Plasma-assisted synthesis and properties of Na3N

Article abstract of DOI:10.1021/ic700406q

Dark-blue sodium nitride, Na₃N, was prepared by the reaction of metallic sodium or liquid Na-K alloy with plasma-activated nitrogen at low pressure. The compound crystallizes in the cubic anti-ReO₃-type structure (space group Pm3m with a = 4.73301(6) A and Z = 1) according to powder and single-crystal X-ray diffraction data. Na₃N decomposes above 104°C into the elements, with ΔH_f estimated at +64(2) kJ/mol.

Full text of DOI:10.1021/ic700406q

Inorg. Chem. 2007, 46, 5146−5148
Plasma-Assisted Synthesis and Properties of Na3N†
Grigori V. Vajenine*
Max-Planck-Institut fu¨r Festko¨rperforschung, Stuttgart D-70569, Germany, and Institut fu¨r
Anorganische Chemie, UniVersita¨t Stuttgart, Stuttgart D-70569, Germany
Received March 2, 2007
Dark-blue sodium nitride, Na3N, was prepared by the reaction of
metallic sodium or liquid Na K alloy with plasma-activated nitrogen
at low pressure. The compound crystallizes in the cubic anti-ReO3-
type structure (space group Pm3m with a 4.73301(6) Å and
1) according to powder and single-crystal X-ray diffraction
data. Na3N decomposes above 104 C into the elements, with
Hf estimated at 64(2) kJ/mol.
tion of nitrogen, produced by the electric discharge” with a
specific yellow long-lived afterglow, also known as Lewis-
Rayleigh afterglow. In the following years, the meaning of
the term actiVe nitrogen was extended to include any excited
form of nitrogen, molecular or atomic, of sufficient lifetime
so that its chemical and physical properties can be studied.5
Activation of nitrogen and other gases by means of
electrical discharge received increasing attention in the 1960s
because it was found useful for semiconductor applications,
such as deposition of silicon nitride.6 This technique under-
went significant development under the term plasma actiVa-
tion.7
Though widely used industrially for thin film deposition
of known nitrides, plasma activation of nitrogen was rarely
employed in studies toward new compounds.8 We could
show recently that plasma-enhanced chemical vapor deposi-
tion, for example, can indeed lead to new compounds as in
the case of Ta2N3.9
−
h
)
Z
)
°
∆
+
Alkali metals, with the exception of lithium, react with
elemental nitrogen neither under ambient conditions nor at
higher temperatures. However, as sodium was used to purify
nitrogen for spectroscopic investigations starting in 1872,1
activation by electrical discharge was shown to lead to a
reaction between sodium and nitrogen as reported by Salet
in 1876,2 followed by a number of investigations through
the 1930s3 including studies on the other alkali metals, arc
discharge in liquid nitrogen,3d and thermal decomposition
of the corresponding azides.3f,j Although compound formation
and release of ammonia in subsequent hydrolysis were
described in these studies, no solid analytical and structural
evidence for nitride formation was presented. Only a few
years ago, Na3N4a and K3N4b could be structurally character-
ized in 100-200-nm-thick films obtained by codeposition
of atomized elements at low temperatures.
In this work, we report plasma-assisted synthesis of single-
phase Na3N on a laboratory scale, as well as its chemical
and physical properties.
Na3N was prepared in a reaction of elemental sodium
(typically pieces of ca. 0.5 g) with gaseous nitrogen activated
by capacitive high-frequency discharge (10-50 W and 13.6
MHz) at pressures of 0.1-4 mbar without external heating.
Dark-blue film of the product forms within minutes on the
metal surface at lower power. Use of higher power leads to
heating of the metal, which at some point, presumably near
the melting point of sodium, leads to growth of a micro-
crystalline dark-blue product on a 10 mg scale consisting to
In 1911, Strutt (later Lord Rayleigh) coined the term actiVe
nitrogen3e in order to describe “a chemically active modifica-
* E-mail: g.vajenine@fkf.mpg.de.
† Dedicated to Professor Roald Hoffmann on the occasion of his 70th
birthday.
(1) (a) Schuster, A. Proc. R. Soc. London 1872, 20, 484-487. (b) Schuster,
A. Philos. Mag. 1872, 44, 537-541. (c) Schuster, A. Ann. Phys. 1872,
223, 106-112.
