New gas phase reaction to iodine azide, IN3: Microwave spectrum and structure

Article abstract of DOI:10.1016/j.molstruc.2003.12.045

The unstable molecule iodine azide has been prepared for the first time by a gas phase reaction in a flow system, and the microwave spectra of two isotopic species, a- and b-type, have been measured: the abundant species in the frequency range from 10 to 40GHz, the I¹⁴N¹⁵N ¹⁴N-species from 20 to 26GHz. Rotational constants and centrifugal distortion parameters were obtained. The molecule has been shown to be planar, and the structure was derived. Also, the I-quadrupole coupling constants have been determined.

Full text of DOI:10.1016/j.molstruc.2003.12.045

Journal of Molecular Structure 695–696 (2004) 189–202
www.elsevier.com/locate/molstruc
New gas phase reaction to iodine azide, IN3:
microwave spectrum and structure
*
H.-O. Munz, H.-K. Bodenseh , M. Ferner
¨ ¨ ¨
Fakultat fur Naturwissenschaften, Abteilung Chemische Physik, Universitat Ulm, Ulm D-89069, Germany
Received 14 October 2003; accepted 17 December 2003
Abstract
The unstable molecule iodine azide has been prepared for the first time by a gas phase reaction in a flow system, and the microwave spectra
of two isotopic species, a- and b-type, have been measured: the abundant species in the frequency range from 10 to 40 GHz, the I14N15N14N-
species from 20 to 26 GHz. Rotational constants and centrifugal distortion parameters were obtained. The molecule has been shown to be
planar, and the structure was derived. Also, the I-quadrupole coupling constants have been determined.
q 2004 Elsevier B.V. All rights reserved.
Keywords: Iodine azide; Microwave spectroscopy; Nuclear quadrupole coupling constants; Molecular structure
1. Introduction
but no other line of the rich spectrum matched these first
ones on the basis of a reasonable structure. We finally
realized that the spectrum originated from an unknown
reaction product of IN3 and a deposit in the cell from
previous measurements. In new, clean metal cells,
aluminum or brass, no lines could be found. Obviously,
the decomposition was faster than the filling of the cell at
the low temperature. We therefore, looked for another
way to produce the molecule.
IN3 has first been synthesized in 1900 by Hantzsch [1]
as a highly explosive solid. The reaction was carried out
in solution by adding iodine, dissolved in diethylether, to
a suspension of silver azide in water; but no good data
could be obtained because of poor purity. Only in 1993,
¨
Klapotke and co-workers [2] published a modified
preparation by using CFCl3 as a solvent, yielding highly
pure, bright yellow crystals, suitable for X-ray crystal-
lography. Also, ab initio calculations were carried out [2],
spectroscopic work was done with IR [3], 14N NMR [4]
and Raman [5]. Furthermore, a structure determination
by electron diffraction was published [6]. Following
2. Experimental procedures
2.1. Chemical preparation of iodine azide
¨
Klapotke’s suggestion, we tried to obtain a rotational
2.1.1. Parent molecule I14N3
spectrum, the first samples being supplied in solution
over dry ice by his group. By pumping-off the solvent
and warming up to 240 8C, we could record a
reasonable-looking microwave spectrum, showing the
typical hyperfine sextets of the iodine nucleus. We then
prepared our own crystals, yielding an identical spectrum.
We seemingly succeeded in assigning 18 transitions [7],
Taking into account the analogy between iodine azide
and the isoelectronic molecule iodine isocyanate, we tried
the same gas phase reaction as used for the production of
INCO [8] according to:
AgN3 þ I2 ! AgI þ IN3
Although, such an approach to obtain a gaseous azide had
never been reported before, the attempt was immediately
successful: iodine vapour was passed in a flow system over
silver azide at 120 8C (optimum temperature) and through
the absorption cell, maintaining a total pressure of 2–3 Pa
for survey spectra and 0.6 Pa for frequency measurements.
*
Corresponding author. Address: c/o Abteilung Quanten-
¨
informationsverarbeitung, Universitat Ulm, Ulm D-89069, Germany.
Tel.: þ49-731-502-2832; fax: þ49-731-502-2839.
E-mail address: hans-karl.bodenseh@chemie.uni-ulm.de (H.-K.
Bodenseh).
0022-2860/$ - see front matter q 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.molstruc.2003.12.045

