.2
132.5
135.6
t(N4N2C1N3)
t(N5N4N2C1)
t(C6N5N4N2)
t(H7C1N2N4)
t(H8C6N5N4)
103.9
ꢀ46.2
ꢀ98.5
ꢀ73.6
177.2
98.3
4.2
102.5
ꢀ44.0
ꢀ98.0
ꢀ74.8
177.1
98.0
2.7
t(N4N3N1N2)
t(C5N1N2N4)
t(C6N2N1C5)
t(H7C5N1N2)
t(H8C6N2N1)
25.4
ꢀ34.9
104.5
ꢀ54.2
179.6
8.9
ꢀ39.9
106.8
ꢀ51.6
177.3
ꢀ106.6
ꢀ80.1
178.1
ꢀ104.7
ꢀ80.8
178.5
in order to optimize conditions for nitrile formation, and the
reaction progress was monitored by variable-temperature H
electronic structure of diazirinyl radicals is only slightly
perturbed by the 3-substituent (see above), 3-phenyl and
3-tolyl diazirinyl radicals were chosen because the parent 3-
bromo-3-aryldiazirines are known to be less volatile than the
corresponding 3-alkyl derivatives, and their synthesis is
straightforward.6 The following is a brief outline of how the
cross-over experiment was designed. Thermolysis of an equi-
molar solution of 15N-3-bromo-3-phenyldiazirine and 3-bromo-
3-tolyldiazirine (in the presence of di-tert-butylhyponitrite and
triethylsilane) results in the simultaneous generation of both
diazirinyl radicals. Once formed, each radical is capable of
self- or cross-dimerization. For cross-dimerization, 15N isotope
incorporation into the tolunitrile product is expected for reaction
via 2 (in certain cases, see Scheme 4), while no 15N isotope
incorporation is expected for reaction via 1.
1
NMR. A solution consisting of 3-bromo-3-phenyldiazirine
(ꢄ0.1 M), with 2 equiv. of TES and 5 equiv. of TBHN in
hexane (1 mL) was degassed and heated at 50 ꢁC under an
atmosphere of nitrogen. Complete disappearance of the
starting 3-bromo-3-phenyldiazine was observed after ꢄ1 h.
However, at least one unknown was clearly visible in the
aromatic region in addition to the desired benzonitrile. The
identity of the unknown(s) has yet to be clearly established
but is presumed to be an a-bromophenylimine based on the
findings of Barton et al. who also provided mechanistic
support for the radical induced dimerization and nitrogen
transfer reactions of substituted diazirines.8 Nevertheless,
addition of 10 equiv. of TES resulted in benzonitrile as the
only product to arise from the parent diazirine (Fig. 4).
Therefore, the 15N/14N ratio of the formed nitriles, deter-
mined by mass spectral analysis, should reveal whether reac-
tion via 1, 2 or a mixture of both is responsible for product
formation.
Experimental design
The cross-over experiment involves the simultaneous thermal
generation of two distinct diazirinyl radicals in solution, one
of which has been isotopically labeled with 15N. As the
The crossover experiment
For cross-dimerization between 3-phenyl and 3-tolydiazirinyl
radicals, only 2 could result in 15N isotopic incorporation into
the tolunitrile product (see Scheme 4). For the case of an
Table 2 Total and relative energies for 1 and 1-TS at various levels of
theory
Total energy/Eh
6-31G(d,p)
CCSD(T)
6-311þþ
6-311þþ
6-311þþ
Species B3LYP
CCSD(T) G(d,p)
G(2d,2p)
G(2df,2p)
1
ꢀ296.16423 ꢀ295.40913 ꢀ295.52388 ꢀ295.58349 ꢀ295.67973
1-TS ꢀ296.15966 ꢀ295.40512 ꢀ295.52037 ꢀ295.58203 ꢀ295.67553
Relative energy/kJ molꢀ1
Fig. 2 Summary of present theoretical work at the B3LYP/6-
31G(d,p) level of theory. Solid shapes denote stationary points while
open shapes denote data extracted from IRC calculations.
1
1-TS
0
12.1
0
10.5
0
9.2
0
3.8
0
10.9
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
758
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 7 5 6 – 7 6 5

View Article Online
Scheme 3
equimolar mixture of 5 (72.7% singly, 27.3% doubly 15N-
solution of 5 and 6 was heated under similar conditions (see
labeled) and 6, the percentages for 15N isotopic retention
(PhCN) and isotopic incorporation (TolCN) for 1 and 2 were
predicted statistically and are shown in Table 3. Possible iso-
tope effects and the methyl substituent effect were neglected.
As can be seen, ꢄ23% 15N isotopic incorporation into TolCN
is expected if 2 is exclusively formed. At the other extreme, 0%
15N isotopic incorporation is expected for reaction via 1. Three
equimolar solutions of 5 and 6 were degassed and heated at
50 ꢁC under an atmosphere of nitrogen for 1 h. The percent
15N isotopic retention and isotopic incorporation (for PhCN
and TolCN) in each case was determined by GC-MS. The
results are shown in Table 3 (trials 1–3). Essentially no 15N iso-
topic incorporation above the natural abundance level (ca.
0.08%) was observed in TolCN. In an attempt to increase
the probability of 15N isotopic transfer, an approximately 4:1
trial 4, Table 3). Again, the reaction proceeded with essentially
zero 15N isotopic incorporation into the tolunitrile.
