Reactions of alkaline earth metals and nitrogen in sealed niobium ampoules: The formation of MgZn2 type intermetallic phases in the presence of nitrogen and the new compound Ba5[NbN4]N

Article abstract of DOI:10.1016/j.jallcom.2004.04.082

Reactions between alkaline earth metals or their corresponding nitrides and NaN₃ as nitrogen donor in cleaned sealed Nb ampoules at different temperatures yields known phases, 'β-Ca₃N₂' and single crystals in the form of transparent, orange plates of Ba ₅[NbN₄]N. The crystal structure of this new compound was determined by means of single crystal X-ray diffraction (C2/m (no. 12), a = 1231.7(3), b = 1094.6(2), c = 853.8(2) pm,β = 113.65(3) and Z = 4). At temperatures ≥ 1000°C, the available nitrogen apparently reacts with the Nb container walls and intermetallic phases are formed. The crystal structures of AeMg₂ (MgZn₂ type with Ae = Ca, Yb, Sr, Eu and Ba) were re-determined by X-ray single crystal structure analyses on crystals obtained in such reactions to ascertain the presence or absence of nitrogen in structural voids. The structures of the ternary nitride and intermetallic phases are described and compared to known compounds.

Full text of DOI:10.1016/j.jallcom.2004.04.082

Journal of Alloys and Compounds 384 (2004) 98–105
Reactions of alkaline earth metals and nitrogen in sealed niobium
ampoules: the formation of MgZn2 type intermetallic phases in the
presence of nitrogen and the new compound Ba5[NbN4]N
Olaf Reckewega,b,∗, Cora Lindc, Arndt Simona, Francis J. DiSalvob
a
Max-Planck-Institut für Festkörperforschung, Heisenbergstr. 1, D-70569 Stuttgart, Germany
b
Baker Laboratory, Department of Chemistry and Chemical Biology, Cornell University, Ithaca, NY 14853-1301, USA
c
Department of Chemistry, MS 602, University of Toledo, Toledo, OH 43606-3390, USA
Received 8 December 2003; received in revised form 12 April 2004; accepted 12 April 2004
Abstract
Reactions between alkaline earth metals or their corresponding nitrides and NaN3 as nitrogen donor in cleaned sealed Nb ampoules at
different temperatures yields known phases, ‘␤-Ca3N2’ and single crystals in the form of transparent, orange plates of Ba5[NbN4]N. The
crystal structure of this new compound was determined by means of single crystal X-ray diffraction (C2/m (no. 12), a = 1231.7(3), b =
1094.6(2), c = 853.8(2) pm, β = 113.65(3)◦ and Z = 4). At temperatures ≥1000 ◦C, the available nitrogen apparently reacts with the Nb
container walls and intermetallic phases are formed. The crystal structures of AeMg2 (MgZn2 type with Ae = Ca, Yb, Sr, Eu and Ba) were
re-determined by X-ray single crystal structure analyses on crystals obtained in such reactions to ascertain the presence or absence of nitrogen
in structural voids. The structures of the ternary nitride and intermetallic phases are described and compared to known compounds.
© 2004 Elsevier B.V. All rights reserved.
Keywords: Alkaline earth metal; Niobium; Nitride; MgZn2; Laves phase; Structure elucidation
1. Introduction
source of readily available nitrogen. The products obtained
were determined as a function of the reaction temperature.
Ytterbium and europium, which frequently show behavior
very similar to alkaline earth metals in nitride compounds,
were included in the list of alkaline earth metals for conve-
nience. The results of this series of reactions are presented
here.
Structural refinements on X-ray single crystal data of
AeMg2 (Ae = Ca, Yb, Sr, Eu and Ba) alloys (which were
among the products of some reactions) were used to address
the potential presence of nitrogen in structural voids. No ni-
trogen is detected in these intermetallic phases.
Many different reaction techniques have been employed
for the syntheses of nitrides. They can roughly be divided
into sealed- and open-vessel, the latter with flow-by or
flow-through of a nitrogen source. Both strategies have
their specific advantages and disadvantages, which will not
be discussed here. Regardless of the strategy employed,
cleaned niobium and tantalum are frequently used as ‘inert’
container materials. However, in many cases the con-
tainer material is incorporated into the reaction products,
e.g. in the case of Ba(Mg3.33M0.67)N4 (M = Nb or Ta)
[1].
In an attempt to better understand the reactivity of the con-
tainer materials, we carried out a series of reactions where
alkaline earth metals and/or alkaline earth metal nitride re-
actants were sealed in niobium ampoules with NaN3 as a
2. Experimental section
All manipulations were carried out under a continuously
purified and monitored argon atmosphere in glove boxes.
The reactions took place in thoroughly cleaned (cleaning
bath consists of 4:2:1 volume ratio of H2SO4 (conc.), HNO3
(fuming) and HF (40%)) arc-welded Nb tubes (99.99%,
Plansee, Germany).
∗
Corresponding author. Tel.: +1-607-255-4164;
fax: +1-607-255-4137.
E-mail address: fjd3@cornell.edu (O. Reckeweg).
0925-8388/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jallcom.2004.04.082