(2) (a) Salet, G. Comp. Rend. 1876, 82, 223-226 and 274-275. (b) Salet,
G. Ann. Phys. 1876, 234, 329-334.
(3) (a) Zehnder, L. Ann. Phys. 1894, 288, 56-66. (b) Mey, K. Ann. Phys.
1903, 316, 127-145. (c) Gehlhoff, G.; Rottgardt, K. Verh. Dtsch. Phys.
Ges. 1910, 12, 492-505. (d) Fischer, F.; Schro¨ter, F. Ber. Dtsch.
Chem. Ges. 1910, 43, 1465-1479. (e) Strutt, R. J. Proc. R. Soc.
London, Ser. A 1911, 85, 219-229. (f) Suhrmann, R.; Clusius, K. Z.
Anorg. Allg. Chem. 1926, 152, 52-58. (g) Okubo, J.; Hamada, H.
Philos. Mag. 1928, 5, 372-380. (h) Okubo, J.; Hamada, H. Philos.
Mag. 1929, 7, 729-736. (i) Moldenhauer, W.; Mo¨ttig, H. Ber. Dtsch.
Chem. Ges. 1929, 62, 1954-1959. (j) Wattenberg, H. Ber. Dtsch.
Chem. Ges. 1930, 63, 1667-1672. (k) Tiede, E. Z. Elektrochem. 1935,
41, 526. (l) Tiede, E.; Knoblauch, H.-G. Ber Dtsch. Chem. Ges. 1935,
68, 1149-1154.
(5) (a) Jennings, K. R.; Linnett, J. W. Quart. ReV. 1958, 12, 116-132.
(b) Mannella, G. G. Chem. ReV. 1963, 63, 1-20. (c) Wright, A. N.;
Winkler, C. A. ActiVe Nitrogen; Academic Press: New York, 1968.
(d) Brown, G. R.; Winkler, C. A. Angew. Chem. 1970, 82, 187-202;
Angew. Chem., Int. Ed. Engl. 1970, 9, 181-196.
(6) (a) Pierson, H. O. Handbook of Chemical Vapor Deposition, 2nd ed.;
Noyes Publications: Berkshire, U.K., 1999. (b) Konuma, M. Plasma
Techniques for Film Deposition; Alpha Science: Oxford, U.K., 2005.
(7) Synonymous descriptions plasma-assisted (PA), plasma-enhanced
(PE), and plasma-induced (PI) are also used interchangeably in the
literature in relation to plasma-activated processes.
(8) (a) Blucher, J.; Bang, K.; Giessen, B. C. Mater. Sci. Eng. 1989, A117,
L1-L3. (b) Prange, R.; Cremer, R.; Neuschu¨tz, D. Surf. Coat. Technol.
2000, 133-134, 208-214. (c) Schiefenho¨vel, N.; Himmel, H.-J.;
Binnewies, M. Angew. Chem. 2003, 115, 6138-6139; Angew. Chem.,
Int. Ed. 2003, 42, 5956-5957. (d) Schiefenho¨vel, N.; Himmel, H.-J.;
Binnewies, M. Z. Anorg. Allg. Chem. 2005, 631, 1507-1511.
(9) Ganin, A. Yu.; Kienle, L.; Vajenine, G. V. Eur. J. Inorg. Chem. 2004,
3233-3239.
(4) (a) Fischer, D.; Jansen, M. Angew. Chem. 2002, 114, 1831-1833;
Angew. Chem., Int. Ed. 2002, 41, 1755-1756. (b) Fischer, D.;
Cancarevic, Z.; Scho¨n, J. C.; Jansen, M. Z. Anorg. Allg. Chem. 2004,
630, 156-160.
5146 Inorganic Chemistry, Vol. 46, No. 13, 2007
10.1021/ic700406q CCC: $37.00
© 2007 American Chemical Society
Published on Web 05/26/2007

COMMUNICATION
over 90 wt % of cubic Na3N contaminated predominantly
by metallic sodium and NaN3 according to powder X-ray
diffraction analysis. External heating of sodium above its
melting point appears to hinder the formation of the nitride;
instead, only the azide is formed.