The head-to-head nitrogen dimerization (1) is preferred by
at least a factor of 100, even though 2 is favored thermodyna-
mically by more than 125 kJ molꢀ1. This confirms the original
proposal by Ingold and Maeda based on their insight that the
unpaired spin density of diazirinyl radicals is delocalized exclu-
sively across the two nitrogen 2pz atomic orbitals with no
contribution from Ca according to EPR characterization and
INDO calculations.1
Unimolecular decomposition pathways of diazirinyl radicals?
Diazirinyl radicals have been proposed as intermediates in the
formation of ‘‘prompt’’ NO during hydrocarbon combus-
tion.3,4 Fenimore originally proposed the formation of atomic
nitrogen as the key step in the thermoneutral but spin forbid-
den reaction (4).
2
4
CH ð PÞ þ N2 ! HCN þ N ð SÞ
ð4Þ
Fig. 4 1H NMR spectra for the thermolysis of 3-bromo-3-phenyldia-
zirine under optimal conditions. Peaks are labeled as ortho (o), meta
(m) and para (p) corresponding to the phenyl substituent protons of
3-bromo-3-phenyldiazirine (A) and benzonitrile (B).
Fig. 3 125 MHz 13C NMR expansion of 5-Ca : (A) simulated, (B)
experimental and (C) difference.
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 7 5 6 – 7 6 5
759

View Article Online
shift (147 kJ molꢀ1 barrier) to form 2HNCN followed by
decomposition (351 kJ molꢀ1 barrier) to NCN (3Sgꢀ) and H,
with the former being readily oxidized to NO under combus-
tion conditions.18 Additional calculations by Takayanagi19
support their results.
2
3
2
CH ð PÞ þ N2 ! c-2HCN2 ! HNCN ! NCN ð SgꢀÞ þ H
ð6Þ
Bise et al. investigated the photodissociation spectroscopy
and dynamics of HNCN in the gas phase.20 They found the
2
2
dominant product channel to be isomerization of HNCN to
c-2HCN2 followed by dissociation to CH þ N2 . Their data
was consistent with the energy surface calculated by Cui and
Morokuma12 suggesting that 2HNCN is accessible from
CH þ N2 and could play a vital role in the formation of
‘‘prompt’’ NO. They concluded that ‘‘Our results therefore
indirectly support the recent proposal by Moskaleva and
Lin18 that the CH þ N2 reaction does not yield N (4S) þ HCN,
but instead results in higher-energy, spin-allowed products
H (2S) þ NCN (X3Sgꢀ) through dissociation of the HNCN
radical.’’ (see reactions (5) and (6), respectively). The recent
detection of NCN in flames by Smith21 provides experimental
support for the Moskaleva and Lin proposal.18 However, there
remains no direct experimental or recent theoretical evidence
for the formation of ground-state atomic nitrogen from
CH þ N2 .
Scheme 4
Because of the importance of nitrogen fixation, we examined
the fate of the phenyldiazirinyl radical under high dilution con-
ditions to check for competing unimolecular reaction/decom-
position pathways. Our search–see details below–found no
evidence for either atomic nitrogen formation, or for PhNCN
derived products expected from the exothermic rearrangement
corresponding to the Moskaleva and Lin pathway. These
unimolecular processes occur too slowly under our conditions
to compete with the facile bimolecular pathway even under
high dilution conditions (10ꢀ4 M). The results are shown in
Table 4.
In all cases, the only product arising from the parent 3-
bromo-3-phenyldiazirine was benzonitrile. Neither 1H-3-phe-
nyldiazirine nor 3H-3-phenyldiazirine (the latter synthesized
according to the method of McMahon et al., see Experimental
section)22 resulting from possible hydrogen atom extraction by
the 3-phenyldiazirinyl radical were observed. The unimolecular
process was also counter-indicated by the absence of a chemi-
cally induced dynamic nuclear polarization (CIDNP) effect in
the benzonitrile formed in the NMR spectrometer. In an
attempt to facilitate the doublet–quartet crossing, the same
reactions were repeated using 3-bromo-3-(4-bromophenyl)di-
azirine. In all cases, the results were identical to those observed
for 3-bromo-3-phenyldiazirine and no ‘‘heavy-atom’’ effect
was found.
Subsequent work by several researchers,9–11 most recently
by Cui and Morokuma,12 have led to a proposed mechanism
involving insertion of CH (2P) into N2 (45.2 kJ molꢀ1 barrier)
to ultimately form the cyclic diazirinyl radical (c-2HCN2) fol-
lowed by doublet–quartet intersystem crossing (113.8 kJ molꢀ1
barrier) and subsequent decomposition (86.2 kJ molꢀ1
barrier) to HCN and N (4S),
2
4
CH ð PÞ þ N2 ! c-2HCN2 ! c-4HCN2 ! HCN þ N ð SÞ
ð5Þ
Although reaction (5) had been widely accepted as the key step
in the formation of ‘‘prompt’’ NO, no direct proof exists to
date. Furthermore, recent dynamic simulations by both Wada
and Takayanagi13 and by Cui et al.14 found the calculated rate
constants for reaction (5) to be smaller than the experimental
rate constants by more than two orders of magnitude.11,15–17
More recently, Moskaleva and Lin have proposed a spin-
allowed process in which c-2HCN2 undergoes a hydrogen atom
Table 3 Predicted and observed 15N-isotopic ratios for the cross-over
experiment
Calculated
We conclude that bimolecular decomposition of the 3-phe-
nyldiazirinyl radical is significantly more favorable than either
of the proposed unimolecular pathways even at low radical
concentrations under our hexane solution conditions.