O. Reckeweg et al. / Journal of Alloys and Compounds 384 (2004) 98–105
99
Mg (chips, 99.98%, Aldrich), Ca (<20 ppmw C,
<30 ppmw H and <100 ppmw O, dendritic pieces, Ames
Laboratory, Materials Preparation Center), Sr (99.95%,
dendritic pieces, Strem), Ba (99.9%, dendritic, Strem), Eu
(99.9%, powder), and Yb (99.9%, sublimed dendritic) were
used as received, NaN3 (99%, powder, Aldrich) was dried at
400 K under dynamic vacuum for 2 h prior to use, Mg3N2,
Ca3N2, Sr2N, Ba2N, EuN and YbN were synthesized fol-
lowing literature methods [2,3]. To verify the nature of
these binary nitrides, powder X-ray diffractograms of all
products were recorded on an Inel powder diffractometer
(ground materials were loaded in 0.2 mm, thin-walled glass
capillaries). No impurities were detected by this method
unless stated otherwise.
Bruker Smart CCD using Mo K␣ radiation. If the quality
was found to be sufficient, a set of intensity data was col-
lected on the same machine. All processing of the data, the
structure solutions and refinements were carried out using
Bruker Smart CCD software [4]. Selected parameters of the
measurements and results of the refinements are summarized
in Tables 1–4. Further details of the crystal structure inves-
tigations may be obtained from the Fachinformationszen-
trum Karlsruhe, Eggenstein-Leopoldshafen, Germany (fax:
+49-7247-808-666; e-mail: crysdata@fiz-karlsruhe.de),
on quoting the depository numbers CSD-412667 (Ba5
[NbN4]N), 412683 (CaMg2), 416684 (SrMg2), 412681
(BaMg2), 412666 (YbMg2) and 412689 (EuMg2).
The compounds AeMg2 (Ae = Ca, Yb, Sr, Eu and Ba)
are moderately air-stable, Ba5[NbN4]N is air and moisture
sensitive and decomposes if exposed to air.
The reactions were carried out in the following manner:
ꢀ
stoichiometric amounts of the starting materials ( m ≈
500 mg, the nitrogen content as indicated by the nominal
starting composition was supplied by the addition of proper
amounts of NaN3 if necessary) as listed in Table 1 were
filled in cleaned niobium ampoules of about 5 cm length
without compacting or grinding the starting materials. The
arc-welded metal tubes were enclosed in evacuated and fused
silica tubes to protect the metal containers from oxidation by
air during heating. The reaction vessels were placed upright
in a box furnace, which was heated within 24 h to reaction
temperatures of 600, 800, 1000 or 1200 ◦C. The samples
were kept at the respective temperature for 3 days. Then
the furnace was switched off and allowed to cool to room
temperature. The higher the reaction temperature was set, the
more brittle the metal container became and the more the Nb
lost its metallic luster on the inner and outer tube surfaces.
4. Discussion
4.1. The starting materials
The identification of the binary nitrides from their re-
spective powder diffractogram was straightforward for
Mg3N2 [2], ␣-Ca3N2 [2], EuN and YbN [9]. Difficulties
were encountered for Ae2N compounds (Ae = Ca, Sr or
Ba; Table 5) [3,10–16]. The measured intensities differed
considerably from the simulated pattern due to preferred
orientation of the crystallites, which usually form as hexag-
onal plates. Additionally, the c lattice constants showed
a strong dependence on the reaction temperature. The c
axis length decreases with increasing temperature until de-
composition takes place. The insertion of small amounts
of hydride in the Ae2N compounds was considered as an
explanation for this phenomena [3], because a considerable
decrease of the c axes can be observed due to the formation
of Ae2NH, which becomes smaller with increasing reaction
temperature. However, if only small amounts of hydride are
present both phases form in equilibrium with each other
and are identified by their respective X-ray powder diffrac-
togram. So the most probable explanation seems to be the
insertion of nitrogen in form of diazenide (N22−) species at
lower temperature, which break up with increasing temper-
ature [17–19]. The dependence of the length of the c axes
on the reaction temperature was already reported for Ca2N
[3,10–12], and inspection of the literature reveals variations
of the c lattice parameters for Sr2N [3,13,14] and Ba2N
[3,8,15,16] as well. Table 5 shows the range of the lattice
constants that have been reported so far.
2.1. Ba5NbN5
ꢀ
Ba2N and NaN3 (mixed in a 1:2 molar ratio ( m ≈
420 mg) and sealed inside an arc-welded niobium container
which was enclosed inside an evacuated silica tube) were
heated to 625 ◦C within 24 h. After 10 days at this temper-
ature, the furnace was cooled to room temperature over a
period of 24 h. The reaction product contained mainly unre-
acted Ba2N as identified by its X-ray powder pattern, metal-
lic Na and some dark orange, transparent plates of only
moderate crystallinity.
3. X-ray investigations
To verify the nature of the products, powder X-ray diffrac-
tograms of all products were measured on an Inel powder
diffractometer (ground materials were loaded and sealed in
0.2 mm, thin-walled glass capillaries). Suitable single crys-
tals of Ba5[NbN4]N and of the Laves phases of the MgZn2
type (AeMg2 with Ae = Ca, Yb, Sr, Eu and Ba) were se-
lected in an argon filled glove box under a microscope and
sealed in thin-walled glass capillaries. The crystals were
checked for their quality by taking rotation frames on a
4.2. The behavior of alkaline earth metals with nitrogen in
sealed niobium containers
The results of our experiments on the reaction of alkaline
earth metals with nitrogen in sealed Nb containers are shown
in Table 6. Table 7 gives an overview of selected physical