190
H.-O. Munz et al. / Journal of Molecular Structure 695–696 (2004) 189–202
The products were trapped behind the cell by LN2. An
intense microwave spectrum was obtained which could
easily be ascribed to iodine azide and assigned in a
straightforward manner. Silver azide was precipitated by
adding an aqueous solution of silver nitrate to a solution of
sodium azide in water, filtered-off, washed with water,
ethanol, and finally with diethylether and carefully dried.
About 0.1 mol were put into a quartz boat which was placed
into a quartz tube of 4 cm in diameter and surrounded by an
oven, 30 cm in length and horizontally arranged. Such a
charge could be used for more than a week. In spite of
intense drying, it was impossible to avoid producing strong
transitions of HN3 so long as IN3 was produced from a given
sample of silver azide. These appeared even at room
temperature as soon as the iodine vapour was added,
whereas water vapour without iodine over silver azide
yielded no trace of HN3 even at 220 8C: obviously IN3 has to
be formed first and then reacts with water to make HN3.
When steady state is reached in the flow system and the cell
is then closed simultaneously at both ends, the strong signals
from IN3 decay with a rate that can be described best by a
second order law: after 10 min half intensity is reached and
after 50 min nothing more can be detected. While IN3-
crystals produced in solution cannot be cooled down to LN2-
temperature without explosion, we had no problem in
trapping the IN3-vapour at that low temperature as a
microcrystalline sublimate. Warming the trap with caution
and flowing the products back through the cell, yields at first
the spectrum of HN3 followed by that of IN3. So the IN3
produced in the gas phase can be handled much more safely
than that produced in solution. Obviously, IN3 condensed
from the gas phase forms a molecular crystal, whereas that
from solution forms layers of I–N–I–N-chains [2].
Perhaps, these chains rearrange only slowly, compared to
the rate of decomposition in the metal cells, to the monomer
molecules. This may also explain why these transitions are
about one order of magnitude weaker and not suited for
searching the unknown spectrum. As we found out by
another preparation in solution, they can be identified if
their position is known. But even if the procedures described
here appear relatively safe: the usual precautions when
handling azides should always be taken! When manipulat-
ing the substances, the whole body was at all times protected
by a thick leather apron, the eyes by safety glasses, the
hands by gloves of chain-mail, and the head by a helmet
with additional safety shield in front of the face. The glass
traps were contained in Dewar vessels made from stainless
steel. As soon as the silver azide had been placed into the
reaction tube in the oven, this part of the spectrograph was
isolated towards the operators by an acrylic glass shield,
2 cm in thickness and wrapped with several layers of
transparent foil against slivering by the impact of a shock
wave. While dealing with iodine azide prepared from
solution, several detonations occurred, but never with iodine
azide prepared from the gas phase.
2.1.2. I14N15N14N
15NH144NO3 was used as starting material. It was heated
for 20 h at 220 8C in an evacuated and sealed glass ampoule
to yield 15NNO. After cooling down this was condensed into
another, evacuated ampoule over NaNH2. Heating to 300 8C
for 22 h resulted in the formation of Naþ(14N15N14N)2
which was converted to the silver salt as described above.
Due to the limited supply of the substituted ammonium
nitrate, the quantities had to be drastically reduced to mmol
scale and the experimental setup was accordingly changed:
instead of the horizontal arrangement of oven and reaction
tube a vertical one was used with the quartz tube only 2 cm
in diameter and 15 cm in length. At about 5 cm from the
lower end of the tube a frit was inserted. Because the tube
was equipped with ground joints at both ends it could be
used instead of a filter and the precipitated silver azide was
sucked directly onto the frit, a layer of quartz wool in-
between. Washing the salt was done through the tube and
the frit, also the first drying. Final drying occurred after
insertion into the gas flow line at the spectrometer. A small
oven of 10 cm in length was placed around the reaction
tube. With this modified setup the optimum temperature for
producing IN3 increased to 160 8C. During the tests to
minimize the sample size, we realized that the reaction was
not reproducible: 300 mg (,2 mmol) silver azide some-
times allowed spectra to be taken through 3 h, but some-
times only for a few minutes; all intermediate values were
possible. We could not find out the reason; it seems to be a
surface effect of the crystals depending upon tiny variations
in the conditions of precipitation and was not important with
the large samples for the parent molecule. So we just had to
trust luck when using the 15N-sample.
2.2. Spectrometer setup
The microwave spectrograph was of conventional Stark-
type operating at 30 kHz. Because pyrolytic or discharge
reactions for producing instable species do not often reach a
real steady state and the yield of the desired product can
change within seconds, relatively fast time averaging is
necessary for good results. Therefore, the apparatus was
equipped with a self-configured and -programmed
ELTECq computer based on a MOTOROLAq 68K
processor. The original idea was to step the frequency in
small increments via a fast switching frequency standard
and to have the microwave source, a MARCONIq sweeper,
following by means of a phase locking circuit. But because
the microwave power incident at the detector diode is not
constant vs. frequency, the switching within a few ms
produces sharp rises or drops of the voltage across the diode.
The harmonics thereof reach easily into the range of the
modulation frequency and produce excessive noise in the
signal channel. Therefore, a method was chosen which
allowed analog sweeping of the microwave source and still
kept all the advantages of digital signal processing. A block
diagram with the basic elements (no attenuators etc.) of

H.-O. Munz et al. / Journal of Molecular Structure 695–696 (2004) 189–202
191
Fig. 1. Simplified block diagram of the spectrograph.
the setup is shown on Fig. 1: a pulse generator triggered a
‘leading’ scope at the front end of the spectrograph and a
monitoring scope (to observe the tuning of the MW-power)
at the far end of the cell, the repetition rate was adjustable to
match the acquisition time the computer needed for one
sweep. The time base (sawtooth) of the leading scope was
fed into a special low frequency amplifier to adjust
amplitude, offset, and polarity (for sweep reversal). The
MW-sweeper with backward wave oscillator plug-ins could
be operated in single frequency mode. In this mode, the
frequency determining high voltage was especially stabil-
ized, yielding high stability of the output frequency with
only small and slow drifts. An FM-input (bandwidth DC-
5 MHz) allowed a modulating voltage to be superimposed
and was used to sweep the MW-frequency with the adjusted
sawtooth for survey spectra or to stabilize the source by the
output voltage of a phase locking device for precise
frequency measurements.
the corresponding harmonic of the frequency standard.
The output had to pass a gate circuit, which was opened
by the pulse generator when starting a sweep. The time
the gate was open could be adjusted and was chosen to
about 1/3 of a sweep duration. Thus, false triggering of
the computer by the upper frequency marker was avoided
in cases when the lower marker was not caught. The
other receiver was tuned to, e.g. 25 MHz and produced
two frequency markers with 50 MHz separation for
evaluating the frequencies of the absorption lines.
Triggering the computer with a real time frequency
marker had the effect that sampling always started at the
same frequency and that drifts of the un-stabilized
sweeper did not blur the accumulated sweeps. The
achieved frequency accuracy of 300–500 kHz was
sufficient for assignments. Using the frequency markers
these spectra could be concatenated with arbitrary
frequency scaling for viewing larger spectral ranges.
† For precise frequency measurements we made use of the
fact that the width of the band-pass filters in signal- and
reference-channel of our phase locking device (FDS 30
by Schomandlq) was about 4 MHz. Instead of offering a
fixed frequency of 10 MHz for reference (internally
tripled to 30) we drove a voltage-controlled oscillator
with the sawtooth voltage from the low frequency
amplifier, thus generating a reference frequency which
started from 31.5 MHz, decreased linearly to 27.5 MHz,
and jumped back. When locked, the MW-frequency had
to follow; and when the lower side IF was used, this
† For survey spectra a range of 55–60 MHz was chosen.
Some db of the MW-power were coupled to a mixing
diode and mixed with the harmonics of the frequency
standard. The resulting intermediate frequencies were
detected by means of two identical commercial, quartz
controlled communications receivers, 0.09–34 MHz.
One of them was used to trigger the computer. Therefore,
it was set to, e.g. 26 MHz and gave a signal when the
swept MW-frequency was 26 MHz below or above

