Photochemistry of fluorinated phenyl nitrenes: Matrix isolation of fluorinated azirines

Article abstract of DOI:10.1021/jo960093w

2,6-Difluorophenylnitrene and pentafluorophenylnitrene were generated in solid argon at 10 K by irradiation of the corresponding phenyl azides and characterized by IR and UV spectroscopy. Selective irradiation with λ = 444 nm results in the formation of the corresponding azirines, while ketene imines are not produced. On λ = 366 nm irradiation the azirines rearrange back to the nitrenes. The assignment of the azirines is confirmed by ab-initio calculations.

Full text of DOI:10.1021/jo960093w

J . Org. Chem. 1996, 61, 4351-4354
4351
P h otoch em istr y of F lu or in a ted P h en yl Nitr en es: Ma tr ix Isola tion
of F lu or in a ted Azir in es
J ens Morawietz and Wolfram Sander*
Lehrstuhl fu¨r Organische Chemie II der Ruhr-Universita¨t Bochum, 44780 Bochum, Germany
Received J anuary 16, 1996X
2,6-Difluorophenylnitrene and pentafluorophenylnitrene were generated in solid argon at 10 K by
irradiation of the corresponding phenyl azides and characterized by IR and UV spectroscopy.
Selective irradiation with λ ) 444 nm results in the formation of the corresponding azirines, while
ketene imines are not produced. On λ ) 366 nm irradiation the azirines rearrange back to the
nitrenes. The assignment of the azirines is confirmed by ab-initio calculations.
Phenylnitrene (1a ) is the prototype of the arylnitrenes
and has been the topic of many mechanistic studies.1
Several reactive intermediates have to be considered to
understand the thermo- and photochemistry of 1a .
Generally, the nitrene is produced from a precursor
moleculesmost frequently phenyl azidesin its singlet
ground state S-1a . Platz et al. demonstrated that S-1a
is a true intermediate in the photochemistry of the azide2
and that the fate of the nitrene depends very much on
the environment.3 At cryogenic temperatures S-1a pref-
erentially relaxes to the triplet ground state T-1a , which
was characterized spectroscopically by IR,4 UV,10 and
ESR6,7 spectroscopy. The singlet-triplet splitting was
determined to 18 kcal/mol.8,9 At higher temperatures in
solution rearrangement to ketene imine 2a becomes the
major path; the activation barrier (Ea) for this rearrange-
ment is 3 kcal/mol.10 In the gas phase the relaxation of
thermally excited nitrene is slow and the rearrangement
to cyanocyclopentadienesthe global energy minimum of
the C6H5N hypersurfacesbecomes feasible (Ea ) 30-51
kcal/mol).11 The IR spectrum of matrix isolated ketene
imine 2a was published first by Chapman et al.5 Later
Sheridan et al. were able to demonstrate that T-1a and
2a are formed in a photostationary equilibrium under
the conditions of matrix isolation.4
however, it has never been observed by direct spectro-
scopic methods. The azirines 3b and 6bsbut not the
nitrene 1bswere observed by IR spectroscopy during the
photolysis of 1-naphthyl azide (4b), matrix isolated in
argon at 10 K.15 Subsequent photolysis at shorter
wavelength resulted in the irreversible rearrangement
of the azirines to ketene imines 2b and 5b. This finding
clearly demonstrates the importance of substituents in
the nitrene photochemistry.
Ortho-disubstituted phenylnitrenes, such as 2,6-dimeth-
ylphenylnitrene (1c16) and pentafluorophenylnitrene (1d),17
were reported by Dunkin et al. to rearrange only very
inefficiently, if at all, during matrix photolysis. This lack
(6) Yager, W. A.; Wasserman, E.; Cramer, R. M. R. J . Chem. Phys.
1962, 37, 1148-1149.
(7) Chapman, O. L.; Sheridan, R. S.; LeRoux, J . P. J . Am. Chem.
Soc. 1978, 100, 6245-6247.
(8) Hrovat, D. A.; Waali, E. E.; Borden, W. T. J . Am. Chem. Soc.
