Matrix isolation of two isomers of N4CO

Article abstract of DOI:10.1002/anie.201005177

A close analogue to diazide: Azido isocyanate, N₃-NCO, is formed by a stepwise photo-induced decomposition of carbonyl diazide, CO(N ₃)₂, and characterized by matrix IR, UV/Vis spectroscopy, and DFT calculations (see scheme). The azido carbonyl nitrene intermediate N₃C(O)N has also been characterized.

Full text of DOI:10.1002/anie.201005177

Communications
DOI: 10.1002/anie.201005177
Nitrogen-Rich Compounds
Matrix Isolation of Two Isomers of N4CO**
Xiaoqing Zeng, Helmut Beckers, and Helge Willner*
Inorganic molecules containing polynitrogen chains are
known as intrinsically unstable high-energetic compounds.
Tetranitrogen (N4) was merely detected as a gaseous meta-
stable molecule with a lifetime of microseconds by neutral-
ization–reionization mass spectrometry,[1] and only a few
derivatives of N4 with open-chain structures—N4O,[2] N4O2,[3]
and [N3NFO][SbF6][4]—have been spectroscopically identi-
fied. A catenated pentanitrogen unit is found in the N5+ ion.[5]
Its structure and chemistry was recently explored by Christe
et al., and the N5 unit exhibits a bent structure with C2v
symmetry as shown by X-ray crystallography.
Concerning the hexanitrogen compounds, numerous the-
oretical calculations have been devoted to the two isomers of
N6, diazide and hexazine, but none of them were identified
experimentally up to now.[6] To the contrary, its isoelectronic
analogue, diisocyanate (OCN-NCO), was prepared by pho-
tolysis and flash pyrolysis of oxalic acid diazide, O2C2(N3)2,
and characterized by matrix IR spectroscopy.[7] Only the C2h
symmetrical conformer was computationally found to be a
minimum,[8] and the initially formed nitrene intermediate was
not observed. Our recent synthesis of pure carbonyl diazide,
OC(N3)2,[9] encouraged us to conduct a similar experiment in
seeking azido isocyanate, N3-NCO, an even closer analogue to
N6.
Gaseous OC(N3)2 exhibits two broad absorptions at lmax
=
232 and 198 nm. Irradiation of Ar matrix-isolated OC(N3)2
with UV light (l = 255 nm, interference filter) was performed,
and the appearance of new IR bands at the expense of those
of OC(N3)2 was observed. Formation of CO was evidenced by
a weak band at 2140.6 cmꢀ1 (Figure S1 in the Supporting
Information), and strong new bands in the region of N3 and
NCO stretching modes appeared upon photolysis. To distin-
guish the different carriers of these bands, subsequent
irradiation of the matrix with visible light (l ꢁ 455 nm) was
performed. No decomposition of OC(N3)2 was observed
under these conditions, and the IR difference spectrum before
and after l ꢁ 455 nm photolysis is shown in Figure 1b. The
good agreement between the observed and calculated spectra
(B3LYP/6-311 + G(3df), using a scaling factor of 0.9679)[10]
for triplet N3C(O)N (Figure 1a, Table 1) suggested its for-
mation upon UV photolysis (l = 255 nm) of OC(N3)2. The
Figure 1. a,c) Calculated IR spectra of a) anti-N3C(O)N (3A’’) and c) N3-
NCO (B3LYP/6-311+G(3df), a scaling factor of 0.9679 was applied).
b) IR difference spectrum recorded before and after visible light
irradiation (lꢁ455 nm) of the UV photolysis (l=255 nm) products of
Ar matrix-isolated OC(N3)2. d) IR difference spectrum recorded before
and after subsequent UV/Vis irradiation (lꢁ335 nm) of (b). Features
marked with asterisks are due to different matrix site occupancies of
OC(N3)2.
nitrene N3C(O)N was depleted upon visible irradiation (l ꢁ
455 nm, Figure 1b).