192
H.-O. Munz et al. / Journal of Molecular Structure 695–696 (2004) 189–202
Fig. 2. Example of a survey spectrum showing the I-quadrupole splitting of the transition J ¼ 101;10 ˆ 91;9: (Lines # 6–9 not assigned).
resulted in a sweep from lower to higher frequencies.
One receiver was set to 31 MHz to trigger the computer
and the other was tuned to, e.g. 29.0 and 28.5 MHz to
produce frequency markers. This arrangement worked
reliably with repetition rates of 9 Hz and sweep durations
of 110 ms. Although sweep reversal was provided, it was
not necessary to use it except for occasional checking:
the parameters could always be adjusted in a way that the
full half-width of a line, be it broad or narrow, extended
over 1/3 of the sweep width and <35 ms were needed to
sweep over this range. Because the time constant of the
lock-in amplifier was set to 1 ms, no time delay could be
observed. The frequency standard was monitored against
the standard signals of DCF77-station, Mainflingen,
and for a name of the spectrum. Then all these data were
transferred to a PC where each project had its own
exchangeable hard disk. A special program calculated the
frequencies of all 1000 points and displayed the spectrum.
The peaks could be marked with a cursor, and these
frequencies were collected in a list. After passing these
extended data to the program ORIGINq by Microcal
Software Inc, the spectrum could be plotted; for frequency
determination a fit to Lorentzian line shape was performed.
Finally, the collected frequencies were sent to the central
computer together with quantum numbers for fitting the
spectrum. Fig. 2 and 3 show examples of a survey spectrum
and of a frequency determination.
Germany, the relative deviations kept ,1028
.
3. Results
As stated, when triggered by the frequency marker of the
first receiver, the computer started displaying the output of
the lock-in amplifier with the repetition rate of 9 Hz and,
when all was correct, the sampling was started. The first
value of the A/D-converter was added to the first memory
location, then the content was read out and checked if it
would not exceed the range of the display. When necessary,
the scale factor was changed and the value was displayed as
the first point of the spectrum on a screen with medium
persistence. This procedure was repeated for all 1000 points
of the sweep and then the computer waited for the next
trigger marker. Sampling could be interrupted as soon as the
S=N-ratio was sufficient, or more sweeps could be added to
the number initially set. Since, the frequency markers were
almost noiseless, it was sufficient to fix the number of
sweeps to 10. The accumulated data of the lock-in output
were never changed and, as original data, stored as integers
with 32 bits. After sampling a menu asked for the setting of
the frequency standard, for the harmonic, for the frequencies
of the markers, the substance, for the experimental details,
3.1. Rotational spectra and spectroscopic constants
The spectrum of the parent molecule was measured from
10 to 40 GHz, the cell temperature kept at 25 8C. Three
hundred and fifteen hyperfine transitions belonging to 89
transitions in Jt; a- and b-type, could be assigned. As
mentioned above, this assignment was not difficult: due to
the large quadrupole splitting of the I-nucleus the typical a-
type pattern of a very near prolate asymmetric top shows up
only from higher J-values upward; but then it is very distinct
as can be seen from Fig. 4 where 30 survey spectra have
been concatenated to show the transition J ¼ 10 ˆ 9: With
these first rough constants, the b-type transitions could be
identified. More details are given in Ref. [9]. Because each
rotational line represents an unresolved multiplet of the up
to 27 hyperfine transitions of the three N-nuclei, the line
widths were about 800 kHz even at very low pressures.
Nevertheless, the assigned transitions could be fitted with a
standard deviation of 62 kHz. The program ‘Q2fit’ by

H.-O. Munz et al. / Journal of Molecular Structure 695–696 (2004) 189–202
193
Fig. 3. Example of a frequency determination (the almost coinciding transitions 63,4 17/2 ˆ 53,3 15/2 and 64,2 9/2 ˆ 54,1 7/2 of I14N15N14N).
Zbigniew Kisie1 was used which performs complete
diagonalization of rotational and quadrupole energies. The
transitions were weighted according to their accuracy; the
calculations were done with Watson’s [10] A-reduction in
Ir-representation. Besides the rotational and the five quartic
distortion constants, xaa; xbb –xcc; and xab could be
determined. The frequencies are given in Table 1, the
spectroscopic constants in Table 3. No transitions of low F-
values could be found which were isolated enough to
measure their Stark-shifts. Therefore, the components of the
dipole moment could not be determined. Because the yield
of IN3 and with it the intensity of the transitions was
fluctuating, even their ratio could not be estimated from the
relative intensities of a- and b-type transitions. Also, due to
the splitting of each transition into six lines the spectrum is
very rich, and we did not succeed in assigning vibrational
satellites.
Because of the small sample of the 15N-substituted
species these measurements had to be restricted to the range
20–30 GHz and not as many sweeps could be taken for the
frequency measurements as would have been desirable.
Therefore, the accuracy of the constants is not comparable
to that of the parent molecule. 59 hyperfine transitions,
belonging to 19 transitions in Jt; could be assigned, but only
four of b-type among them. They are listed in Table 2. The
best results with a standard deviation of 126 kHz were
obtained by fitting only the rotational constants, just DJ and
DJK for the distortion constants, and two of the three
quadrupole quantities: xcc should not be affected by the
rotation of the a–b-plane around the c-axis, and the small
change of xbb by the substitution could be estimated with
sufficient accuracy. Therefore, xbb 2 xcc was held constant,
and the remaining three distortion constants were also fixed
to the values of the parent molecule. The slightly reduced
accuracy of the rotational constants is not detrimental to
the structure calculations. All constants are included in
Table 3. Due to the constraint for the trace of the coupling
tensor the values for xbb and xcc have been slightly changed
by the fitting process.
The substitution of the central N-atom had been
complete: no lines of the parent molecule could be detected.
3.2. Nuclear quadrupole coupling
As mentioned, only the coupling of the 127I-nucleus is
considered, the splitting due to the 14N-nuclei remaining
unresolved. Because xab results directly from the spectrum
along with xaa and xbb; the planar problem can easily be
diagonalized to yield the principal values of the quadrupole
coupling tensor. They are given in Table 4 together with the
angle w between a-inertial axis and the z-principal
quadrupole axis. This angle, a byproduct from the
diagonalization, is an essential quantity for the structure
1
Institute of Physics, Polish Academy of Sciences, Al. Lotnikow 32/46,
Fig. 4. Wide range spectrum of the transition J ¼ 10 ˆ 9 (30 survey spectra
Warsaw, Poland.
concatenated).