1992, 114, 8698-8699.
A third C6H5N isomer, azirine 3a , has been claimed
as an intermediate in early trapping experiments;12-14
(9) McDonald, R. N.; Davidson, S. J . J . Am. Chem. Soc. 1993, 115,
10857-10862.
(10) Leyva, E.; Platz, M. S.; Persy, G.; Wirz, J . J . Am. Chem. Soc.
1986, 108, 3783-3790.
(11) (a) Wentrup, C. Tetrahedron 1974, 30, 1301-1311. (b) Wentrup,
C.; Crow, W. D. Tetrahedron 1970, 26, 3965-3981.
(12) Huisgen, R.; Vossius, D.; Appl, M. Chem. Ber. 1958, 91, 1-12.
(13) Doering, W. von E.; Odum, R. A. Tetrahedron 1966, 22, 81.
(14) Huisgen, R.; Appl, M. Chem. Ber. 1958, 91, 12.
(15) Dunkin, I. R.; Thomson, P. C. P. J . Chem. Soc., Chem. Commun.
1980, 499-501.
X Abstract published in Advance ACS Abstracts, May 15, 1996.
(1) For recent reviews see: (a) Platz, M. S.; Leyva, E.; Haider, K.
Org. Photochem. 1991, 11, 367-429. (b) Platz, M. S. Acc. Chem. Res.
1995, 28, 487-492.
(2) Marcinek, A.; Leyva, E.; Whitt, D.; Platz, M. S. J . Am. Chem.
Soc. 1993, 115, 8609-8612.
(3) Cullin, D. W.; Soundararajan, N.; Platz, M. S.; Miller, T. A. J .
Phys. Chem. 1990, 94, 8890-8896.
(4) Hayes, J . C.; Sheridan, R. S. J . Am. Chem. Soc. 1990, 112, 5879-
5881.
(16) Dunkin, I. R.; Donnelly, T.; Lockhart, T. S. Tetrahedron Lett.
1985, 26, 359-362.
(17) Dunkin, I. R.; Thomson, P. C. P. J . Chem. Soc., Chem. Commun.
1982, 1192-1193.
(5) McMahon, R. J .; Abelt, C. J .; Chapman, O. L.; J ohnson, J . W.;
Kreil, C. L.; LeRoux, J . P.; Mooring, A. M.; West, P. R. J . Am. Chem.
Soc. 1987, 109, 2456-2469.
S0022-3263(96)00093-X CCC: $12.00 © 1996 American Chemical Society

4352 J . Org. Chem., Vol. 61, No. 13, 1996
Morawietz and Sander
Ta ble 1. In fr a r ed Da ta of 2,6-Diflu or op h en yln itr en e
(1e), Ma tr ix Isola ted in Ar a t 10 K
ν (cm-1
)
Ia
ν (cm-1
)
Ia
ν (cm-1
)
Ia
1585.1
1577.7
1459.4
1452.2
1411.2
1312.4
1281.1
0.34
0.59
0.28
1
0.14
0.27
0.13
1242.0
1229.3
1002.8
991.6
850.4
836.6
772.4
0.44
0.36
0.09
0.63
0.11
0.08
0.13
767.7
708.3
699.6
553.2
501.2
0.33
0.13
0.09
0.08
0.05
a
Relative intensities.
of intramolecular reactivity was attributed to the steric
hindrance of the ortho positions in 1c and 1d . This is in
line with the conclusion of Young and Platz from product
studies that for S-1d at -78 °C ISC to T-1d is the
primary process, while at room temperature ring expan-
sion to the ketene imine 2d is predominant.18
F igu r e 1. 1. Infrared spectrum of 7-aza-1,5-difluoro-2,4,6-
bicyclo[4.1.0]heptatriene (3e), matrix isolated in Ar at 10 K.
Bottom: difference spectrum showing the photochemical
conversion of 1e to 3e. Irradiation (λ ) 444 nm) of 1e, matrix
isolated in Ar at 10 K. Absorptions assigned to 1e decrease
and absorptions assigned to 3e appear. Top: vibrational
spectrum of 3e calculated at the RMP2/6-31G(d) level of
theory. The band positions are scaled with 0.95, and a Lorentz
line shape with 2 cm-1 at half height is assumed.