In principle, two nitrene conformers can be formed which
adopt a syn and anti orientation of the azide group with
ꢀ
respect to the C N bond. In fact, weak satellites at
1602.6 cmꢀ1 (Ar matrix) and 1602.9 cmꢀ1 (Ne matrix) of the
[*] Dr. X. Zeng, Dr. H. Beckers, Prof. Dr. H. Willner
Bergische Universitꢀt Wuppertal, FB C–Anorganische Chemie
42097 Wuppertal (Germany)
ꢀ1
=
strong bands for the C O stretching vibration at 1581.7 cm
(Ar matrix) and 1582.5 cmꢀ1 (Ne matrix) might suggest the
formation of a second nitrene conformer. However, the
absence of any such features for the other nitrene bands
indicates the presence of only one conformer. The preference
for anti-N3C(O)N (3A’’) comes from various DFT calculations
(Figure S2 and Table S2), which predict this conformer to be
energetically slightly more stable than the syn conformer. The
Fax: (+49)202-439-3053
E-mail: willner@uni-wuppertal.de
[**] This work was supported by the Deutsche Forschungsgemeinschaft
and the Fonds der Chemischen Industrie. X.Z. acknowledges a
fellowship from the Alexander von Humboldt Foundation.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201005177.
482
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 482 –485

Table 1: Calculated and experimental IR wavenumbers (cmꢀ1) of N3C(O)N.
Calculated[a]
Experimental[b]
Assignment[c]
syn
anti
Ar matrix
3A’’
Ne matrix
anti
3A’’
1A’
3A’’
1A’
3A’’
2197 (509)
1551 (264)
1257 (362)
1101 (72)
821 (19)
614 (30)
607 (28)
2224 (429)
1698 (445)
1327 (123)
1247 (147)
884 (16)
626 (15)
582 (25)
2212 (412)
1558 (193)
1236(454)
1138 (94)
871 (17)
652 (42)
643 (37)
2228 (426)
1727 (486)
1321 (65)
1227 (157)
889 (7)
2159.7 (55)
1581.7 (35)
1209.1 (100)
1138.5 (9)
894.6 (6)
664.8 (17)
646.5 (19)
2159.9 (68)
1582.5 (47)
1204.0 (100)
1138.8 (11)
897.6 (7)
n1, n(N3)asym
n2, n(CO)
n3, n(N3)sym
n4, n(CN)
ꢀ
n5, n(C N3)
641 (12)
582 (19)
662.7 (11)
651.8 (14)
n6, d(N3)ip
n10, d(N2CO)oop
[a] Calculated (B3LYP/6-311+G(3df)) frequencies (scaled by 0.9679). Absolute intensities [kmmolꢀ1] are given in parenthesis. A complete list of all
the calculated IR fundamentals of singlet and triplet N3C(O)N species is given in Table S1. Syn/anti orientation of the azide group with respect to the
ꢀ
C N bond (Figure S2). [b] Band positions at the most intense matrix sites with relative integrated intensities [%] in parenthesis. [c] Assignments
correspond to the more stable anti conformer of N3C(O)N (3A’’).
anti preference is also consistent with Scheme S1 that
assumed the favorable formation of anti-N3C(O)N from the
more abundant syn–syn conformer of OC(N3)2.[9]
B3LYP/6-311 + G(3df) method. This band shows a regular
average vibrational spacing of 466 cmꢀ1 (Table S4), which
resembles those of triplet carbonyl nitrene FC(O)N at lmax
=
Further experiments were performed using 15N-labeled
OC(N3)2 which was prepared from 1-15N sodium azide. Three
isotopomers of doubly 15N-labeled OC(N3)2 were identified in
the matrix spectrum, from which four isotopologues of
N3C(O)N with equal molar ratio were expected to be
formed upon photolysis (Scheme S2). The expected doublet-
like splitting with nearly equal intensity was clearly observed
for the N3 vibrational modes n1, n3, and n5 in the 15N-labeling
experiment (Figure S3). An 15N-isotopic shift was also
observed for the CO stretching mode (n2), indicating vibra-
tional coupling to the N3 stretching modes (Table S3).