194
H.-O. Munz et al. / Journal of Molecular Structure 695–696 (2004) 189–202
Table 1
Rotational transitions of I14N3 (frequencies in MHz)
0
0
00
00
Kc
J0
Ka
Kc
J00
Ka
F0
7/2
F00
5/2
Frequency
o–c
3
1
3
2
1
2
10744.606
10797.318
10805.166
10881.389
10834.824
10948.090
11035.041
11051.404
11114.356
11137.270
11228.576
14395.629
14427.230
14473.617
14499.276
14608.735
14653.825
14660.925
14694.976
14598.691
14607.473
14617.819
14657.356
14758.561
14595.205
14603.994
14614.365
14653.825
14755.630
14510.656
14521.354
14701.463
14715.340
14864.659
14875.237
14900.773
14932.696
18035.426
18058.165
18086.921
18095.954
18357.708
18363.398
18363.398
18317.586
18336.769
18402.416
18310.586
18320.977
18330.209
18355.667
18396.388
18266.643
18284.163
18438.863
18606.417
18616.230
18634.310
18649.475
21660.085
21662.623
20.107
0.174
9/2
7/2
7/2
7/2
20.018
0.007
11/2
3/2
9/2
3
0
3
2
0
2
1/2
20.060
20.137
0.071
7/2
5/2
11/2
9/2
9/2
7/2
20.004
20.052
20.082
20.051
20.043
0.029
3
4
1
1
2
4
2
3
1
1
1
3
7/2
5/2
9/2
7/2
11/2
7/2
9/2
5/2
9/2
7/2
11/2
13/2
7/2
9/2
0.053
11/2
5/2
0.048
4
4
0
2
4
2
3
3
0
2
3
1
20.134
20.110
0.010
9/2
7/2
9/2
9/2
13/2
9/2
11/2
9/2
0.039
20.102
0.022
9/2
7/2
11/2
7/2
9/2
20.002
20.043
0.055
5/2
13/2
9/2
11/2
9/2
4
2
3
3
2
2
0.059
9/2
7/2
0.118
11/2
7/2
9/2
20.030
20.095
0.019
5/2
13/2
11/2
9/2
11/2
9/2
4
4
3
1
1,2
3
3
3
3
1
0,1
2
0.026/0.030
7/2
0.109/0.115
0.056/0.052
0.112
7/2
5/2
11/2
7/2
11/2
5/2
0.032
9/2
7/2
0.015
9/2
9/2
0.066
13/2
9/2
11/2
7/2
20.027
5
1
5
4
1
4
0.030
20.044
0.043
11/2
13/2
15/2
15/2
11/2
13/2
11/2
13/2
15/2
11/2
9/2
9/2
11/2
13/2
13/2
9/2
20.007
0.077
0.218a
5
5
5
0
2
2
5
3
4
4
4
4
0
2
2
4
2
3
11/2
9/2
0.029
0.043
11/2
13/2
9/2
20.011
0.027
20.096
20.120
0.074
7/2
13/2
7/2
11/2
5/2
0.009
15/2
11/2
13/2
15/2
7/2
13/2
9/2
0.001
5
5
3
1
2,3
4
4
4
3
1
1,2
3
0.053/0.067
11/2
13/2
9/2
0.056/0.070
20.062/20.073
20.063
11/2
13/2
15/2
9/2
9/2
20.042
0.007
11/2
13/2
7/2
0.058
6
1
6
5
1
5
20.027
20.078
11/2
9/2
(continued on next page)

H.-O. Munz et al. / Journal of Molecular Structure 695–696 (2004) 189–202
195
Table 1 (continued)
J0
Ka
Kc
J00
Ka
F0
F00
Frequency
o–c
0
0
00
00
Kc
13/2
7/2
11/2
5/2
21679.796
21679.831
21699.046
21702.328
21980.661
21986.197
21994.470
22002.820
22020.565
22022.834
22010.974
22027.484
22027.484
22060.091
22062.169
21999.288
22016.121
22016.121
22051.501
21976.146
21995.590
21997.640
22044.854
22081.052
21963.378
21989.930
22117.140
21888.697
21925.469
22343.503
22346.438
22348.710
22354.123
22361.395
22372.798
25280.638
25283.584
25294.464
25296.668
25310.382
25311.399
25651.709
25658.898
25660.558
25672.644
25668.425
25676.553
25684.997
25703.615
25736.434
25645.380
25667.727
25669.692
25722.630
25761.773
25616.543
25647.436
25661.535
25581.336
26077.898
26080.522
26090.590
0.057
20.040
0.002
15/2
17/2
9/2
13/2
15/2
7/2
20.021
0.021
6
6
0
2
6
5
5
0
2
5
11/2
7/2
9/2
20.042
20,008
0.006
5/2
13/2
17/2
15/2
13/2
15/2
9/2
11/2
15/2
13/2
11/2
13/2
7/2
0.027
20.055
20.072
20.062
0.250a
20.052
20.027
20.021
0.029
4
5
3
4
7/2
5/2
17/2
13/2
15/2
9/2
15/2
11/2
13/2
7/2
6
6
2
3
5
5
2
3
0.134a
17/2
13/2
15/2
11/2
9/2
15/2
11/2
13/2
9/2
0.009
3,4
2,3
2,3
1,2
0.017/0.052
20.003/0.032
0.030/0.065
0.009/0.041
20.052/20.022
20.091
7/2
17/2
15/2
11/2
17/2
13/2
15/2
11/2
9/2
15/2
13/2
9/2
6
4
5
4
20.008
20.019
0.056
15/2
11/2
13/2
9/2
6
6
5
1
1,2
5
5
5
5
1
0,1
4
20.075
20.049
7/2
20.033
0.078
13/2
7/2
11/2
5/2
20.050
20.058
15/2
17/2
11/2
13/2
9/2
13/2
15/2
9/2
20.029
0.018
7
1
7
6
1
6
11/2
7/2
0.022
0.114
15/2
17/2
19/2
11/2
9/2
13/2
15/2
17/2
9/2
0.041
20.014
0.025
7
7
0
3
7
6
6
0
3
6
0.010
7/2
20.020
0.001
13/2
15/2
15/2
13/2
17/2
11/2
19/2
15/2
17/2
13/2
11/2
19/2
15/2
17/2
13/2
15/2
9/2
11/2
13/2
13/2
11/2
15/2
9/2
0.019
4,5
3,4
20.030/0.048
20.091/20.014
20.085/20.009
20.145/20.072
20.142/20.074
20.005
17/2
13/2
15/2
11/2
9/2
7
7
4
5
3,4
2,3
6
6
4
5
2,3
1,2
0.004
20.015
20.103/20.096
20.015
20.002
17/2
13/2
15/2
11/2
13/2
7/2
20.011
0.060
7
7
6
1
1,2
6
6
6
6
1
0,1
5
20.029
20.084
0.028
15/2
17/2
13/2
15/2
20.021
(continued on next page)

