Photolysis of compressed sodium azide (NaN3) as a synthetic pathway to nitrogen materials

Article abstract of DOI:10.1021/jp025963u

The potential syntheses of pure nitrogen materials may involve photolysis of azide compounds. In this work, compressed solid samples of NaN₃ were photolyzed in gem anvil cells at pressures up to 5.0 GPa. Reaction rates and mechanisms for the photolytic processes were studied using single-shot time-resolved (microsecond scale) absorption spectroscopy. Two broad absorption bands due to reaction intermediates were observed at 300 and 700 nm. Tentative assignments of the reaction intermediates and their reaction rates were used to help elucidate the potential for high-pressure syntheses of novel nitrogen structures. Finally, recovered products were analyzed with infrared and Raman vibrational spectroscopy.

Full text of DOI:10.1021/jp025963u

944
J. Phys. Chem. A 2003, 107, 944-947
Photolysis of Compressed Sodium Azide (NaN3) as a Synthetic Pathway to Nitrogen
Materials
,†
†
Suhithi M. Peiris* and Thomas P. Russell
Chemistry DiVision, NaVal Research Lab, Washington, D.C. 20375
ReceiVed: April 17, 2002; In Final Form: September 27, 2002
The potential syntheses of pure nitrogen materials may involve photolysis of azide compounds. In this work,
compressed solid samples of NaN were photolyzed in gem anvil cells at pressures up to 5.0 GPa. Reaction
3
rates and mechanisms for the photolytic processes were studied using single-shot time-resolved (microsecond
scale) absorption spectroscopy. Two broad absorption bands due to reaction intermediates were observed at
300 and 700 nm. Tentative assignments of the reaction intermediates and their reaction rates were used to
help elucidate the potential for high-pressure syntheses of novel nitrogen structures. Finally, recovered products
were analyzed with infrared and Raman vibrational spectroscopy.
3
Introduction
reported. In this reference, a transient visible absorption band
•
-
observed at 700 nm, was assigned to N6
More recently, the pure nitrogen ion (N5 ) was produced by
Christe and Wilson. They report that the N5 cation needs to
be accompanied by large anions such as AsF5 or SbF5 . The
N5 geometry is stabilized by a resonance between two
structures, each with four N-N bonds with bond orders of 1,
.5, 2.5, and 3. The synthesis of N5 also includes the reaction
.
+
High-energy density materials that burn with release of
environmentally clean byproducts are of interest as more
efficient fuels/propellants that generate larger thrusts for super-
sonic and space applications. They also have potential applica-
tions as explosives and other energy storage systems. The most
promising material would be an all-nitrogen compound in a
structure with high lattice energy, which would also be a
relatively inert solid at room temperature. Conceivably such a
solid would decompose, producing only N2, a relatively in-
nocuous product (air being composed of 70% N2).
4
+
-
-
+
+
1
+
- 4
of N2 F with the azide anion N3 .
The efforts reported here focus on the photolytic synthesis
of N materials from solid azides at high pressure. Photolysis of
azide ions occurring by one-photon absorption at 230 nm or
two- and three-photon absorption (at blue and green wave-
Past efforts to synthesize all-N materials have established the
azide radical (N3 ) as a useful starting material.
mention of an all-N component of a synthesized material in
•
1-4
The earliest
8
-10
lengths) are known to yield azide radicals.
Under highly
confined conditions such as elevated pressure azide radicals may
bond with other azide species or nitrogen species to form larger
high-N molecules. This paper presents our results on NaN3 and
proposes a photolysis reaction mechanism for this relatively inert
and widely used material.
5
published literature, is in 1956 by Huisgen and Ugi. On the
1
5
basis of N studies they postulated that a thermally unstable
product resulting from a reaction of benzenediazonium chloride
with lithium azide in methanol at -40 °C, was phenylpentazole.
Subsequent work showed that electron donating substituents in
the para position of the benzene ring stabilized the pentazole
Procedure
6
ring system. A series of para-substituted derivatives were then
isolated at low temperatures. The most stable of these substituted
pentazoles, 4-(dimethylamino)phenylpentazole, decomposes only
at 50 °C. A crystal structure for this all-N ring was reported
Samples were loaded in Merrill-Bassett type pressure cells
containing either diamond or sapphire anvils with 350-800 µm
1
1
wide culet flats. One hundred micrometer thick 301 stainless
steel gaskets drilled with 250-550 µm diameter holes were
placed between the anvils. A ruby sphere about 10 µm in
diameter was placed with the sample to measure the initial
7
much later. The measured NsN bond distances in the pentazole
ring are in the range 1.30-1.35 Å. These distances are
intermediate between single NsN bonds (1.45 Å) and double
NdN bonds (1.25 Å), and therefore, the pentazole ring, with
its 6π electrons and nearly equivalent bond lengths, is considered
aromatic.