Dimer
PhC15N/TolC15N (%)
1
2
60.2/0.0
47.7/22.7
Experimentala
Trial
Table 4 Reaction of 3-bromo-3-phenyldiazirine at increasing dilution
Conditionsb
PhC15N/TolC15N (%)
[PhC(Br)N2]0
TBHNb
TESb
t(s)c
a
1
2
3
4
A
A
A
B
56.6(60.3)/0.9
58.0(57.7)/0.8
59.4(56.2)/0.7
57.8(57.2)/0.3
8.6 ꢂ 10ꢀ2
1.0 ꢂ 10ꢀ2
1.0 ꢂ 10ꢀ3
0.85 ꢂ 10ꢀ4
5
5
10
3600
3300
4500
600
50
1000
5
50
11 385
a
Values in parenthesis denote electron impact (EI) data, otherwise
b
chemical ionization (CI) data was used. A: 0.05 M 5, 0.05 M 6, 1.0
a
b
Molarity in hexane. Equivalents with respect to 3-bromo-3-
c
phenyldiazirine. Total reaction time for disappearance of 3-bromo-
3-phenyldiazirine.
M TES and 0.5 M TBHN in hexane at 50 ꢁC for 1 h; B: 0.08 M 5,
0.02 M 6, 1.0 M TES and 0.5 M TBHN in hexane at 50 ꢁC for 1 h.
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
760
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 7 5 6 – 7 6 5

View Article Online
Kinetic simulations
Therefore, reaction (11) was included and the rate constant,
11 , was optimized. The best fits to the experimental data over
k
The reactions believed to be relevant to the system of interest
are shown in eqns. (7)–(16). Reactions (8), (12), (15) and (16)
are those responsible for formation and decomposition of
the 3-phenyldiazirinyl radical dimer, while reactions (7), (9),
(10), (13) and (14) represent competing side-reactions includ-
ing: thermolysis of the parent 3-bromo-3-phenyldiazirine, radi-
cal recombination and b-scission of the tert-butoxy radical and
radical recombination and disproportionation of the triethyl-
silyl radical, respectively.
the entire concentration range occurred when a value of
1.0 ꢂ 106 Mꢀ1 sꢀ1 was used for k11 which is in good agreement
with ka ¼ 1.6 ꢂ 106 Mꢀ1 sꢀ1 reported by Baignee et al. for the
reaction of tert-butoxy radicals with cyclohexane.25 The
experimental and simulated data for the highest and lowest
initial concentrations of 3-bromo-3-phenyldiazirine are shown
in Fig. 5. The experimental and simulated data are in satisfac-
tory agreement and are consistent with the proposed reaction
scheme (reactions (7)–(16)). The approximately zero-order
kinetics (Fig. 5) show that the observed reaction rate is
controlled by the rate of generation of the triethylsilyl radical.
Based on the reactions (7)–(16), the following steady-state
equations can be derived for the tert-butoxy, triethylsilyl and
3-phenyldiazirinyl radicals. For the tert-butoxy radical:
PhCðBrÞN2 ! PhCBr þ N2
ð7Þ
ð8Þ
ð9Þ
ð10Þ
ð11Þ
ðCH3Þ CONNOCðCH3Þ ! 2ðCH3Þ COꢃ þ N2
3
3
3
2ðCH3Þ COꢃ ! ðCH3Þ COOCðCH3Þ
3
3
3
ðCH3Þ COꢃ ! CH3ðC
B
OÞCH3 þ CH3
ꢃ
½ðCH3Þ COꢃꢅ ¼ 1=ð2k9Þfꢀk10 ꢀ k11½C6H14ꢅ
3
3
ss
ðCH3Þ COꢃ þ C6H14 ! ðCH3Þ COH þ C6H13
ꢃ
ꢀ k12½ðCH3CH2Þ3SiHꢅ0 þ ðk210 þ 2k10k11½C6H14ꢅ
3
3
ðCH3Þ COꢃþ ðCH3CH2Þ SiH !
2
þ 2k10k12½ðCH3CH2Þ3SiHꢅ0 þ k211½C6H14ꢅ
3
3
ðCH3Þ COH þ ðCH3CH2Þ Siꢃ ð12Þ
2
þ 2k11k12½C6H14ꢅ½ðCH3CH2Þ3SiHꢅ0 þ k212½ðCH3CH2Þ3SiHꢅ0
3
3
2ðCH3CH2Þ Siꢃ ! ðCH3CH2Þ SiSiðCH2CH3Þ
ð13Þ
1=2
3
3
3
þ 4k9k8½ðCH3Þ3CONNOCðCH3Þ3ꢅ0Þ
for the triethylsilyl radical:
g
ð17Þ
2ðCH3 CH2Þ Siꢃ !
3
ðCH3CH2Þ3SiH þ CH3CH
B
SiðCH2CH3Þ2 ð14Þ
½ðCH3CH2Þ Siꢃꢅ ¼ 1=ð2k13 þ 2k14Þfꢀk15½PhCðBrÞN2ꢅ
3
ss
0
ðCH3CH2Þ Siꢃ þ PhCðBrÞN2 !
3
ꢃ
2
þ ðk215½PhCðBrÞN2ꢅ þ 4k12½ðCH3Þ3CO ꢅss
ꢃ
ꢂ ½ðCH3CH2Þ SiHꢅ0 k13 þ 4k12½ðCH3Þ COꢃꢅ
ðCH3CH2Þ3SiBr þ c-PhCN2
ð15Þ
ð16Þ
3
0
3
ss
ꢃ
2c-PhCN2 ! 2PhCN þ N2
1=2
ꢂ ½ðCH3CH2Þ3SiHꢅ0k14Þ
g
and for the 3-phenyldiazirinyl radical:
ð18Þ
The corresponding rate constants (at 50ꢁ) derived from
values reported in the literature are shown in Table 5.