196
H.-O. Munz et al. / Journal of Molecular Structure 695–696 (2004) 189–202
Table 1 (continued)
J0
Ka
Kc
J00
Ka
F0
F00
Frequency
o–c
0
0
00
00
Kc
13/2
19/2
11/2
13/2
15/2
11/2
17/2
21/2
19/2
13/2
19/2
15/2
17/2
21/2
13/2
21/2
15/2
17/2
19/2
11/2
17/2
13/2
15/2
19/2
19/2
13/2
21/2
15/2
17/2
13/2
19/2
23/2
21/2
17/2
23/2
19/2
15/2
21/2
13/2
13/2
19/2
21/2
17/2
23/2
17/2
19/2
15/2
21/2
25/2
23/2
17/2
15/2
23/2
19/2
21/2
25/2
19/2
17/2
15/2
21/2
23/2
25/2
19/2
11/2
17/2
9/2
26103.118
26120.640
26120.640
28897.876
28900.668
28908.052
28911.078
28921.391
28921.391
29288.265
29312.285
29322.085
29330.433
29342.620
29693.680
29721.076
29723.313
29790.561
29794.161
29801.557
29381.258
29381.258
39354.829
29365.656
29356.190
29383.090
29442.683
32512.972
32515.539
32520.676
32523.844
32531.473
32531.874
32998.613
33000.385
33009.256
33056.793
33090.699
33091.270
33524.780
33527.377
33534.035
33536.562
33557.225
36126.223
36128.505
36132.439
36135.388
36141.247
36141.783
36628.862
36630.110
36630.676
36642.934
36647.591
36655.497
37246.767
37246.767
37246.767
37247.879
37252.824
37258.240
39737.815
20.024
0.014
0.162a
8
1
8
7
1
7
11/2
13/2
9/2
20.058
20.087
0.040
15/2
19/2
17/2
11/2
17/2
13/2
15/2
19/2
11/2
19/2
13/2
15/2
17/2
9/2
0.038
0.080
0.064
8
8
0
1
8
7
7
7
0
1
7
6
20.019
20.051
20.080
20.054
0.053
20.009
20.031
20.038
20.016
0.028
20.137
0.070
8
8
8
2
3
4
6
7
7
7
2
3
4
5
15/2
11/2
13/2
17/2
17/2
11/2
19/2
13/2
15/2
11/2
17/2
21/2
19/2
15/2
21/2
17/2
13/2
19/2
11/2
11/2
17/2
19/2
15/2
21/2
15/2
17/2
13/2
19/2
23/2
21/2
15/2
13/2
21/2
17/2
19/2
23/2
17/2
15/2
13/2
19/2
21/2
23/2
17/2
20.083
20.089/0.063
20.117/0.032
20.021
20.109
0.009
5,6
4,5
4,5
3,4
8
9
5
1
3,4
9
7
8
5
1
2,3
8
20.013
20.001
20.042
0.009
0.080
0.048
9
0
9
8
0
8
20.057
20.002
20.036
20.086
20.012
20.054
0.019
9
1
1
8
8
9
1
1
7
9
0.033
20.001
20.067
20.013
20.013
0.014
10
10
0.026
0.048
0.051
0.001
10
10
11
0
1
1
10
9
0
1
1
9
0.030
20.030
20.025
20.013
0.046
0.046
0.175a
20.003a
20.109a
20.038
0.037
9
9
8
20.015
20.030
11
10
10
(continued on next page)

H.-O. Munz et al. / Journal of Molecular Structure 695–696 (2004) 189–202
197
Table 1 (continued)
J0
Ka
Kc
J00
Ka
F0
F00
Frequency
o–c
0
0
00
00
Kc
21/2
17/2
23/2
27/2
25/2
19/2
3/2
19/2
15/2
21/2
25/2
23/2
17/2
1/2
39739.805
39743.099
39745.634
39750.498
39751.074
40293.380
28785.962
29016.202
29374.651
29551.941
29570.640
29591.536
29686.730
29735.045
29818.028
29841.541
30046.310
30087.360
30130.604
30180.239
30190.550
30465.511
30525.525
30591.126
30598.557
30599.370
30618.088
30904.142
30989.737
31019.217
31058.502
31423.104
31460.936
31484.422
31516.287
31554.762
31576.586
32039.836
32063.766
32102.396
32138.230
32158.315
32176.868
32710.761
32731.796
32773.950
32808.932
32842.276
33456.826
33547.678
33582.086
33593.038
34265.983
34280.931
34334.801
34359.349
34395.550
35153.897
35168.391
35223.752
35247.210
35271.881
0.006
0.093
0.011
0.045
0.010
11
2
0
1
1
11
1
10
2
0
0
0
10
2
20.007
20.116
20.052
20.003
20.035
0.029
3
2
3
3
3/2
1/2
9/2
9/2
4
5
1
1
3
4
4
5
0
0
4
5
7/2
7/2
11/2
9/2
11/2
9/2
20.065
0.010
5/2
5/2
15/2
9/2
15/2
9/2
0.023
20.075
20.004
20.129
0.050
13/2
7/2
13/2
7/2
6
7
1
1
5
6
6
7
0
0
6
7
17/2
9/2
17/2
9/2
0.066
15/2
13/2
9/2
15/2
13/2
9/2
0.129
0.038
0.008
19/2
17/2
15/2
11/2
13/2
21/2
11/2
15/2
17/2
13/2
23/2
15/2
21/2
17/2
19/2
15/2
25/2
17/2
23/2
19/2
21/2
17/2
27/2
19/2
25/2
23/2
19/2
27/2
23/2
25/2
21/2
31/2
23/2
29/2
27/2
23/2
33/2
25/2
31/2
27/2
19/2
17/2
15/2
11/2
13/2
21/2
11/2
15/2
17/2
13/2
23/2
15/2
21/2
17/2
19/2
15/2
25/2
17/2
23/2
19/2
21/2
17/2
27/2
19/2
25/2
23/2
19/2
27/2
23/2
25/2
21/2
31/2
23/2
29/2
27/2
23/2
33/2
25/2
31/2
27/2
0.024
0.083
0.176
0.083
0.087
8
9
1
1
7
8
8
9
0
0
8
9
0.100
0.082
0.031
0.033
0.012
0.044
20.051
0.123
0.100
0.063
10
11
1
1
9
10
11
0
0
10
11
0.007
0.069
20.015
20.019
0.009
20.026
20.002
0.027
10
0.099
0.049
0.087
12
13
1
1
11
12
12
13
0
0
12
13
0.052
0.033
0.086
0.014
0.072
20.034
20.032
0.051
0.012
14
1
13
14
0
14
0.008
20.065
0.043
20.007
0.024
(continued on next page)