Here we report on the matrix photochemistry of the
fluorinated phenylnitrenes 1d and 1e, which on irradia-
tion reversibly rearrange to the corresponding azirines
3d and 3c in high yields.
Resu lts a n d Discu ssion
2,6-Diflu or op h en yln itr en e (1e). Irradiation of ma-
trix-isolated 2,6-difluorophenyl azide (4e)19 in Ar at 10
K at λ > 280 nm results in the rapid decomposition of
the azide. In the IR spectrum a new set of bands appears
with intense absorptions at 1578, 1452, 1242, and 768
cm-1, which we assign to 2,6-difluorophenylnitrene (1e)
(Table 1, Figure 1). The UV spectrum of nitrene 1e in
argon at 10 K exhibits maxima at 312, 284, and 236 nm
(Figure 2) and is in accordance with literature data (EPA
glass, 77 K).10 Prolonged irradiation at λ ) 444 nm
(irradiation into the long-wavelength tail of the 312 nm
absorption or weak absorption which is not detected in
the vis spectra) slowly leads to the appearance of new
medium to strong IR absorptions at 1679, 1610, 1259,
and 1180 cm-1 (Table 2, Figure 1). On the basis of
comparison with IR spectra of other azirines15,20,21 and
on an ab-initio calculations, the newly formed compound
is assigned the structure of 7-aza-1,5-difluoro-2,4,6-
bicyclo[4.1.0]heptatriene (3e) and the IR absorption at
1679 cm-1 to the CdN stretching vibration νCdN (vide
infra). There were no strong absorptions in the 1890
cm-1 or 2200 cm-1 region, indicative for cyclic ketene
imines 2 or cyanocyclopentadienes, respectively. In the
UV region azirine 3e absorbs at 252 nm, but not around
350 nm (broad maximum), as frequently observed in
ketene imines 2.22 On 366 nm irradiation the 1e f 3e
rearrangement is reversible and the nitrene partially
recovered.
F igu r e 2. 2. UV-vis spectra showing the reversible conver-
sion of 1e to 3e. A: 2,6-Difluorophenylnitrene (1e), matrix
isolated in Ar at 10 K. B: Spectrum after irradiation with λ )
444 nm. C: Spectrum after prolonged irradiation with λ ) 366
nm.
In matrix-isolated 2H-azirine νCdN is observed at 1669
cm-1 20 in 2-phenylazirine at 1755 cm-1 21
, in 3b at 1708
,
(18) Young, M. J . T.; Platz, M. S. J . Org. Chem. 1991, 56, 6403-
6406.
cm-1, and in 6b at 1730 cm-1 (Scheme 1).15 This reveals
that substitution shifts νCdN to higher frequencies, while
the increased strain in 3b and 6b results in a red-shift
compared to the phenylazirine. In 3e the CdN stretching
frequency of 1679 cm-1 is close to that of the parent 2H-
azirine.
(19) Kanakarajan, K.; Haider, K.; Czarnik, A. W. Synthesis 1988,
566-568.
(20) Maier, G.; Schmidt, C.; Reisenauer, H. P.; Endlein, E.; Becker,
D.; Eckwert, J .; Andes, B.; Hess, J r.; Schaad, L. J . Chem. Ber. 1993,
126(10), 2337-2352.
(21) Chapman, O. L.; Le Roux, J . P. J . Am. Chem. Soc. 1978, 100,
282-285.
The calculated (RMP2/6-31G(d) level of theory) IR
spectrum is in qualitative agreement with the experi-
(22) Poe, R.; Schnapp, K.; Young, M. J . T.; Grayzar, J .; Platz, M. S.
J . Am. Chem. Soc. 1992, 114, 5054-5067.