572 nm (Dn˜ = 465 cmꢀ1).[11] The strongest transition for
N3C(O)N (3A’’) is predicted at 484 nm (f = 0.0208), consistent
with the observation of the strong broad band at lmax
=
430 nm (Figure 2). In addition, three weak transitions
observed at 344, 325, and 308 nm are also assigned to
N3C(O)N (3A’’) (Table S4) based on their photolytic behav-
ior. During the photolysis of the diazide, NCO radicals were
also formed as evidenced by their characteristic visible
!
absorptions at 438 nm (0 0) and the vibrational progression
at 414, 398, and 383 nm (Figure 2c).[11]
Once the formation of N3C(O)N (3A’’) from OC(N3)2 by
UV photolysis (l = 255 nm) was ascertained, it may photo-
isomerize to the title compound N3-NCO through Curtius-
type rearrangement. This is indeed proved by the appearance
of new IR bands upon visible irradiation (l ꢁ 455 nm) of the
matrices containing N3C(O)N (3A’’) (Table 1). Experimental
IR band positions are compared to calculated ones for N3-
The UV photolysis (l = 255 nm) products of Ne matrix-
isolated OC(N3)2 exhibits several absorptions below 800 nm
(Figure 2). Upon subsequent visible photolysis (l ꢁ 455 nm),
the major broad band at 430 nm and further weak features at
around 660 nm and 330 nm disappeared simultaneously, while
two additional bands at 438 and 272 nm increased. This
photolysis behavior indicates that these transitions are due to
different species, and the former bands (660, 450, and 330 nm)
NCO in Table 2. Two strong bands at 2220.8 (n˜calcd
=
2244 cmꢀ1) and 2099.1 cmꢀ1 (n˜calcd = 2141 cmꢀ1) were assigned
to the asymmetric stretches of the NCO and N3 moieties,
respectively. The first band split into three components in 15N-
labeling experiments (Figure S4) with 15N-isotopic shifts of
were assigned to N3C(O)N. The weak absorption at lmax
=
660 nm for N3C(O)N (3A’’) is consistent with a calculated
vertical transition at 740 nm (f = 0.0004) using the TD-DFT
Table 2: Calculated and experimental IR wavenumbers [cmꢀ1] of N3-
NCO.
Calculated[a]
Experimental[b]
Assignment
Ar matrix
Ne matrix
2267.8 (33)
2220.8 (79)
2099.1 (100)
2266.7 (38)
2223.8 (74)
2102.2 (100)
n3 + n5
2244 (810)
2141 (1066)
1380 (3)
1197 (39)
861 (5)
n1, n(NCO)asym
n2, n(N3)asym
n3, n(NCO)sym
n4, n(N3)sym
1160.7 (4)
663.7 (6)
1165.9 (4)
668.5 (5)
ꢀ
n5, n(N N)
652 (63)
603 (2)
n6, d(NCO)ip
n7, d(N3)ip
[a] Calculated (B3LYP/6-311+G(3df)) frequencies (scaled by 0.9679).
Absolute intensities [kmmolꢀ1] are given in parenthesis. A list of all the
calculated IR data of N3-NCO is given in Table S6. [b] Band positions of
the most intense matrix sites with relative integrated intensities [%] in
parenthesis.
Figure 2. UV/Vis spectrum of Ne matrix-isolated OC(N3)2: a) before
photolysis, b) after UV photolysis (l=255 nm), c) after subsequent Vis
photolysis (lꢁ455 nm).
Angew. Chem. Int. Ed. 2011, 50, 482 –485
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
483

Communications
1.0, 8.6, and 14.3 cmꢀ1, while the second band shows two
theoretically using various DFT methods (Tables S2, S6, and
S7). Only the former one was observed experimentally, and
the nitrene–isocyanate rearrangement was found to be
exothermic by 157.3 kJmolꢀ1 at the B3LYP/6-311 + G(3df)
level of theory (Table S1). N3-OCN was found to be higher in
energy than N3-NCO by 173.4 kJmolꢀ1 at the same level of
theory. The calculated structures of the interpseudohalogens
N3-NCO (Figure 3) and N3-OCN are similar to those of the
components with 15N-isotopic shifts of 7.6 and 23.7 cmꢀ1
(Table S5). Additional two weak bands at 1160.7 and
663.7 cmꢀ1 were assigned to the symmetric stretching mode
of the N3 group and the in-plane bending mode of NCO
ꢀ
group, respectively. The stretching mode of the N N bond
linking the two pseudo halogen moieties (n˜calcd = 861 cmꢀ1)
was predicted to be very weak (Table 2) and has not been
observed. A strong band at 2267.8 cmꢀ1 showed a similar
photolytic behavior, indicating that the carrier of this band is
also N3-NCO. It is tentatively assigned to the (n3 + n5)
combination band, which probably gains intensity through a
Fermi resonance with n1 (2220.8 cmꢀ1). Attempts to accumu-
late N3-NCO in the cryogenic matrix by prolonged photolysis
failed, because it decomposed to form carbon monoxide when
exposed to UV light of l = 255 nm.