198
H.-O. Munz et al. / Journal of Molecular Structure 695–696 (2004) 189–202
Table 1 (continued)
J0
Ka
Kc
J00
Ka
F0
F00
Frequency
o–c
0
0
00
00
Kc
29/2
25/2
27/2
33/2
29/2
31/2
27/2
37/2
29/2
39/2
31/2
37/2
33/2
35/2
31/2
41/2
33/2
39/2
35/2
37/2
25/2
35/2
33/2
31/2
29/2
31/2
25/2
19/2
25/2
31/2
21/2
23/2
33/2
31/2
27/2
29/2
27/2
33/2
29/2
31/2
37/2
33/2
35/2
33/2
35/2
37/2
61/2
63/2
53/2
55/2
61/2
57/2
59/2
63/2
55/2
57/2
67/2
57/2
65/2
63/2
59/2
61/2
29/2
25/2
27/2
33/2
29/2
31/2
27/2
37/2
29/2
39/2
31/2
37/2
33/2
35/2
31/2
41/2
33/2
39/2
35/2
27/2
23/2
33/2
31/2
29/2
27/2
29/2
27/2
21/2
27/2
33/2
23/2
25/2
35/2
33/2
29/2
31/2
29/2
35/2
31/2
33/2
39/2
35/2
37/2
35/2
37/2
39/2
63/2
65/2
55/2
57/2
63/2
59/2
61/2
65/2
57/2
59/2
69/2
59/2
67/2
65/2
61/2
63/2
35282.922
36121.197
36192.370
36214.521
36240.191
36250.450
37170.792
37184.640
38305.797
38319.711
38379.910
38399.813
38428.017
38437.054
39529.332
39543.708
39605.310
39623.903
39654.122
39662.416
31283.887
31292.693
31348.907
31360.913
35654.230
35680.099
38451.533
38471.176
29644.438
29706.264
29714.045
39039.674
39052.717
39145.919
39172.959
39182.039
36551.088
36571.031
36599.185
31699.768
31717.613
31747.826
31755.055
29433.019
29480.402
29486.681
31566.466
38104.262
38118.468
38121.718
38136.847
38145.284
38153.726
34939.296
34964.426
35057.788
31865.860
31888.219
31898.883
31903.813
31906.237
31911.152
0.027
0.001
15
1
14
15
0
15
0.076
20.018
0.073
0.012
16
17
1
1
15
16
16
17
0
0
16
17
20.053
20.038
0.035
20.020
0.056
20.012
0.057
20.001
0.106
18
15
1
0
17
15
18
14
0
1
18
14
0.017
0.053
0.002
0.083
0.003
0.044
0.122
20.101
0.117
16
11
14
0
2
1
16
10
13
15
12
13
1
1
2
15
11
12
0.085
20.151
20.003
0.098
20.015
0.017
20.014
20.054
20.052
20.122
20.027
20.086
0.015
14
2
12
15
1
15
15
17
2
2
13
15
16
18
1
1
16
18
0.044
20.131
20.075
20.098
20.124
20.124
0.066
18
2
16
19
1
19
20.135
20.059
0.010
28
29
3
3
26
26
29
30
2
2
27
29
0.081
20.063
20.049
20.036
0.099
0.079
30
31
3
3
27
28
31
32
2
2
30
31
0.037
0.040
0.137
0.090
20.065
20.032
20.112
20.032
20.112
Three hundred and fifteen transitions in F (89 in Jt) with Fmax ¼ 67=2 used in the fit; rms ¼ 0.062 MHz.
a
Partially masked; not used in the fit.

H.-O. Munz et al. / Journal of Molecular Structure 695–696 (2004) 189–202
199
Table 2
Rotational transitions of I14N15N14N (frequencies in MHz)
0
0
00
00
Kc
J0
Ka
Kc
J00
Ka
F0
9/2
F00
7/2
Frequency
o–c
6
2
5
5
2
4
21678.095
21678.095
21711.824
21713.535
21845.916
21683.221
21685.735
21688.488
21689.038
21723.663
21706.644
21742.972
21742.972
21842.430
21767.341
21779.411
21788.928
21860.884
24898.129
24901.031
24914.184
24927.983
24928.999
25673.242
25682.884
25685.105
25695.831
25705.582
25221.691
25252.559
25354.012
25400.136
25250.665
25302.764
25328.011
25367.349
25395.550
25258.821
25289.205
25266.554
25267.158
25273.760
25282.001
25290.464
25309.294
25342.014
25350.440
25284.608
25313.385
25318.724
25427.764
25298.378
25302.059
25315.070
25329.391
25334.704
25363.052
25395.550
25440.091
25624.294
0.170
0.084
15/2
7/2
13/2
5/2
0.224
17/2
7/2
15/2
7/2
20.008
0.074
6
6
0
2
6
4
5
5
0
2
5
3
17/2
15/2
9/2
15/2
13/2
7/2
0.088
20.029
20.015
0.245
15/2
17/2
9/2
13/2
15/2
7/2
0.062
6
6
3
4
4
2
5
5
3
4
3
1
0.061
17/2
9/2
15/2
7/2
20.009
0.119
7/2
5/2
20.272
20.198
20.548
0.066
6
3
1
2
0
1
3
5
5
5
3
1
1
1
1
2
4
7/2
5/2
6
15/2
31/2
31/2
11/2
13/2
15/2
17/2
19/2
15/2
17/2
13/2
11/2
19/2
15/2
17/2
11/2
19/2
15/2
11/2
11/2
19/2
9/2
15/2
33/2
29/2
9/2
15
13
7
14
13
7
16
12
6
15
12
6
0.199
20.025
20.069
20.001
0.011
11/2
13/2
15/2
17/2
13/2
15/2
11/2
9/2
0.046
7
1
6
6
1
5
0.215
20.281
20.112
20.202
0.065
17/2
13/2
15/2
9/2
7
7
5
4
2
3
6
6
5
4
1
2
20.076
20.041
0.057
17/2
13/2
11/2
9/2
20.178
0.048
20.344
20.161
20.016
20.130
20.241
0.049
17/2
7/2
7
7
0
3
7
5
6
6
0
3
6
4
11/2
19/2
9/2
9/2
17/2
7/2
20.046
20.105
0.069
13/2
15/2
13/2
17/2
11/2
19/2
9/2
11/2
13/2
11/2
15/2
9/2
0.094
0.118
0.198
17/2
7/2
20.018
0.059
7
7
2
2
6
5
6
6
2
2
5
4
15/2
9/2
13/2
7/2
0.050
0.262
19/2
11/2
13/2
15/2
17/2
9/2
17/2
11/2
11/2
13/2
15/2
7/2
0.131
20.097
20.099
0.079
0.159
0.093
19/2
13/2
9/2
17/2
13/2
7/2
20.159
20.427
20.130
20.148
20.003
7
4
6
2
3
1
6
6
4
6
1
2
0
7
19/2
37/2
17/2
37/2
20
18
21
21
Fifty nine Transitions in F (19 in Jt) with Fmax ¼ 37=2 used in the fit; rms ¼ 126 kHz.