Reactions (7)–(16) along with the corresponding rate constants
listed in Table 5 were loaded into the Chemical Kinetics Simula-
tor (CKS)30 program, and simulations were run for each set of
reaction conditions listed in Table 4. Initially, hydrogen atom
abstraction from the solvent by the tert-butoxy radical, reaction
(11), was omitted. However, in all cases, the simulated reaction
times were found to be too short compared with the experimental
data. It was found that lowering the effective concentration of
the tert-butoxy radical increased the reaction times in a manner
consistent with the experimental data.
ꢃ
ꢃ
1=2
½PhCN2 ꢅss ¼ ð1=k16Þðk15k16½ðCH3CH2Þ3Si ꢅss½PhCðBrÞN2ꢅ0Þ
ð19Þ
The decrease in 3-phenyldiazirinyl radical steady-state concen-
ꢃ
trations ([PhCN2 ]ss) on dilution calculated for each set of
reaction conditions are shown in Table 6.
All concentrations correspond to molarity in hexane.
ꢃ
Assuming that even at the highest dilution ([PhCN2 ]ss
¼
2.2 ꢂ 10ꢀ9 M), bimolecular decomposition is dominant
with little or no contribution from the unimolecular pathway,
as suggested by the experimental data, a minimum activation
energy of ꢄ70 kJ molꢀ1 can be calculated for unimolecular
decomposition of the 3-phenyldiazirinyl radical. An additional
limit is established by assuming that the unimolecular pathway
to form benzonitrile is the only path for generation. When this
is the case, further simulations have shown that a minimum
rate for unimolecular decomposition of 5.0 ꢂ 10ꢀ3 sꢀ1 is
required to fit the experimental data. This would suggest a
maximum activation energy of at least 90 kJ molꢀ1 for the loss
of atomic nitrogen from the 3-phenyldiazirinyl radical. In
other words, according to the above results the activation
energy for nitrogen atom loss lies in the 70–90 kJ molꢀ1 range
which is in excellent agreement with theoretically determined
value of 82.0 kJ molꢀ1 reported by Cui and Morokuma.12
Furthermore there was no evidence of products involving
phenyl migration from C to N–the intermediate from the
Moskaleva and Lin mechanism.
Table 5 Reported rate constants for reactions of interest
Equation
k (calc. 323 K)
Ref.
k7
k8
1.7 ꢂ 10ꢀ5 sꢀ1
0.57 ꢂ 10ꢀ4 sꢀ1
7
23
24
24
25
26
27
28
29
1
a
k9
k10
2.5 ꢂ 1010 (Mꢀ1 sꢀ1 b
)
6.7 ꢂ 104 Mꢀ1 ꢀ1
s
c
k11
k12
1.0 ꢂ 106(Mꢀ1 sꢀ1 d
)
8.1 ꢂ 106 Mꢀ1 ꢀ1
s
e
k13
k14
1.7 ꢂ 109 Mꢀ1 ꢀ1
s
3.2 ꢂ 108 Mꢀ1 ꢀ1
s
f
g
k15
k16
2.7 ꢂ 109 Mꢀ1 sꢀ1
2.5 ꢂ 1010 (Mꢀ1 sꢀ1 b
)
h
a
Assumed diffusion controlled based on k295 ¼ 2.8 ꢂ 109 Mꢀ1 sꢀ1
.
b
Diffusion limit calculated to be 2.5 ꢂ 1010 Mꢀ1 sꢀ1 using the
Stokes–Einstein equation kdc ¼ 8000RT/Z with Z ¼ 2.82 ꢂ 10ꢀ4 for
Diazirines as radical traps and nitrogen atom transfer agents
hexane. Value optimized to fit experimental data k300 ¼ 1.6 ꢂ 106
c
d
Our results corroborate and expand on the original work of
Ingold1 and of Barton et al.8 in demonstrating that diazirines
and diazirinyl radicals are excellent radical traps and nitrogen
atom transfer agents. We have confirmed that trapping occurs
at nitrogen because of the high spin density at those centers as
originally proposed by Ingold. Adduct formation lengthens
Mꢀ1 sꢀ1 reported for tert-BuOꢃ þ c-C6H12
.
Based on 2(CH3)3Siꢃ !
e
f
(CH3)3SiSi(CH3)3 .
Based on 2(CH3)3Siꢃ ! (CH3)3SiH þ CH2
¼
Si(CH3)3 . Based on (CH3CH2)Siꢃ þ PhCH2Br ! (CH3CH2)SiBr þ
g
ꢃ
PhCH2 . Assumed diffusion controlled based on k195 ¼ 2.0 ꢂ 108
h
Mꢀ1 sꢀ1
.
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 7 5 6 – 7 6 5
761

View Article Online
Fig. 5 Experimental and simulated data for reaction of 3-bromo-3-phenyldiazirine in the presence of triethylsilane and di-tert-butyl hyponitrite at
50 ꢁC. (A) 0.086 M 4, 0.86 M TES and 0.43 M TBHN in hexane. (B) 0.000085 M 4, 0.97 M TES and 0.0043 M TBHN in hexane. (C, D) Simulated
reaction of (A) and (B), respectively. Solid and dashed lines correspond to 3-bromo-3-phenyldiazirine and benzonitrile, respectively.
and weakens the internal N–N bond which is then readily ther-
molysed. Dissociation results in transfer of one of the diaziri-
nyl nitrogens to the incoming radical. This intrinsically high
reactivity of diazirinyl radicals as nitrogen atom transfer
agents suggests that a bimolecular pathway may also be avail-
able in gas phase combustion leading to ‘‘prompt’’ NO. For a
bimolecular path, e.g. reaction (20), to compete with unimole-
cular decomposition pathways in the gas phase, the nitrogen
atom acceptor ‘‘X’’ must be present in high abundance.