Figure 3. Calculated structure of N3-NCO at the B3LYP/6-311+G(3df)
level of theory. Bond lengths and angles are given in [ꢁ] and [deg],
respectively.
Upon irradiation of matrix-isolated N3-NCO with UV/Vis
light (l ꢁ 335 nm), CO was formed as the only IR detectable
species (Figure 1c, 2140.6 cmꢀ1). Its photolytic behavior
implies an electronic band of N3-NCO above 335 nm. This
was confirmed by TD-DFT calculation, which predicted a
weak transition at 366 nm (f = 0.0004). An additional absorp-
tion at 272 nm can be assigned to N3-NCO, whose intensity
increased with visible irradiation (l ꢁ 455 nm) of N3C(O)N
(3A’’), and decreased after subsequent UV/Vis irradiation
(l ꢁ 335 nm). It corresponds to a predicted vertical transition
at 279 nm (f = 0.0005) for N3-NCO.
isoelectronic analogues N6 and OCN-NCO[8a] concerning
[6c]
the anti arrangement of their two pseudohalogen moieties
ꢀ
with respect to the central N N bond, which minimize their
mutual repulsion. At the B3LYP/6-311 + G(3df) level of
ꢀ
theory the predicted length of the weak N N bond in N3-
NCO is 1.404 ꢀ, which is between those predicted for OCN-
NCO (1.372 ꢀ) and the experimentally unknown diazide N3-
N3 (1.439 ꢀ) at the same level. The latter bond is, however,
The photochemistry studied herein is summarized in
Scheme 1. Starting from OC(N3)2 two isomers of N4CO were
isolated, the azido carbonyl nitrene, N3C(O)N, which re-
ꢀ
still considerably shorter than the corresponding N O bond
ꢀ
of N3-OCN (1.501 ꢀ). Even if formed, the very weak N O
bond in N3-OCN might facilitate its rearrangement to the
more stable N3-NCO under the experimental conditions.
Experimental Section
Scheme 1. Photochemistry of OC(N3)2 in Ar matrices at 16 K.
Caution! Carbonyl diazide, OC(N3)2, was found to be an extremely
explosive and shock sensitive compound in liquid and solid state.
Although we did not experience any explosions during this work, safety
precautions must be taken, including face shields, leather gloves, and
protective leather clothing, particularly in the case of handling pure
OC(N3)2 in solid and liquid state.
arranged upon visible light irradiation (l ꢁ 455 nm) through a
Curtius-type rearrangement to azido isocyanate, N3-NCO.
Further loss of N2 by near-UV irradiation yielded CO and N2.
No other intermediates, for example, the long-sought diazir-
inone, N2CO,[12] or its isomers,[13] were observed under our
experimental conditions.
The novel carbonyl nitrene was proved to adopt a triplet
ground state by its characteristic IR and UV spectra.
However, according to CBS-QB3 calculations (Table S1),
the calculated singlet–triplet energy gap is rather small (DES–T
< 20 kJmolꢀ1). It is worth to mention that alkyl and aryl
carbonyl nitrenes usually have closed-shell singlet ground
states.[14] Such singlet carbonyl nitrenes shows structural and
spectroscopic properties which are very different from those
of the triplet species, as the former ones have much shorter,
Sample preparation: Carbonyl diazide, OC(N3)2, was prepared
from FC(O)Cl and NaN3 according to literature procedure[9] and
purified by repeated fractional condensation in vacuum. For the
preparation of 15N-labeled OC(N3)2, 1-15N sodium azide (98 at% 15N,
EURISO-TOP GmbH) was used. The purity of the sample was
verified by FT-IR spectroscopy.