1
2
pressure applied.
In our initial experiments we found that NaN3 reacted with
air to form an undesired Na2CO3 passivation layer. When
Later, during spectroscopic studies of the photoionization of
azides in solution, pure nitrogen molecules were postulated.
working with micro quantities, upon reaction of NaN , the
passivation layer was sufficiently significant to be observed by
3
8
Optical absorption studies showed the formation of transient
species in solution. A prominent absorption detected at 278 nm
was identified as the N3• radical in its 2πg ground state.
IR spectra. Consequently in later experiments, NaN powders
were stored, handled, and loaded into pressure cells in a dry
3
N -purged glovebox. To ensure good light transmittance and
2
•-
Subsequently, a spectroscopic and theoretical study of N6 was
optically thin NaN samples, the azide was loaded between two
3
layers of NaCl powder.
*
Corresponding author. E-mail address: PeirisSM@ih.navy.mil, Work
The experimental arrangement to measure time-resolved
absorption is shown in Figure 1. Details of this experimental
Phone: (301) 744-4252.
†
Now at: Research and Technology Department, Naval Surface Warfare
1
3,14
method can be found elsewhere.
A single pulse from a
Center - Indian Head Division, 101, Strauss Avenue, Indian Head, Maryland
2
0640.
flashlamp-pumped dye laser (514 nm, 7-8 µs pulse duration,
1
0.1021/jp025963u This article not subject to U.S. Copyright. Published 2003 by the American Chemical Society
Published on Web 01/17/2003

Photolysis of Compressed NaN3
J. Phys. Chem. A, Vol. 107, No. 6, 2003 945
the sample to air. Once removed from the pressure cell, the
samples were again analyzed (IR, Raman, UV-vis, and XPS)
prior to and after exposure to air at ambient pressure.
Results
The “prephotolysis” absorption spectra of unreacted NaN3
show an absorbance rising to 250 nm (This is the lower detection
limit of our UV spectroscopy system). This absorption edge is
-
assigned to the absorption of the N3 in the NaN3 (230 nm).
Once the NaN3 photolysis is started with the laser pulse at time
0, changes in UV absorption (relative to prephotolysis absorption
of NaN3), show an increase in absorption intensity centered near
3
00 nm.
The right plot in Figure 2 represents a spectral sequence of
Figure 1. Experimental setup of the single-shot time-resolved spec-
troscopy system used to study reaction chemistry at high pressure. Both
the laser pulse used to initiate photolysis and the 25 µs UV-vis source
pulse used as incident absorbance were focused into the sample within
the diamond cell.
this increase in absorption, showing time-resolved “change in
absorbance” data dispersed by the UV spectrometer and
measured at time intervals of 0.857 µs, obtained from a single-
shot reaction at pressure. Notice the change in absorption
gradually rising from about 360 nm to peak at about 300 nm.
Previous photolysis studies of azides in solution showed that
2
3
-50 J/cm fluence) is used to photolyze the samples. Simul-
taneously, a source lamp for incident absorption provides light
covering the wavelength region 250-900 nm. The pulse
duration of this absorption source lamp is 25 µs. A dichroic
beam splitter separates the light transmitted through the sample
and anvils between two wavelength regions of 250-420 nm
and 550-850 nm. The transmitted light in each wavelength
region is further dispersed in wavelength by a spectrometer and
is streaked in time normal to the wavelength axis by a
Hamamatsu streak camera. CCD detectors then measure the
intensities of the wavelength- and time-resolved light.
•
9,10
the N3 radical absorbs light at 277 nm.
The growth of the
absorbance feature at 300 nm is postulated to be the formation
•
of the N3 radical, since the azide in these experiments is in the
solid state and at some pressure, when compared to the azide
solutions used in the previous studies. Similarly, the second
spectrometer dispersing visible wavelengths shows an absorption
feature around 750 nm. Previous work by Workentine et al.