Calculations performed at the B3LYP/6-31G(d,p) level of the-
ory have shown that both C–N and N–N dimers have low-bar-
rier transition states consistent with the observed products. A
cross-over experiment was employed to determine the relative
contributions of each dimer towards product formation.
Under our hexane solution conditions, the N–N dimer was
found to be preferred by at least a factor of 100 over the C–
N dimer. The reaction energetics of the N–N dimer surface
were further explored using coupled-cluster theory. At the
CCSD(T)/6-311þþG(2df,2p) level of theory, the barrier for
bimolecular decomposition of diazirinyl radicals was found
to be þ10.9 kJ molꢀ1. Subsequent reactions were run at
increasing dilution, to favor the spin-forbidden unimolecular
decomposition. In all cases, no evidence for unimolecular
decomposition was observed. Likewise, reactions were
repeated incorporating a para-bromo substituent into the 3-
aryldiazirinyl radicals. No ‘‘heavy-atom’’ effect was observed
suggesting that bimolecular decomposition of the 3-aryldiazir-
inyl radicals is significantly more favorable even at low radical
concentrations under our reaction conditions. Kinetic simula-
tions were employed and suggest that the activation energy for
unimolecular decomposition lies in the 70–90 kJ molꢀ1 range.
A possible bimolecular path to ‘‘prompt’’ NO formation
involving the diazirinyl radical has been postulated.
c-2HCN2 þ ‘‘X’’ ! HCN þ ‘‘2NX’’
ð20Þ
Obviously, under combustion conditions, ‘‘X’’ cannot be
another diazirinyl radical. However ‘‘X’’ may be a number
of other species with sufficient abundance in flames such as
CO or alkyl radicals to permit reaction without the need for
a spin flip. Preliminary calculations indicate that diazirinyl
radical may combine with CO to form HCN and NCO
exothermally via a very low barrier process.
Conclusion
The dimerization and decomposition of 3-aryldiazirinyl radi-
cals have been explored both theoretically and experimentally.
Table 6 Calculated steady-state 3-phenyldiazirinyl radical
concentrations
Experimental
ꢃ
(1) General methods
[PhC(Br)N2]0
[TBHN]0
[TES]0
[PhCN2 ]ss
All starting materials were obtained from Aldrich (Milwaukee,
WI) and were used as received unless otherwise noted. Anhy-
drous solvents were dried over either sodium metal or calcium
chloride and distilled under an atmosphere of N2 . All moist-
ure-sensitive reactions were conducted in oven- (130 ꢁC) or
8.6 ꢂ 10ꢀ2
1.0 ꢂ 10ꢀ2
1.0 ꢂ 10ꢀ3
0.85 ꢂ 10ꢀ4
4.3 ꢂ 10ꢀ1
5.0 ꢂ 10ꢀ2
5.0 ꢂ 10ꢀ3
4.2 ꢂ 10ꢀ3
8.6 ꢂ 10ꢀ1
5.0 ꢂ 10ꢀ1
10 ꢂ 10ꢀ1
9.7 ꢂ 10ꢀ1
22 ꢂ 10ꢀ9
6.3 ꢂ 10ꢀ9
2.4 ꢂ 10ꢀ9
2.2 ꢂ 10ꢀ9
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
762
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 7 5 6 – 7 6 5

View Article Online
of 4-methylbenzamidine and 4-bromobenzamidine hydro-
chloride hydrate, ammonium chloride was used.
flame-dried glassware under an atmosphere of N2 . 1H and
13C NMR spectra were recorded on a Bruker Avance DRX
500 (500 MHz) or Varian Unity INOVA 300 (300 MHz) spec-
trometer. Chemical shifts are reported in parts per million
from tetramethylsilane with the solvent resonance as the inter-
nal standard (CHCl3 ; dH 7.26). Data are reported as follows:
chemical shift, multiplicity (s ¼ singlet, d ¼ doublet,
t ¼ triplet, q ¼ quartet, m ¼ multiplet, br ¼ broad), coupling
constant (Hz). 13C NMR spectra were recorded with complete
proton decoupling. Chemical shifts are reported in parts per
million from tetramethylsilane with the solvent resonance as
the internal standard (CDCl3 ; d 77.0). GC–MS analysis was
performed on a Hewlett Packard Engine mass spectrometer
(Hewlett Packard 5890 series II GC, 5989B MS) using electron
impact (EI) and chemical ionization (CI) by the Purdue
University Campus-Wide Mass Spectrometry Center (electron
energy, 70 eV; CI reagent gas, isobutene; injector temperature,
250ꢁ; ion source temperature, 200ꢁ capillary column, DB1 or
DB5).
(c) 3-Bromo-3-aryl diazirines. 3-Bromo-3-phenyl, 15N-3-
bromo-3-phenyl, 3-bromo(4-methylphenyl) and 3-bromo(4-
bromophenyl) diazirines were prepared from the corresponding
3-aryl amidine hydrochloride hydrates according to the
method of Graham.6 A general description is as follows: A
three-neck, 1-L round-bottom flask was equipped with a
mechanical stirrer and a thermometer. It was charged with
the 15N-benzamidine hydrochloride hydrate (20 mmol),
sodium bromide (21 g), DMSO (250 mL) and water (50
mL). Once the salts had dissolved, hexane (100 mL) was
added, and the reaction mixture was cooled to ꢀ5 ꢁC with
an ice/salt bath.