Matrix Isolation of Fluorinated Azirines
J . Org. Chem., Vol. 61, No. 13, 1996 4353
Ta ble 2. In fr a r ed Da ta of
7-Aza -1,5-d iflu or o-2,4,6-bicyclo[4.1.0]h ep ta tr ien e (3e),
Ma tr ix Isola ted in Ar a t 10 K a n d Ca lcu la ted a t th e
RMP 2/6-31G(d ) Level
Ta ble 3. In fr a r ed Da ta of P en ta flu or op h en yln itr en e
(1d ), Ma tr ix Isola ted in Ar a t 10 K
ν (cm-1
)
Ia
ν (cm-1
)
Ia
ν (cm-1
)
Ia
1572.4
1564.4
1502.7
1464.0
1455.6
1359.4
0.12
0.14
1
0.42
0.58
0.33
1334.6
1281.4
1205.5
1151.7
1025.5
1008.5
0.08
0.15
0.08
0.05
0.59
0.23
1000.6
982.3
916.1
610.7
0.15
0.84
0.05
0.03
matrixa
calcdb
Ic
vib
ν (cm-1
)
Ic
ν (cm-1
)
assignmentd
30
29
28
27
26
25
24
23
22
21
20
19
18
17
16
15
14
13
12
11
10
9
3095
3090
3076
1637
1574
1483
1403
1307
1250
1187
1177
1123
1002
913
888
873
804
771
719
668
588
543
0.02
0.04
0.03
0.64
0.85
0.29
0.38
0.38
0.38
1
0.32
0.67
0.06
0.24
0.05
0.1
0.24
0.36
0.23
0.13
0.03
0.01
0.02
0.04
0.01
0.05
CH str
CH str
CH str
CdC,CdN str
CdC str
CdC str
c
a
1679.3
1609.8
1519.7
1412.2
1318.0
1258.9
1179.7
0.28
1
0.25
0.5
0.45
0.75
0.65
Relative intensities.
Ta ble 4. In fr a r ed Da ta of
P er flu or o-7-a za -2,4,6-bicyclo[4.1.0]h ep ta tr ien e (3d ),
Ma tr ix Isola ted in Ar a t 10 K
C-C str
C-F str
ν (cm-1
)
Ia
ν (cm-1
)
Ia
ν (cm-1
)
Ia
C-F str
c
1664.2
1589.3
1395.3
1374.5
1308.7
0.18
0.22
0.45
0.32
0.17
1222.6
1187.8
1095.3
1007.6
947.5
0.12
0.18
0.2
1
0.18
937.3
816.0
660.6
524.8
0.14
0.08
0.03
0.01
1138.3
1006.7
939.2
896.1
856.9
796.0
737.4
688.2
0.45
0.18
0.33
0.1
0.18
0.28
0.23
0.15
6-ring ip
6-ring ip
6-ring oop
CH oop
3-ring ip
c
c
c
a
Relative intensities.
of the IR spectrum with the data published by Dunkin
et al.17 (Table 3). Photolysis with λ ) 444 nm slowly
leads to a decrease of the IR absorptions of the nitrene
and to the appearance of new peaks of medium to strong
intensity at 1664, 1395, and 1008 cm-1 (Table 4). This
photoreaction is reversible on λ ) 366 nm irradiation.
We conclude from the similarities in the photochemistry
of 2,6-difluoro- (1e) and pentafluorophenylnitrene (1d )
that perfluoro-7-aza-2,4,6-bicyclo[4.1.0]heptatriene (3d )
is formed. The absence of 3d in the earlier experiments
is easily explained by the long-wavelength irradiation
required for the formation of 3d . Under the conditions
necessary to cleave the azide 4d the photostationary
equilibrium of 1d and 3d is far on the side of the nitrene.