•
-
reports the formation of N6 , which has a broad featureless
3
absorption centered near 700 nm. Hence, this second absorption
•
-
feature in the visible is postulated to be N6
.
First, “prephotolysis” UV-vis transmission was recorded
through the pressurized NaN3 sample before initiating reaction
with the laser. This transmission was considered a reference
intensity Io(t,λ). The transmitted light during photolysis was
similarly recorded to yield I(t,λ). A “change in absorbance” (∆A)
was calculated with respect to the reference intensity.
Figure 2 also shows how reaction profiles or kinetics are
estimated from the change in absorbance data. Change in
absorbance within a certain wavelength region around a
particular wavelength is averaged and plotted against time. For
these samples, 6 nm wavelength regions around the wavelengths
of 300 and 700 nm were chosen. In Figure 2, the averaged region
around 300 nm is marked by a rectangle in the plot on the left.
On the right side of Figure 2 is a plot of the reaction profile
calculated by averaging the change in absorbance within the
area marked by a rectangle and plotted against time. This plot
is considered a reaction profile.
∆
A(t,λ) ) log(I /I)
o
Photodiodes monitored the light outputs of the initiating laser
and the UV-vis source. These together with a digital signal
analyzer were used to determine the start of the laser pulse
assigned to be zero time (t ) 0). Finally, a “postphotolysis”
UV-vis transmission spectrum of the pressurized reaction
remnants in the cell was recorded. The residue in the cell after
photolysis was analyzed in situ at pressure with UV-vis, IR,
and Raman spectroscopy. Subsequently, the cells were opened
in a dry-nitrogen glovebag to unload the cell without exposing
Figure 3 shows reaction profiles of averaged change in
absorbance versus time. This sample was at an initial pressure
of 1.1 GPa (11 kbar). The UV absorbance (at 300 nm) shown
as circles, increases to 5.2 µs, and then decreases, indicating
•
the appearance and disappearance of N3 radicals. The visible
absorbance (at 700 nm) shown as stars, achieves a maximum
Figure 2. On the left is a plot of typical wavelength- and time-resolved changes in absorbance measured around 300 nm during the photolysis of
NaN at 1.1 GPa. The reaction profile is estimated by averaging the change in absorbance at each time over the wavelength region denoted by the
3
rectangle. On the right is the reaction profile or change in absorbance averaged around 300 nm versus time. The dotted lines are included to show
the averaged absorbance values on the left plot that yielded the data on the right plot.

946 J. Phys. Chem. A, Vol. 107, No. 6, 2003
Peiris and Russell
Figure 5. Micro-Raman spectrum of the red liquid residue compressed
within the diamond cell. The Raman lines originating from the red
liquid disappear when the cell pressure is reduced to ambient.
TABLE 1: Observed Raman Frequencies of the Red Liquid
Residue and Calculated Frequencies of Other Possible
Species
Figure 3. Change in absorbance averaged at 300 nm from the UV
spectrometer, and the change in absorbance averaged at 700 nm from
the vis spectrometer versus time, to show typical reaction profiles at
species
symmetry
raman modes (cm-1)
ref.
1
.1 GPa. The solid lines are not the fitted curve. They are added to
red liquid
?
D
C
C
C
781, 1140, 1378, 1894, 2185
122, 1267, 1358
939, 1252, 1299, 2151
547, 721, 1231, 2191
835, 1145, 1954
our data
16
17
15
15
show the reaction profile. The laser pulse is included to show its time
profile.
NaN
3
3d
-
N
7
2V
ClN
Cl
3
s
(linear)
(boat)
2
N
6
1
the laser fluence pulse rises and falls off when the intensity of
the laser pulse drops, indicating that the production of N3•
radicals is primarily photolytic. However, as demonstrated in
•
-
Figure 4, the N6 temporal profile appears to depend on
pressure. In this pressure range, the time to maximum absor-
•-
bance of N6 decreases from 8 to 6.0 µs, suggesting that higher
initial pressures result in higher rates of formation of N6•
-
.