A fresh solution of sodium hypobromite (0.47 mol) was pre-
pared by the slow addition of bromine (24 mL, 0.47 mol) to a
stirred and cooled (ꢀ5 ꢁC) solution of sodium hydroxide (50 g)
and sodium bromide (100 g) in water (360 mL). The cool
sodium hypobromite solution was slowly added (ꢄ10 mL
minꢀ1) to the vigorously stirred benzamidine reaction mixture.
The reaction temperature was maintained below 15 ꢁC with an
ice-bath. The mixture was allowed to stir for 1 h, diluted with
water (200 mL) and transferred to a large separatory funnel.
The hexane layer was separated, and the aqueous layer was
extracted with hexane (4 ꢂ 50 mL). The combined hexane frac-
tions were dried (MgSO4), filtered and concentrated under
reduced pressure. The crude residue was purified by flash-chro-
matography through silica gel (230–400, 2.5 ꢂ 25 cm) using
hexane (ꢄ200 mL) as the eluent. The solvent was removed
under reduced pressure to yield the desired product as an oil
in typical yields of 35–45%. 3-Bromo-3-phenyldiazirine: 1H
NMR (500 MHz, CDCl3) d 7.14 (m, 3H), 7.37 (m, 2H): 13C
NMR (125 MHz, CDCl3) d 38.0, 126.6, 128.5, 129.4, 136.2.
(2) Computational methods
All calculations were carried out using the GAUSSIAN 98
suite of programs and basis sets.31 The geometries of all species
considered were optimized using density functional theory
with the B3LYP functional.32 Calculations were initially per-
formed using the 6-31G(d,p) basis set. The optimized geome-
tries of bound species were determined to be true minima by
the absence of imaginary frequencies; those of transition states
were identified by their single imaginary frequency. The reac-
tion coordinate of each molecule was also characterized to
ensure that the true transition-state structure was found. An
intrinsic reaction coordinate (IRC) calculation was performed
to verify if the transition state proceeded to the correct reac-
tants and products. Geometries were then re-optimized using
coupled-cluster methods, including single, double and con-
nected triple excitations (CCSD(T)), with the latter being
handled perturbatively.33 To improve the accuracy of the ener-
getics, subsequent single-point calculations were performed on
the CCSD(T)/6-31G(d,p) optimized geometries using the
6-311þþG(d,p), 6-311þþG(2d,2p), and 6-311þþG(2df,2p)
basis sets.
1
15N-3-Bromo-3-phenyldiazirine: H NMR (500 MHz, CDCl3)
d 7.14 (m, 3H), 7.39 (m, 2H): 13C NMR (125 MHz, CDCl3) d
38.0 (d, JCN ¼ 10.9 Hz), 126.6, 128.5, 129.4, 136.6. 3-Bromo-3-
(4-methylphenyl) diazirine: 1H NMR (300 MHz, CDCl3) d
2.36 (s, 3H), 7.01 (m, 2H), 7.16 (m, 2H): 13C NMR (75
MHz, CDCl3) d 21.1, 38.2, 126.5, 129.1, 133.8, 139.5. 3-
Bromo-3-(4-bromophenyl) diazirine: 1H NMR (500 MHz,
CDCl3) d 7.00 (m, 2H), 7.50 (m, 2H): 13C NMR (125 MHz,
CDCl3) d 36.9, 123.9, 128.2, 131.7, 135.7.
(3) Synthesis
(d) 2,4,6-Triphenyl-1,3,5-triazabicyclo[3.1.0]hexane.
was
prepared according to the method of McMahon et al.22 A 1
L round-bottom flask was equipped with a mechanical stirrer.
It was charged with MeOH (300 mL) and cooled to 0 ꢁC with
an ice-bath and saturated with ammonia for 1 h. The solution
was cooled to ꢀ10 ꢁC with an ice/salt bath, saturated with
ammonia for an additional 30 min then cooled to ꢀ40 ꢁC with
an acetonitrile/liquid N2 bath. A solution of tert-butyl alcohol
(17.5 mL) and tert-butyl hypochlorite (17.5 mL, 0.145 mol)35
was added dropwise over the period of 20 min. The solution
was warmed to ꢀ10 ꢁC, and benzaldehyde (25 mL, 0.246
mol) was added dropwise over the period of 10 min. The solu-
tion was allowed to stir for an additional 1.5 h at ꢀ10 ꢁC and
gradually warmed to room temperature overnight without stir-
ring. The precipitate that formed was filtered off and dried
under reduced pressure to yield the desired product (1.9 g,
7.4%): 1H NMR (300 MHz, CDCl3) d 3.03 (br s, 1H), 3.25
(s, 1H), 5.24–5.70 (m, 2H), 7.24–7.81 (m, 15H): 13C NMR
(75 MHz, CDCl3) d 51.9, 80.4, 81.4, 127.8, 127.9, 128.3,
128.4, 128.6, 128.7, 136.0, 136.2, 140.3.
(a) Di-tert-butylhyponitrite (TBHN). was prepared accord-
ing to the method of Mendehall (51.5%).34 1H NMR (500
MHz, CDCl3) d 1.39 (s, 18H): 13C NMR (125 MHz, CDCl3)
d 27.8, 81.2.
(b) 3-Aryl amidine hydrochloride hydrates. Benzamidine
hydrochloride hydrate was purchased directly from Aldrich.
15N-benzamidine, 4-methylbenzamidine and 4-bromobenzami-
dine hydrochloride hydrate were prepared from the corres-
ponding nitriles according to the method of Schaefer and
Peters.5 A general description is as follows: A three-neck, 1-L
round-bottom flask was equipped with a mechanical stirrer.