200
H.-O. Munz et al. / Journal of Molecular Structure 695–696 (2004) 189–202
Table 3
Table 5
Spectroscopic constants of iodine azide isotopomers
Moments of inertia of iodine azide from Watson’s determinable constants
I14N3
I14N15N14
N
I14N3
I14N15N14
N
Rotational constants (MHz)
Ia
Ib
Ic
D
16.4214
267.2556
283.9788
0.3019
16.4285
271.5425
288.2756
0.3046
A
B
C
30775.669 (10)
1891.0338 (5)
1779.6352 (4)
30762.32 (24)
1861.1795 (19)
1753.1099 (45)
Centrifugal distortion constants (kHz)
Ref. [10].
DJ
DJK
DK
dj
0.6017 (15)
0.375 (35)
244.246 (94)
1800.0a
245.028 (35)
1800.0 (25)
0.07993 (20)
2.76 (10)
0.07993a
needed: three bond lengths and two bond angles. But the
measurements yielded only four independent moments of
inertia. However, because the quadrupole coupling tensor
is so symmetric we could reasonably assume that its
principal axis ‘z’ coincides with the direction of the I–N–
bond, and therefore the angle w is another independent
quantity for the structure determination. From the six
moments of inertia Ib and Ic from each species were used
because these are less sensitive to vibrational effects than
the small values of Ia: Two kinds of calculation have been
carried out:
dk
2.76a
k
20.9923
20.9925
Quadrupole coupling constants (MHz)
xaa
21912.39 (17)
2838.16 (24)
21656.93 (20)
537.11 (27)
21917.31 (39)
2832.28a
xbb 2 xcc
xab
xbb
21646.4 (20)
542.6 (13)
1374.8
xcc
1375.28 (15)
a
Values fixed to parent molecule.
calculations. The principal values should be the same within
reasonable error limits for both isotopomers. This is not
quite the case due to the lack of sufficient b-type transitions
of the 15N-species. However, the relative differences
are rather small, 0.4–0.04%, confirming the correctness
of the assignments. Of course, for comparisons and
discussions the precise values of the parent molecule are
used. The asymmetry parameter is remarkably close to zero,
indicating a very symmetric charge distribution around
the I–N-bond.
† The five equations expressing the four moments of
inertia and the angle w of the 14N-species in terms of the
parameters have been solved for these parameters.
† In addition the angle w of the 15N-species has been
used to give a system of six equations. Thus, a least
squares calculation for the five parameters was
possible.
Both methods gave exactly the same results for the
structural parameters. They are listed in Table 6 under
‘Method I’. But, as pointed out in Section 4, the resulting
structure is rather unusual for an azido group, and the
assumption of the coincidence of xzz with the I–N-direction
is obviously not quite correct. Therefore, another calcu-
lation was carried out: instead of making use of the angle w;
the distance Nb–Nc (numbering shown on Fig. 5) was fixed
to the value as found in the well determined hydrazoic acid
[13] and the four moments of inertia solved for the
remaining four parameters. The results are also shown in
Table 6 as ‘Method II’, the angle w now different, and
depicted in Fig. 5.
3.3. Structure
From Watson’s [10] determinable constants the
moments of inertia were calculated as given in Table 5.
2
The inertial defect of 0.3019 amu A proves the planar
˚
arrangement of the four atoms and is typical for such a
type of molecules with Cs-symmetry and consisting of
only a few ‘heavy’ atoms: INCO 0.4441 [8], BrNCO
0.401 [11], NC–N3 0.347 [12], FN3 0.229 [14], and ClN3
[15]. To describe this planar geometry five parameters are
Table 4
Principal values of the quadrupole coupling tensor of iodine azide
isotopomers
4. Discussion
I
14N3
I14N15N14
N
For a structure comparison only three other covalent
azides are available with reliable microwave structures
without assumptions: HN3 [13], FN3 [14], and ClN3 [15].
These are included in Table 6 along with the results of the
electron diffraction work [6] and of the ab initio
calculations [2]. As already pointed out in Refs. [14,16],
xxx
xyy
xzz
wa
h
1372.80 (9)
1375.28 (15)
1367.6 (5)
1374.8 (5)
22748.08 (35)
26.765 (2)
22742.4 (20)
26.619 (8)
0.0009 (2)
0.0026 (11)
a
\ between quadrupole z-axis and inertia a-axis.