The products remnant in the anvil cell upon completion of
photolysis, were studied with optical light microscopy, IR, and
Raman spectroscopy. In all samples, optical microscopy indi-
cated a silver-colored solid. When exposed to air the silver solid
reacted to form a liquid (as expected for 2Na + 2H2O f 2NaOH
+
H2), indicating that it is most likely Na metal. In all samples,
Raman spectroscopy indicated N2 formation. Optical microscopy
of the cells in which enough azide had been photolyzed to enable
clear observation of products, revealed a red liquid residue. No
strong absorbance from this liquid was observed in the
wavelength regions 270-420 and 550-850 nm. Micro-Raman
spectroscopy of the red liquid yields five Raman-active vibra-
Figure 4. Time (from the start of the laser pulse) to maximum
absorbance averaged at 300 nm and time (from the start of the laser
pulse) to maximum absorbance averaged at 700 nm, plotted versus the
initial pressure in the anvil cell. The magnitudes of the bracketed lines
indicate uncertainties in the measurement of time and pressure. The
solid lines are least-squares fits to the time-to-peak data from the two
wavelength regions.
-1
tional modes at 782, 1140, 1378, 1834, and 2185 cm . (See
Figure 5 and Table 1.) Unfortunately, no new IR modes were
observed, probably because of the very weak intensity of N-N
stretches or the very small product quantities present. Once
pressure is released, the red liquid vanishes, indicating that the
red liquid is stable only at high pressure and either evaporates
or decomposes to gaseous products upon release of pressure.
at 7.8 µs, indicating the formation of N6•-. The solid line shows
the laser pulse, which peaks after 4 µs and then falls off by 9
µs.
Discussion
The pressure dependence of the postulated reaction interme-
•
•-
diates N3 and N6 was determined up to 4.0 GPa (40 kbar)
and are shown in Figure 4. The error bars associated with the
time measurements arise from the rate of the streak camera.
The larger the time-axis error bars in Figure 4, the slower the
streak camera was streaked. The errors associated with the
pressure measurements are within the symbol used for data. A
linear-least-squares fit of a straight line to the UV-peak data
shows a very shallow slope that fits within the error of the time
measurements, indicating that the time to UV peak does not
depend on pressure, at least within the pressure range investi-
gated here. Instead, UV absorbance always starts rising when
During photolysis, a UV absorption peak is observed around
300 nm. Because N3 absorbs at 277 nm, this UV absorbance
•
peak is interpreted as absorbance by azide radical being
9
,10
•
produced.
Because N3 production coincides with the laser
pulse and is independent of pressure it is considered a photolytic
process. The photolytic mechanism for N3 formation is as
follows.
•
N3- + hν f N + e
•
-
3
+
-
Na + e f Na

Photolysis of Compressed NaN3
J. Phys. Chem. A, Vol. 107, No. 6, 2003 947
Once the laser pulse ends, N3• radicals are no longer
1145, and 1954 cm . Then the two “azide like” modes of 2185
-1
•
-1
continuously produced and a decrease in the N3 concentration
and 1378 cm would be attributed to either liquidlike azide or
azide ions distorted from its usual linear symmetry. Alterna-
tively, all five of the observed Raman modes could belong to
•
is observed. The concentration of N3 can decrease via any of
three pathways.
-
N7 , a W-shaped structure that includes “azide-like” linear
•
-
-
N + e f N
(a)
(b)
(c)
branching.
3
3
•
•
N f N + N
Conclusions
3
2
N + N f N6•-
•
3
-
3
The reaction mechanism and kinetics of the high-pressure
photolysis of NaN3 was investigated as a potentially novel
method of synthesizing all-N materials. Time- and wavelength-
resolved absorbance measurements indicated the formation of
two intermediates. The first intermediate absorbing in the UV
300 nm), peaks about 5 µs after the laser pulse initiates reaction
when the laser fluence is maximum, and is proposed to be N3
radicals. The second intermediate, absorbing around 750 nm,
is tentatively identified as N6 . The kinetics of the production
of this intermediate is enhanced by pressure. Therefore, the
proposed reaction mechanism for the photolysis of NaN3 is
described by the equations below.
Because metallic Na, N2 and other products are observed, N3•
most likely continues to react via pathways (b) or (c), instead
of simply converting back to the anions by pathway (a).
Because a second intermediate with absorbance centered near
(
•
7
50 nm is observed during reaction, and this intermediate is
•-
proposed to be N6 , pathway (c) is proposed as the mechanistic
path. In addition, enhanced confinement or higher pressure
•
-
•
-
•-
favors the reaction of N3 + N3 to produce N6 , as shown in
Figure 4. Therefore, pathway (c) is more consistent with La
Chatelier’s principle than pathway (b), in that the production
kinetics of a single species by the combination of two different
species should (and does) increase at higher pressures. Figure
+
-
•
Na + N f Na + N3
3
also shows that the product N6•- is a transient intermediate
and further reacts to produce other species.