3-ring ip
608.5
557.1
535.7
500.3
0.03
0.03
0.03
0.08
c
c
c
c
c
c
8
7
6
5
511
484
423
409
430.0
0.15
a
3e, generated by irradiation (λ ) 444 nm, argon, 10 K) of 4e.
b
Calculated at the RMP2/6-31G(d) level, frequencies scaled by
0.95. c Relative intensities. Approximate description. The as-
signment of experimental to calculated vibrations is tentative and
based on band positions and relative intensities: ip, in plane
deformation; oop, out of plane deformation; c, combination.
d
Sch em e 1
Con clu sion
Fluorinated phenylnitrenes 1d and 1e are not as
photochemically inert as previously anticipated. The
main difference to the photochemistry of other phenylni-
trenes is that the cyclic ketene imines 2 are not formed
in the low-temperature photolysis. This does not rule out
that 2 might be thermally formed at higher tempera-
tures, as was observed by Platz et al.18,22 in trapping
experiments and by time-resolved spectroscopy. These
authors argue that ring expansion proceeds through the
π2 configuration of the nitrene.1b Since this configuration
is destabilized by the fluorination, ring expansion be-
comes less favorable compared to other reactions of the
nitrene. The formation of the azirine can be looked upon
as an electrophilic attack of the nitrene center at a formal
double bond of the phenyl ring. Thus, a similar reasoning
would explain why fluorinated nitrenes are thermally
stable toward formation of azirines 3. In our experiment
the formation of 3 is a photochemical process, and the
energy of a 444 nm photon (64.4 kcal/mol) should be
enough to generate the π2 nitrene.
The question remains why azirines 3d and 3e do not
rearrange to the corresponding ketene imines 2. The
yield of intermediates obtained during matrix photolysis
depends very much on the irradiation conditions (wave-
length and bandwidth of the light source), since in many
cases photostationary equilibria of the intermediates are
formed. If the photochemistry is reversible, it is difficult
to determine the primary photo product in a sequence of
photochemical steps. Thus, broad band UV photolysissas
mental spectrum (Figure 1, Table 2). If the usual scaling
of the calculated frequencies with 0.95 is used, νCdN is
found at 1637 cm-1, at significantly lower frequency than
the experimental value. Since the RMP2 frequency of
the CdN stretching vibration of 2H-azirine is calculated
at 1597 cm-1 (scaled by 0.95), it becomes clear that RMP2
theory is not adequate to correctly predict νCdN
.
Further ab-initio calculations on 3-fluoro-2H-azirine
and azirine 3a indicate that a fluoro substituent in the
3-position results in a red shift of the CdN-stretching
mode of approximately 40-60 cm-1. This nicely explains
the red shift of 39 and 61 cm-1 in 3e compared to 3b and
6b, respectively.
P en ta flu or op h en yln itr en e (1d ). The photochemis-
try of pentafluorophenyl azide (4d ) is very similar to that
of 4e. Irradiation of 4d ,22 matrix isolated in argon at 10
K, with λ > 311 nm results in the formation of pentafluo-
rophenylnitrene 1d , which was identified by comparison

4354 J . Org. Chem., Vol. 61, No. 13, 1996
Morawietz and Sander
by using a Bruker IFS66 FTIR spectrometer with a standard
it was used in the original experimentssof 4a leads to
ketene imine 2a 21 and of 4d to nitrene 1d ,17 respectively,
and selective narrow band irradiation is required to
obtain 1a 4 or 3d .
resolution of 1 cm-1 in the range 400-4000 cm-1
. UV/vis
spectra were recorded on a Hewlett-Packard 8452A diode array
spectrophotometer with a resolution of 2 nm. Irradiations
were carried out with use of Osram HBO 500 W/2 or Ushio
USH-508SA mercury high-pressure arc lamps in Oriel hous-
ings equipped with quartz optics. IR irradiation from the
lamps was absorbed by a 10-cm path of water. For broad-
band irradiation Schott cutoff filters were used (50% transmis-
sion at the wavelength specified). For narrow-band irradiation
interference filters in combination with dichroic mirrors (“cold
mirrors”) were used.
1H-NMR spectra were taken on a Bruker AM 400. Mass
spectra (EI 70eV) were taken on a Finnigan MAT 8430.