It was charged with benzonitrile (3.7 mL, 36.2 mmol) and
MeOH (anhydrous, 500 mL). Once dissolved, the mixture
was heated to ꢄ35 ꢁC using an oil bath. A catalytic amount
of sodium metal (ꢄ0.1 g) was added, and the reaction was
allowed to stir under N2-atmosphere for 48 h. 15N-ammonium
chloride (2.0 g, 36.8 mmol) was added, and stirring was contin-
ued for an additional 24 h. The reaction mixture was concen-
trated under reduced pressure, and the crude residue was
slurried with Et2O (3 ꢂ 100 mL), dissolved in EtOH (ꢄ100
mL) and filtered. The solvent was removed under reduced
pressure and the product (typically a white solid in yields of
50–60%) was used without further purification. In the case
(e) 3-H-3-Phenyldiazirine. This was prepared according to
the method of McMahon et. al.22 2,4,6-Triphenyl-1,3,5-triaza-
bicyclo[3.1.0]hexane (2.0 g, 6.3 mmol) was suspended in
MeOH (20 mL) and the mixture was cooled to 0 ꢁC with an
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 7 5 6 – 7 6 5
763

View Article Online
ice-bath. A solution of tert-butyl alcohol (2.0 mL) and tert-
butylhypochlorite (0.75 mL, 6.2 mmol)35 was added dropwise
over a period of 10 min. The resulting solution was allowed to
stir at 0 ꢁC for 1 h and at room temperature for an additional
20 min. The solution was poured into 0.05 M aqueous sodium
metabisulfite (300 mL), extracted with hexane (2 ꢂ 50 mL),
concentrated under reduced pressure and purified by flash-
chromatography through silica gel (230–400, 2.5 ꢂ 25 cm)
using hexane (ꢄ200 mL) as the eluent. The solvent was
removed under reduced pressure to yield the desired product
as an oil (0.24 g, 32.3%): 1H NMR (500 MHz, CDCl3) d
2.05 (s, 1H), 6.92 (m, 2H), 7.32 (m, 3H): 13C NMR (125
MHz, CDCl3) d 23.3, 125.1, 128.0, 128.3, 136.3.
Acknowledgements
We thank Ms A. Rothwell of the Purdue University Mass
Spectrometry Service for the mass spectra and Dr K. Hallenga
for help with the NMR pulse sequences. Partial financial
support was generously provided by Texaco and by BASF
through the Purdue Industrial Associates program.
References
1
2
3
4
Y. Maeda and K. U. Ingold, J. Am. Chem. Soc., 1979, 101,
837–840.
C. P. Fenimore, in 13th Symp. (Int.) Combust., The Combustion
Institute, Pittsburgh, PA, 1971, p. 373.
C. T. Bowman, in 24th Symp. (Int.) Combust. Proc., The
Combustion Institute, Pittsburgh, PA, 1993, pp. 859–878.
J. Warnatz, U. Maas and R. W. Dibble, in Combustion: Physical
and Chemical Fundamentals, Modeling and Simulation,
Experiments, Pollutant Formation, Springer-Verlag, Berlin–
Heidelberg–New York, 1996, pp. 219–236.
(4) NMR reactions
(a) Pulse sequence. A double pulse field gradient spin echo
sequence (DPFGSE) was employed for selective excitation of
the aromatic region and subsequent dephasing and elimination
of the unwanted solvent signals (see Fig. 6).36
A variable delay list was incorporated to record a spectrum
(typically 1–4 scans) every 300 s for a total of 19 spectra.
5
F. C. Schaefer and G. A. Peters, J. Org. Chem., 1961, 26,
412–418.
6
7
8
W. H. Graham, J. Am. Chem. Soc., 1965, 87, 4396–4397.
M. T. H. Liu, Chem. Soc. Rev., 1982, 11(2), 127–140.
D. H. R. Barton, J. Cs. Jaszberenyi, E. A. Theodorakis and
J. H. Reibenspies, J. Am. Chem. Soc., 1993, 115, 8050–8059.
D. R. Yarkony, J. Phys. Chem., 1993, 97, 4407–4412.
(b) Sample preparation. All samples were prepared in a
glove-box under an atmosphere of N2 . A general description
is as follows: A 0.086 M stock solution of 3-bromo-3-phenyl-
diazirine was prepared in hexane (degassed under reduced
pressure). 1 mL was transferred to a 1-dram vial, and triethyl-
silane (10 equiv.) and tert-butylhyponitrite (5 equiv.) were
added. Once dissolved, ꢄ700 mL were transferred to a 5 mm
open-cap NMR tube with a rubber/PTFE septum, and a con-
centric-tube of acetone-d6 was added to provide a lock signal.
9
10 J. M. L. Martin and P. R. Taylor, Chem. Phys. Lett., 1993, 209,
143–150.
11 J. A. Miller and S. P Walch, Int. J. Chem. Kinet., 1996, 29,
253–259.
12 Q. Cui and K. Morokuma, Theor. Chem. Acc., 1999, 102,
127–133.
13 A. Wada and T. Takayanagi, J. Chem. Phys., 2002, 116,
7065–7072.
(c) NMR experiment. A general description is as follows:
The reaction sample was inserted into the spectrometer,
locked, shimmed and the offset for the 180ꢁ Q3 gaussian cas-
cade shaped soft pulse was optimized. The sample was ejected
and a 100% ethylene glycol sample was inserted. The euro-
therm was set to 323 K (50 ꢁC), and the sample was allowed
to equilibrate for ꢄ7 min. A 1D spectrum was acquired, and
chemical shift difference between the CH2 and OH groups.37
The reaction sample was inserted and locked, and the experi-
ment was immediately started. Once complete, each spectrum
was processed individually using standard procedures. A sepa-
rate experiment showed that the sample had equilibrated at the
desired temperature after ꢄ200 s, and the temperature was
stable within ꢆ0.3 K over the specified time interval.