H.-O. Munz et al. / Journal of Molecular Structure 695–696 (2004) 189–202
201
Table 6
Structure comparison of azides (distances in pm, angles in degrees)
a
IN3 Method II
b
IN3 El.-diffr.
d
e
f
Parameter
IN3a Method I
IN3cab initio
FN3
HN3
ClN3
r(X–Na)
215.8
117.9
111.5
109.1
172.2
26.765
209.7
123.8
113.4g
109.9
172.3
28.009
212.0
126.0
114.7
106.6
169.6
213.3
125.8
116.7
110.2
171.4
144.4
125.3
113.2
103.8
170.9
101.5
124.3
113.4
108.8
171.3
174.5
125.2
113.3
108.8
171.9
r(–Na–Nb)
r(Nb–Nc)
\(X–Na–Nb)
\(Na–Nb–Nc)
wh
a
This work.
b
Ref. [6].
c
Ref. [2].
d
Ref. [14].
e
Ref. [13].
f
Ref. [15].
g
Assumed.
h
Angle between quadrupole z-axis and inertia a-axis.
the geometry of the azido group is rather insensitive to the
adhering atom or group. It is obvious that the azido group
in iodine azide calculated with ‘Method I’ differs
considerably from that in the other compounds: both N–
N-bonds are notably shorter and therefore, the whole
group by 9 pm. On the other hand, the I–N-bond is longer
than reasonably expected. If we instead make use of the
mentioned stability of the azido group and consider the
short distance Nb–Nc as the most stable one, hold it fixed,
and solve for the remaining parameters as done in
‘Method II’, we obtain a very reasonable structure. This
geometry is remarkably close to that of ClN3. It should be
mentioned that the two angles appear almost invariant
against all kinds of assumptions. Especially, the deviation
from linearity of the N3-group in favour of a trans-
configuration has again been clearly established: no real
solution for the parameters could be found if the azido-
group was forced to a linear arrangement. No error limits
for the parameters have been given in this rigid rotor
calculation because these are determined by the unknown
vibrational effects. However, the distances should be
accurate within ^1.3 pm and the angles within ^1.58
estimated from the uncertainty of the assumption and a
contribution from the non-rigidity. The difference in w of
1.248 between diagonalization of the coupling tensor and
from the structure calculation according to ‘Method II’
seems negligible at first glance; but because the influence
on the structure is so serious, it could be real and a hint
of a divergence of the directions of xzz and the I–N-
connecting line.
It is also most interesting to compare the bond
characteristics of the two isoelectronic molecules iodine
azide and iodine isocyanate as they become manifest in the
principal values of the quadrupole coupling tensor and their
derived quantities. This is done in Table 7: ionic character
and p-character were calculated according to the theory of
Townes and Dailey [17] as worked up in Ref. [18], shielding
included. The close analogy is evident; but the numbers
show that iodine azide has even less double bond and ionic
character than the isocyanate.
Table 7
Comparison of the I–N-bond characteristics in iodine azide and iodine
isocyanate
INCOb
a
IN3
xxx/MHz
xyy/MHz
xzz/MHz
h
1372.80
1375.28
1499.5
1486.58
22986.1c
20.004
23
22748.08
0.0009
15
0.1
id (%)
pe (%)
0.4
a
This work.
b
c
d
e
Ref. [8].
Sign error in Ref. [8].
Ionic character.
p-Bond character.
Fig. 5. Geometry of iodine azide: parameters in Table 6 under ‘Method II’.

202
H.-O. Munz et al. / Journal of Molecular Structure 695–696 (2004) 189–202
¨
[6] M. Hargittai, J. Molnar, T.M. Klapotke, I.C. Tornieporth-Oetting, M.
Acknowledgements
Kolonits, I. Hargittai, J. Phys. Chem. 98 (1994) 10095.
¨
[7] H.-O. Munz, H.-K. Bodenseh, T.M. Klapotke, 14th Colloquium on
¨
We are very indebted to Prof. Dr T.M. Klapotke, Institut
fu¨r Anorganische Chemie der LMU Mu¨nchen, for
suggesting this investigation, for providing materials, and
for helpful discussions. We also thank Dr dr.hab. Z. Kisiel
for his fitting program. The supports of the Deutsche
Forschungsgemeinschaft and the Fonds der Chemischen
Industrie are gratefully acknowledged.
High Resolution Molecular Spectroscopy, September 11–15, Dijon,
France, 1995, #D16.
[8] H.M. Jemson, W. Lewis-Bevan, N.P.C. Westwood, M.C.L. Gerry,
J. Mol. Spectrosc. 119 (1986) 22.
[9] H.-O. Munz, Dissertation Ulm, 1999.
[10] J.K.G. Watson, in: J.R. Durig (Ed.), Vibrational Spectra and
Structure: a Series of Advances, vol. 6, Elsevier, Amsterdam, 1977,
pp. 1–89.
[11] H.M. Jemson, W. Lewis-Bevan, N.P.C. Westwood, M.C.L. Gerry,
J. Mol. Spectrosc. 118 (1986) 481.
[12] C.C. Costain, H.W. Kroto, Can. J. Phys. 50 (1972) 1453.
[13] B.P. Winnewisser, J. Mol. Spectrosc. 82 (1980) 220.
[14] D. Christen, H.G. Mack, G. Schatte, H. Willner, J. Am. Chem. Soc.
110 (1988) 70.
References
[1] A. Hantzsch, Ber. Dtsch. Chem. Ges. 33 (1900) 522.
¨
[2] P. Butzek, T.M. Klapotke, P.v.R. Schleyer, I.C. Tornieporth-
[15] R.L. Cook, M.C.L. Gerry, J. Chem. Phys. 53 (1970) 2525.
[16] N. Heineking, M.C.L. Gerry, Z. Naturforsch. 44a (1989) 669.
[17] C.H. Townes, B.P. Dailey, J. Chem. Phys. 17 (1949) 782. B.P. Dailey,
C.H. Townes, J. Chem. Phys. 23 (1955) 118.
Oettingand, P.S. White, Angew. Chem. Int. Ed. Engl. 32 (1993) 275.
¨
[3] A. Schulz, I.C. Tornieporth-Oetting, T.M. Klapotke, Inorg. Chem. 34
(1995) 4343.
¨
[4] P. Geissler, T.M. Klapotke, H.-J. Kroth, Spectrochim. Acta, Part A 51
[18] W. Gordy, R.L. Cook, Microwave molecular spectra, in: A.
Weissberger (Ed.), Technique of Organic Chemistry, second ed.,
vol. IX, Interscience Publishers, New York, 1970, pp. 551–616.
(1995) 1075.
¨
[5] M.-J. Crawford, T.M. Klapotke, Internet J. Vibr. Spectrosc., http://
www.ijvs.com/volume3/edition6/section2.html