Matrix isolation and photolysis: Matrix IR spectra were recorded
on a FT-IR spectrometer (IFS 66v/S Bruker) in reflectance mode
using a transfer optic. A KBr beam splitter and an MCT detector were
used in the region of 5000 to 550 cmꢀ1. For each spectrum 200 scans at
a resolution of 0.25 cmꢀ1 were co-added. The gaseous sample was
mixed by passing the argon gas through a glass U-trap containing ca.
10 mg of OC(N3)2, which was kept in an ethanol bath at a temperature
of ꢀ658C. By adjusting the flow rate of Ar or Ne (2 mmolhꢀ1), a small
amount of the resulting mixture (OC(N3)2/inert gas ca. 1:1000
estimated) was deposited within 30 min onto the matrix support
(Rh-plated Cu block) at 16 K (Ar matrix) or 5 K (Ne matrix) in high
vacuum. Details of the matrix apparatus have been described
elsewhere.[15] Photolysis experiments were carried out with a high-
pressure mercury arc lamp (TQ 150, Heraeus) by conducting the light
=
double bond-like C N bonds and unusually small NCO
angles of about 908 (Figure S2).[11]
In general, the photo-rearrangement of the nitrene
N3C(O)N might give access to two different chain-like
isomers, N3-NCO and N3-OCN, which have been explored
484
www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 482 –485

through water-cooled quartz lenses and an interference filter l =
255 nm or cut off filters l ꢁ 335 and l ꢁ 455 nm (Schott).
Schaefer III, J. Am. Chem. Soc. 1983, 105, 1760 – 1764; c) M.
Tobita, R. J. Bartlett, J. Phys. Chem. A 2001, 105, 4107 – 4113.
[7] G. Maier, M. Naumann, H. P. Reisenauer, J. Eckwert, Angew.
Chem. 1996, 108, 1800 – 1801; Angew. Chem. Int. Ed. Engl. 1996,
35, 1696 – 1697.
[8] a) A. Schulz, T. M. Klapꢁtke, Inorg. Chem. 1996, 35, 4791 – 4793;
b) T. Pasinszki, Phys. Chem. Chem. Phys. 2008, 10, 1411 – 1418.
[9] X. Q. Zeng, M. Gerken, H. Beckers, H. Willner, Inorg. Chem.
2010, 49, 9694 – 9699 .
Theoretical studies: The structures of all stationary points were
fully optimized with DFT (B3LYP,[16] BP86,[17] MPW1PW91[18]) and
CBS-QB3[19] methods. The 6-311 + G(3df) basis set was used for all
the DFT calculations. Time-dependent (TD) DFT (B3LYP/6–311 +
G(3df)) calculations[20] were performed for the prediction of UV/Vis
spectra. Quantum chemical calculations were carried out using the
Gaussian03 software package.[21]
[10] M. P. Andersson, P. Uvdal, J. Phys. Chem. A 2005, 109, 2937 –
2941.
Received: August 18, 2010
[11] X. Q. Zeng, H. Beckers, H. Willner, D. Grote, W. Sander, Chem.
Eur. J., DOI: 10.1002/chem.201002955.
Revised: October 1, 2010
Published online: December 5, 2010
[12] a) R. A. Moss, G. Chu, R. R. Sauers, J. Am. Chem. Soc. 2005,
127, 2408 – 2409; b) C. J. Shaffer, B. J. Esselman, R. J. McMahon,
J. F. Stanton, R. C. Woods, J. Org. Chem. 2010, 75, 1815 – 1821.
[13] G. Maier, H. P. Reisenauer, J. Eckwert, M. Naumann, M. D.
Marco, Angew. Chem. 1997, 109, 1785 – 1787; Angew. Chem. Int.
Ed. Engl. 1997, 36, 1707 – 1709.
[14] a) J. Liu, J. L. Mandel, C. M. Hadad, M. S. Platz, J. Org. Chem.
2004, 69, 8583 – 8593; b) E. A. Pritchina, N. P. Gritsan, A.