14 Q. Cui, K. Morokuma, J. M. Bowman and S. J. Klippenstein,
J. Chem. Phys., 1999, 110, 9469–9482.
15 D. Lindackers, M. Burmeister and P. Roth, in 23rd Symp. (Int.)
Combust. Proc., The Combustion Institute, Pittsburgh, PA, 1991,
pp. 251–258.
16 A. J. Dean, R. K. Hanson and C. T. Bowman, in 23rd Symp.
(Int.) Combust. Proc., The Combustion Institute, Pittsburgh,
PA, 1991, pp. 259–266.
17 K. H. Becker, H. Geiger and P. Wiesen, Int. J. Chem. Kinet., 1996,
28, 115–123.
18 (a) L. V. Moskaleva, W. S. Xia and M. C. Lin, Chem. Phys. Lett.,
2000, 331, 269–277; (b) L. V. Moskaleva and M. C. Lin, Z. Phys.
Chem., 2001, 215, 1043–1050.
19 T. Takayanagi, Chem. Phys. Lett., 2003, 368, 393–398.
20 R. T. Bise, A. A. Hoops and D. M. Neumark, J. Chem. Phys.,
2001, 114, 9000–9011.
21 G. P. Smith, Chem. Phys. Lett., 2003, 367, 541–548.
22 R. J. McMahon, C. J. Abelt, O. L. Chapman, J. W. Johnson,
C. L. Kreil, J. LeRoux, A. M. Mooring and P. R. West, J. Am.
Chem. Soc., 1987, 109, 2456–2469.
23 H. Kiefer and T. G. Traylor, Tetrahedron Lett., 1966, 49,
6163–6168.
1
the actual temperature was determined by measuring the H
(d) Crossover experiment. A general description is as fol-
lows: All samples were prepared in a glove-box under an atmo-
sphere of N2 as previously described with the exception that
the 1-dram vials, themselves, were heated in an oil-bath at
50 ꢁC for 1.5 h. Once complete, the samples were submitted
for GC-MS analysis. The percent isotope incorporation was
determined after correcting for natural isotope abundance
using standard procedures.
24 Y. P. Tsentalovich, L. V. Kulik, N. P. Gritsan and A. V.
Yurkovskaya, J. Phys. Chem. A, 1998, 102, 7975–7980.
25 A. Baignee, J. A. Howard, J. C. Scaiano and L. C. Stewart, J. Am.
Chem. Soc., 1983, 105, 6120–6123.
26 C. Chatgilialoglu, J. C. Scaiano and K. U. Ingold, Organometal-
lics, 1982, 1, 466–469.
27 P. P. Gaspar, A. D. Haizlip and K. Y. Choo, J. Am. Chem. Soc.,
1972, 94, 9032–9037.
28 B. J. Cornett, K. Y. Choo and P. P. Gaspar, J. Am. Chem. Soc.,
1980, 102, 377–378.
29 C. Chatgilialoglu, K. U. Ingold and J. C. Scaiano, J. Am. Chem.
Soc., 1982, 104, 5123–5127.
30 Chemical Kinetics Simulator Version 1.01
Business Machines Corporation, 1996.
# International
31 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A.
Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Montgomery,
Jr., R. E. Stratmann, J. C. Burant, S. Dapprich, J. M. Millam,
A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J. Tomasi,
V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli,
C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P. Y.
Ayala, Q. Cui, K. Morokuma, D. K. Malick, A. D. Rabuck,
Fig. 6 DPFGSE sequence. SP1 denotes 180ꢁ Q3 gaussian cascade
shaped soft pulse (2 ppm excitation window) centered on the aromatic
region (ꢄ7.2 ppm). GP1 and GP2 denote 1 ms gradient pulses at 30%
and 15% maximum gradient strength, respectively.
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
764
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 7 5 6 – 7 6 5

View Article Online
33 G. D. Purvis III and R. J. Bartlett, J. Chem. Phys., 1982, 76,
K. Raghavachari, J. B. Foresman, J. Cioslowski, J. V. Ortiz, B. B.
Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R.
Gomperts, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham,
C. Y. Peng, A. Nanayakkara, C. Gonzalez, M. Challacombe,
P. M. W. Gill, B. G. Johnson, W. Chen, M. W. Wong, J. L.
Andres, M. Head-Gordon, E. S. Replogle and J. A. Pople,
GAUSSIAN 98 (Revision A.7), Gaussian, Inc., Pittsburgh, PA,
PA, 1998.
1910; K. Raghavachari, G. W. Trucks, J. A. Pople and M.
Head-Gordon, Chem. Phys. Lett., 1989, 157, 479; J. D. Watts,
J. Gauss and R. J. Bartlett, J. Chem. Phys., 1993, 98,
8718–8733.
34 G. D. Mendehall, Tetrahedron Lett., 1983, 24, 451–452.
35 M. J. Mintz and C. Walling, Org. Synth., 1973, Collect. vol. V, 184.
36 T. L. Hwang and A. J. Shaka, J. Magn. Reson. A, 1995, 112, 275–279.
37 A. L. Van Geet, Anal. Chem., 1968, 40, 2227–2229.
32 A. D. Becke, J. Chem. Phys., 1993, 98, 5648–5652.
T h i s j o u r n a l i s Q T h e O w n e r S o c i e t i e s 2 0 0 4
P h y s . C h e m . C h e m . P h y s . , 2 0 0 4 , 6 , 7 5 6 – 7 6 5
765