N + N f N6•-
•
-
3
3
3
•
-
If N6 is produced, then several plausible mechanisms
The final products observed are metallic Na, N2, and in certain
samples, a red liquid residue stable only at pressure. The Raman
modes of the red liquid suggest it is either N7 or Cl2N6 where
•
-
involving the reaction of N6 may be suggested to constitute
the red liquid product detected.
-
•
Cl radicals are obtained from the pressure medium NaCl.
•
•-
-
N + N
f N + N7
(1)
(2a)
(2b)
3
6
2
Acknowledgment. S.M.P. thanks Dr. G. Pangilinan for his
assistance in the lab and for obtaining the Raman spectra.
•
•
N + NaCl f NaN + Cl
3
3
References and Notes
N6•- + 2Cl f Cl N + e
•
-
2
6
(
1) Vogler, A.; Wright, R. E.; Kunkley, H. Angew. Chem., Int. Ed.
Engl. 1980, 19, 717.
2) Gorini, J. A. C.; Farras, J.; Feliz, M.; Oliverella, S.; Sole, D.;
Vilarrasa, J. J. Chem. Soc., Chem. Commun. 1986, 1986, 959.
(3) Workentine, M. S.; Wagner, B. D.; Negri, F.; Zgierski, M. Z.;
Lusztyk, J.; Siebrand, W.; Wagner, D. M. J. Phys. Chem. 1995, 99, 94.
4) Christe, K. O.; Wilson, W. W. Angew. Chem., Int. Ed. Engl. 1999,
38, 2004.
(5) Huisgen, R.; Ugi, I. Angew. Chem. 1956, 68, 706.
(6) Ugi, I.; Perlinger H.; Behringer, L. Chem. Ber. 1958, 91, 2324.
7) Wallis, J. D.; Dunitz, J. D. J. Chem. Soc., Chem. Commun. 1983,
983, 910.
(8) Hayon, E.; Simic, M. J. Am. Chem. Soc. 1970, 92, 7486.
9) Burak, D.; Shapira, D.; Treinin, J. Phys. Chem. 1970, 74, 568.
10) Butler, J.; Land, E. J. Radiat. Phys. Chem. 1984, 23, 265.
11) Russell, T. P.; Piermarini, G. J. ReV. Sci. Instrum. 1997, 68, 1835.
(12) Piermarini, G. J.; Block, S.; Barnett J. D.; Forman R. S. J. Appl.
Phys. 1975, 46, 2774.
13) Russell, T. P.; Allen, T. M.; Gupta, Y. M. J. de Physique IV 1995,
, 553.
(14) Russell, T. P.; Allen, T. M.; Gupta, Y. M. Chem. Phys. Lett. 1997,
267, 351.
(15) Mowery, R. C. Private communication, 1999.
-
One potential chainlike structure is N7 , while a potential
(
15
-
ring structure could be Cl2N6. N7 could be produced by the
•
•-
reaction of N3 and N6 , as shown in reaction 1. Cl2N6 is a
feasible product as the pressure medium used is NaCl and the
(
•
•
N3 produced could react with NaCl to release Cl radicals as
•
-
in reaction 2a. Then the N6 can be stabilized in the Cl2N6
ring structure via reaction with Cl radicals as in reaction 2b.
•
(
A wide range of N-materials was considered as possible
structures of the red liquid residue. The most possible structures
were selected by comparing the calculated Raman frequencies
of these structures against the experimentally obtained Raman
frequencies of the red liquid. (The calculations were performed
using the Gaussian 98 and ACES II programs, employing either
the Hartree-Fock self-consistent-field method or the density-
functional theory (B3LYP) method, with basis sets of either
1
(
(
(
(
5
6
-311G** or 6-31G**, and scaled with the recommended
1
5
scaling factors. ) These frequencies are presented in Table 1.
Considering Cl2N6, three of the observed frequencies (781, 1140,
(
(
16) Bryant, J. I. J. Chem. Phys. 1964, 40, 3195.
17) Michels, H. H.; Montgomery, J. A.; Christe, K. O.; Dixon, D. A.
J. Phys. Chem. 1995, 99, 187.
-
1
and 1894 cm ) could correspond to the calculated modes 835,