Maltsev, T. Bally, T. Autrey, Y. Liu, Y. Wang, J. P. Toscano,
Phys. Chem. Chem. Phys. 2003, 5, 1010 – 1018.
Keywords: azides · matrix isolation · nitrenes ·
photoisomerization · vibrational spectroscopy
.
[1] F. Cacace, G. de Petris, A. Troiani, Science 2002, 295, 480 – 481.
[2] a) H. W. Lucien, J. Am. Chem. Soc. 1958, 80, 4458 – 4460; b) A.
Schulz, I. C. Tornieporth-Oetting, T. M. Klapꢁtke, Angew.
Chem. 1993, 105, 1697 – 1699; Angew. Chem. Int. Ed. Engl.
1993, 32, 1610 – 1612; c) J. M. Galbraith, H. F. Schaefer III, J.
Am. Chem. Soc. 1996, 118, 4860 – 4870.
[3] a) M. P. Doyle, J. J. Maciejko, S. C. Busman, J. Am. Chem. Soc.
1973, 95, 952 – 953; b) T. M. Klapꢁtke, A. Schulz, I. C. Tornie-
porth-Oetting, Chem. Ber. 1994, 127, 2181 – 2185.
[15] H. G. Schnꢁckel, H. Willner in Infrared and Raman Spectrosco-
py, Methods and Applications (Ed.: B. Schrader), VCH, Wein-
heim, 1994.
[16] A. D. Becke, J. Chem. Phys. 1993, 98, 5648 – 5652.
[17] J. P. Perdew, Phys. Rev. B 1986, 33, 8822 – 8824.
[18] C. Adamo, V. Barone, J. Chem. Phys. 1998, 108, 664 – 675.
[19] a) M. R. Nyden, G. A. Petersson, J. Chem. Phys. 1981, 75, 1843 –
1862; b) G. A. Petersson, M. A. Al-Laham, J. Chem. Phys. 1991,
94, 6081 – 6090; c) G. A. Petersson, T. Tensfeldt, J. A. Mont-
gomery, J. Chem. Phys. 1991, 94, 6091 – 6101; d) J. A. Montgom-
ery, J. W. Ochterski, G. A. Petersson, J. Chem. Phys. 1994, 101,
5900 – 5909.
[20] a) R. E. Stratmann, G. E. Scuseria, M. J. Frisch, J. Chem. Phys.
1998, 109, 8218 – 8224; b) J. B. Foresman, M. Head-Gordon, J. A.
Pople, M. J. Frisch, J. Phys. Chem. 1992, 96, 135 – 149.
[21] M. J. Frisch et al., Gaussian03 (RevisionB.05), see the Support-
ing Information.
[4] W. W. Wilson, R. Haiges, J. A. Boatz, K. O. Christe, Angew.
Chem. 2007, 119, 3083 – 3087; Angew. Chem. Int. Ed. 2007, 46,
3023 – 3027.
[5] See, for example: a) K. O. Christe, W. W. Wilson, J. A. Sheehy,
J. A. Boatz, Angew. Chem. 1999, 111, 2112 – 2118; Angew. Chem.
Int. Ed. 1999, 38, 2004 – 2009; b) R. Haiges, S. Schneider, T.
Schroer, K. O. Christe, Angew. Chem. 2004, 116, 5027 – 5032;
Angew. Chem. Int. Ed. 2004, 43, 4919 – 4924; c) D. A. Dixon, D.
Feller, K. O. Christe, W. W. Wilson, A. Vij, V. Vij, H. D. B.
Jenkins, R. M. Olson, M. S. Gordon, J. Am. Chem. Soc. 2004,
126, 834 – 843; d) A. Vij, W. W. Wilson, V. Vij, F. S. Tham, J. A.
Sheehy, K. O. Christe, J. Am. Chem. Soc. 2001, 123, 6308 – 6313.
[6] See, for example: a) H. Huber, Angew. Chem. 1982, 94, 71 – 72;
Angew. Chem. Int. Ed. Engl. 1982, 21, 64 – 65; b) P. Saxe, H. F.
Angew. Chem. Int. Ed. 2011, 50, 482 –485
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
485