The formation of Na3N can be further facilitated by the
use of a liquid Na-K alloy as an alkali-metal source.10
Typically, 2 g of such an alloy with compositions from
Na0.5K0.5 to Na0.2K0.8 was used. The reaction progress could
be followed by a pressure decrease down to 0.01 mbar,
resulting in the disappearance of discharge when the system
was closed. Qualitatively, the nitrogen pressure has little
effect on the absorption rate, whereas the reaction rate scales
nearly linearly with the power of the high-frequency source.
The remaining liquid alloy was removed from the product
by a pipet, and the product was washed with fresh alloy and
separated from the liquid phase in a centrifuge. A typical
experiment at 50 W over 8 h yields ca. 100 mg of Na3N
contaminated only by ca. 1 wt % of elemental sodium
according to powder X-ray diffraction analysis. Single
crystals of Na3N up to 0.3 mm in size could also be isolated
from the reaction products obtained with or without ad-
ditional potassium.
The product is very air-sensitive. Caution! Rapid exo-
thermal oxidation of excess alkali metals on the surface of
Na3N particles can lead to decomposition of the nitride
accompanied by a Violent combustion. Samples of eVen a
few milligrams should be disposed of carefully!
Controlled oxidation of the alkali-metal excess by a small
amount of dry oxygen followed by manual removal of the
oxidized layer could also be used to purify the nitride. Na3N
itself appears to be stable with respect to pure oxygen.
Careful hydrolysis of the nitride yields only sodium
hydroxide and ammonia, indirectly confirmed by X-ray
diffraction analysis of the products after neutralization by
HCl. Gravimetric analysis of the resulting NaCl and NH4Cl
confirmed the Na3N stoichiometry.12
Figure 1. Crystal structure of Na3N with thermal ellipsoids drawn at 90%
probability.
with Na-N distances of 2.3665 Å. Anisotropic refinement
of sodium thermal displacement parameters indicated pre-
ferred thermal motion perpendicular to the N-Na-N axis
[U22 ) U33 ) 0.0611(7) Å2] compared to the direction along
the Na-N contacts [U11 ) 0.0180(6) Å2, values from single-
crystal data]. However, no structural phase transition was
observed for Na3N down to -182 °C according to powder
X-ray diffraction, in agreement with the earlier study.4a Only
insignificant rest electron density (a peak of 0.275 e/Å3) was
found in the large void in the middle of the unit cell, ruling
out possible occupation of the A site according to the
perovskite ABX3 structure type. Powder X-ray diffraction
data collected on a single-phase sample of Na3N could also
be used for structure refinement (Figure 2), with the results
in agreement with the single-crystal data.
Previous studies reported different colors for sodium
nitride, ranging from gray, metallic, dark brown, dark blue
to red, while single-phase films of Na3N were observed to
be red-brown.4a All of our samples are dark blue as prepared
from sodium and nitrogen or after cleaning of the surface of
the reaction product from the liquid Na-K alloy. During
hydrolysis of the nitride, no color change is seen: dark-blue
microcrystals could be observed for minutes under an optical
microscope slowly evolving gas and finally disappearing in
No signs of decomposition of Na3N could be observed at
room temperature over several weeks. X-ray diffraction
analysis at higher temperatures using a Simon-Guinier
camera13 with a heating rate of 0.15 °C/min shows decom-
position at 104 °C, which agrees well with the previously
reported value of 96 °C.4a Upon cooling of these samples,
metallic sodium was identified as the major decomposition
product, contaminated by small amounts of NaN3. Dif-
ferential thermal analysis (DTA) measurements with a higher
rate of heating (3 °C/min) indicate an onset of exothermal
decomposition at ca. 140 °C. The latter analysis also yielded
a preliminary value for ∆Hf(Na3N) of +64(2) kJ/mol.
X-ray diffraction analysis of both single-crystal and
powder samples of Na3N obtained with or without additional
potassium confirmed the previously reported anti-ReO3-type
structure4a (Figure 1) in the space group Pm3hm with a )
4.73301(6) Å based on powder diffraction data [compared
to the literature value of 4.7325(2) Å4a] and atoms in the
(10) A reaction between a liquid Na-K alloy and active nitrogen was
already observed earlier, and the formation of sodium nitride and azide
was reported without further characterization.3b,k,l Such a liquid alloy
has also been shown to be an effective flux in the synthesis of Na-
Ba subnitrides and suboxides.11
(11) (a) Steinbrenner, U.; Simon, A. Angew. Chem. 1996, 108, 595-597;
Angew. Chem., Int. Ed. Engl. 1996, 35, 552-554. (b) Simon, A.;
Steinbrenner, U. J. Chem. Soc., Faraday Trans. 1996, 92, 2117-2123.