2,6-Diflu or op h en yl a zid e (4e) was prepared from 2,6-
difluoroaniline according to a literature procedure:19 1H-NMR
(400 MHz, CDCl3) δ 6.88-6.95 (m, 2H), 7.00-7.08 (m, 1H);
IR (Ar, 10 K) 2160.4 (0.07), 2144.5 (0.07), 2123.3 (1), 1594.0
(0.03), 1502.1 (0.14), 1487.7 (0.04), 1482.7 (0.08), 1474.5 (0.03),
1337.9 (0.05), 1330.4 (0.11), 1297.9 (0.04), 1279.6 (0.04), 1246.9
(0.05), 1154.4 (0.02), 1128.8 (0.02), 1064.1 (0.01), 1050.7 (0.01),
1030.1 (0.01), 1023.0 (0.19), 774.6 (0.06), 723.2 (0.01), 705.4
(0.02), 614.3 (0.02), 520.4 (0.01), 504.8 (0.01) cm-1 (rel inten-
sity); UV/vis (Ar, 10 K) λmax 278 (0.19), 248 (0.92) nm (A); MS
m/z 155 (M+), 127.
The ab-initio calculated relative energies at the RMP2/
6-31G(d) + ZPE level of theory favor the azirine 3e with
1.2 kcal/mol relative to the cyclic ketene imine 2e. This
is the reverse ordering compared to the unsubstituted
system. Here, azirine 3a is 0.2 kcal/mol higher in energy
than 2a . However, these differences in the relative
stabilities are in the order of the error of the ab-initio
calculations and should not be overinterpreted. More
experimental and theoretical work has to be done to
explain the effect of fluorine substitution on the thermo-
and photochemistry of phenylnitrenes.
Exp er im en ta l Section
Ca lcu la tion s. The geometries of 2a , 2e, 3a , 3e, 2H-azirine,
and 3-fluoro-2H-azirine have been optimized at the RMP2/6-
31G(d) level of theory. The vibrational frequencies and
intensities of 2e, 3a , 3e, 2H-azirine, and 3-fluoro-2H-azirine
have been calculated at the RMP2/6-31G(d) level. The ab-
initio calculations were carried out using standard basis sets
and the program package Gaussian 9223 on SGI and IBM
workstation computers.
P en ta flu or op h en yl a zid e (4d ) was prepared from pen-
tafluorophenyl hydrazine according to a literature procedure:
22
IR (Ar, 10 K) 2214.0 (0.05), 2197.9 (0.09), 2148.4(0.12),
Ma tr ix Sp ectr oscop y. Matrix isolation experiments were
performed by standard techniques with an APD DE-204SL and
an APD DE-202 Displex closed cycle helium cryostat. Matrices
were produced by deposition of argon (Linde, 99.9999%) on
top of a CsI (IR) or sapphire (UV/vis) window with a rate of
approximately 0.15 mmol/min. Infrared spectra were recorded
2124.4 (0.97), 1529.5 (0.42), 1521.5 (1), 1516.8 (0.87), 1474.7
(0.08), 1354.8 (0.05), 1332.3 (0.08), 1251.8 (0.12), 1107.7 (0.31),
1086.6 (0.03), 1027.8 (0.16), 1014.4 (0.19), 1002.3 (0.23), 944.0
(0.32), 925.5 (0.03), 804.8 (0.05), 672.9 (0.02) cm-1 (rel inten-
sity); UV/vis (Ar, 10 K) λmax 246 (0.82) nm (A); MS m/z 209
(M+), 181.
(23) Frisch, M. J .; Trucks, G. W.; Head-Gordon, M.; Gill, P. M. W.;
Wong, M. W.; Foresman, J . B.; J ohnson, B. G.; Schlegel, H. B.; Robb,
M. A.; Replogle, E. S.; Gomperts, R.; Andres, J . L.; Raghavachari, K.;
Binkley, J . S.; Gonzalez, C.; Martin, R. L.; Fox, D. J .; Defrees, D. J .;
Baker, J .; Stewart, J . J . P.; Pople, J . A. Gaussian 92, Revision C;
Gaussian, Inc.: Pittsburgh, PA, 1992.
Ack n ow led gm en t. This work was financially sup-
ported by the Deutsche Forschungsgemeinschaft and
the Fonds der Chemischen Industrie.
J O960093W