(c) Vajenine, G. V.; Steinbrenner, U.; Simon, A. C. R. Acad. Sci.,
Ser. IIc: Chim. 1999, 2, 583-589. (d) Vajenine, G. V.; Simon, A.
Eur. J. Inorg. Chem. 2001, 1189-1193. (e) Vajenine, G. V.; Simon,
A. Angew. Chem. 2001, 113, 4348-4352; Angew. Chem., Int. Ed.
2001, 40, 4220-4222. (f) Simon, A. In Molecular Clusters of the
Main Group Elements; Driess, M., No¨th, N., Eds.; Wiley-VCH:
Weinheim, Germany, 2004; pp 246-266.
(12) Hydrolysis of a 30 mg sample followed by absorption of the evolving
NH3 and neutralization by HCl yielded amounts of NaCl and NH4Cl
corresponding to 85(1) wt % of sodium and 14(1) wt % of nitrogen
(expected 83.1 and 16.9 wt %). According to powder X-ray diffraction,
NH4Cl was phase-pure, while NaCl contained ca. 1-2 wt % of KCl,
suggesting contamination of the Na3N sample by a small amount of
the Na-K alloy. See also the Supporting Information.
1
special 1a 0,0,0 (N) and 3d /2,0,0 (Na) positions. Nitrogen
atoms are surrounded by octahedra of sodium atoms; the
latter are each coordinated linearly by two nitrogen atoms
(13) Simon, A. J. Appl. Crystallogr. 1971, 4, 138-145.
Inorganic Chemistry, Vol. 46, No. 13, 2007 5147

COMMUNICATION
binary oxides: more negative ∆Hf° is usually accompanied
by a larger band gap.20 Though no similar study for binary
nitrides has appeared so far, a comparison of the ∆Hf° values
for Li3N (-164.56 kJ/mol21), Cu3N (+74.4 kJ/mol21), and
Na3N (+64 kJ/mol) suggests that the band gap in Na3N
should be closer to the value for Cu3N (∼0.5 eV according
to theoretical estimates22 compared to the optical band gap
of 1.5-1.65 eV in thin films23) rather than to that for Li3N
(2.1 eV19).
Preliminary band structure calculations on Na3N at the
density functional theory (DFT) level result in an unrealistic
overlapping of the valence and conduction bands by 0.6 eV.
Because DFT is known to underestimate band gaps by ca.
1-2 eV,24 a band gap of around 0.5-1 eV can be expected
for Na3N.
Na3N is likely to be the most ionic binary nitride stable at
room temperature, yet it appears to have a band gap typical
for a semiconductor. The open anti-ReO3-type structure
suggests the possibility of intercalation, Na+ diffusion, and
structural transitions under high pressure. The metastable
compound might act as an efficient nitride anion donor in
exchange reactions. Plasma-assisted synthesis gives access
to Na3N in amounts sufficient for further study of its physical
and chemical properties, which are being currently investi-
gated.
Acknowledgment. Thanks are due to C. Hoch for
collecting the single-crystal diffraction data, W. Ho¨lle for
carrying out temperature-dependent Guinier measurements,
R. Eger and V. Smetana for help with DTA investigations,
and Prof. A. Simon for continuous support.
Figure 2. Results of Rietveld refinement for a single-phase sample of
Na3N: a comparison of measured (open circles) and calculated intensities
with a difference curve below and vertical lines marking the reflection
positions.
a colorless surrounding of wet sodium hydroxide. These
observations lead us to conclude that the deep color of the
Na3N samples is not due to nonstoichiometry or alkali-metal
dispersion in Na3N but rather is likely to be an intrinsic
property.
Such deep coloring of a supposedly strongly ionic
compound between electropositive sodium and electro-
negative nitrogen might seem counterintuitive at first.
However, the formal ionic description (Na+)3N3- assumes a
nitride anion, for which the lattice energy should compensate
the sum of the first three (negative) electron affinities.
Electron-electron repulsion in the highly charged anion leads
to high-lying filled N 2p states comprising the valence band
with only a relatively small band gap to the empty Na 3s
states of the conduction band.
Supporting Information Available: Experimental details,
X-ray diffraction data on single-crystal and powder samples, a CIF
file for Na3N, DTA investigations, and further results of chemical
analysis. This material is available free of charge via the Internet
at http://pubs.acs.org.
IC700406Q
(14) Rauch, W. Z. Phys. 1940, 116, 652-656.
A comparison of ionic lithium and sodium compounds is
useful here. Excitation of an electron across the band gap
may be simplistically viewed as a charge transfer from anion
to cation with the energy cost of the electron affinity and
lattice energy reduced by the ionization energy of the
cation.14 In pairs of related compounds, LiF/NaF, Li2O/Na2O,
and Li3N/Na3N, the effect of the anion’s electron affinity is
cancelled, while the first ionization energies of lithium (5.39
eV) and sodium (5.14 eV)15 are quite similar. Consequently,
the reduction of the band gaps in the first two pairs from
lithium to sodium compounds (14.5 eV in LiF16 to 11.7 eV
in NaF17 and from ∼7 eV in Li2O14,18 to ∼4 eV in Na2O14)
can be qualitatively explained by higher lattice energy and
thus better electrostatic stabilization of the 2p anion states
in the case of lithium compared to the isostructural sodium
compounds with longer cation-anion distances. Because red
R-Li3N has a band gap of 2.1 eV,19 one could also expect a
smaller band gap and thus a deeper color for Na3N, even
though the two are not isostructural.
(15) Lide, D. R., Ed. Chemical Rubber Company Handbook of Chemistry
and Physics, 79th ed.; CRC Press: Boca Raton, FL, 1998.
(16) Piacentini, M. Solid State Commun. 1975, 17, 697-700.
(17) Nakai, S.; Sagawa, T. J. Phys. Soc. Jpn. 1969, 26, 1427-1434.
(18) (a) Nasu, S.; Tanifuji, T.; Kawatsura, K.; Uchida, K.; Kikuchi, A. J.
Nucl. Mater. 1978, 73, 127-128. (b) Uchida, K.; Noda, K.; Tanifuji,
T.; Nasu, S.; Kirihara, T.; Kikuchi, A. Phys. Status Solidi A 1980, 58,
557-566. (c) Liu, L.; Henrich, V. E.; Ellis, W. P.; Shindo, I. Phys.
ReV. B 1996, 54, 2236-2239. (d) Ishii, Y.; Murakami, J.; Itoh, M. J.
Phys. Soc. Jpn. 1999, 68, 696-697.
(19) (a) Brendecke, H.; Bludau, W. J. Appl. Phys. 1979, 50, 4743-4746.
(b) Beister, H. J.; Haag, S.; Kniep, R.; Stro¨ssner, K.; Syassen, K.
Angew. Chem. 1988, 100, 1116-1118; Angew. Chem., Int. Ed. Engl.
1988, 27, 1101-1103.
(20) (b) Portier, J.; Campet, G.; Kwon, C. W.; Etourneau, J.; Subramanian,
M. A. Int. J. Inorg. Mater. 2001, 3, 1091-1094. (b) Portier, J.; Hilal,
H. S.; Saadeddin, I.; Hwang, S. J.; Subramanian, M. A.; Campet, G.
Prog. Solid State Chem. 2004, 32, 207-217.
(21) Blachnik, R., Ed. D’Ans Lax Taschenbuch fu¨r Chemiker und Physiker,
4th ed.; Springer: Berlin, 1998.
(22) (a) Hahn, U.; Weber, W. Phys. ReV. B 1996, 53, 12684-12693. (b)
Yu, W.; Zhao, J.; Jin, C. Phys. ReV. B 2005, 72, 214116-1-214116-
8.
(23) (a) Kim, K. J.; Kim, J. H.; Kang, J. H. J. Cryst. Growth 2001, 222,
767-772. (b) Borsa, D. M.; Boerma, D. O. Surf. Sci. 2004, 548, 95-
105.
(24) (a) Johnson, K. A.; Ashcroft, N. W. Phys. ReV. B 1998, 58, 15548-
15556. (b) Aryasetiawan, F.; Gunnarsson, O. Rep. Prog. Phys. 1998,
61, 237-312.
An interesting correlation between the band gap and
enthalpy of formation has been pointed out recently for
5148 Inorganic Chemistry, Vol. 46, No. 13, 2007