Table 1
Parameters of the X-ray single crystal structure determinations and selected bond length of MgZn2 type compounds
Compounds
CaMg2
YbMg2
412666
SrMg2
EuMg2
412689
BaMg2
412681
CSD-number
Space group, Z
Lattice constant
a (pm)
c (pm)
a (pm)
412683
412684
P63/mmc (#194), 4
625.28(6)
1014.35(9)
623
623.38(3)
1010.01(6)
624
648.45(8)
1045.58(14)
643.9
638.77(9)
1032.23(18)
640
667.86(4)
1061.33(7)
664.9
c (pm), Refs. [5–7]
V (Å3)
1012
343.45
176.0
1.89
1.715
1008
339.91
376.0
27.57
4.331
1049.4
380.75
248.0
14.21
2.377
1035
364.75
348.0
17.25
3.653
1067.6
409.97
320.0
9.74
3.013
F(0 0 0)
µ (mm−1
)
)
ρ (g cm−3
Refined parameters
10
10
10
11
11
Measured/unique reflections
Absence/restr./inconsistent refl. 0/0/0
Ranges
2417/257
6103/268
38/0/52
3555/473
0/0/0
5666/278
0/0/10
27455/347
35/0/186
h, k, l
2θmax
−9 to 7, −9 to 8, −15 to 15 −9 to 9, −9 to 9, −14 to 14 −9 to 9, −10 to 11, −16 to 16 −9 to 9, −8 to 9, −14 to 14 −10 to 10, −10 to 10, −16 to 16
<63.96
<64.97
<81.87
<65.12
<67.17
Rint
0.0270
0.0673
0.0548
0.0340
0.0879
GooF
1.362
1.405
1.051
0.0571/0.0516
0.0225/0
0.47 (1.87 A from Mg2),
−0.71 (1.20 A from Mg1)
1.303
1.189
R1/wR2 (all data)a
Weight factors x, yb
Residual electron density
0.0347/0.0527
0.0154/0.1917
0.27 (1.92 A from Mg2),
−0.23 (1.88 A from Mg2)
0.0349/0.0694
0.0057/8.2440
2.42 (0.79 A from Yb),
−2.19 (0.48 A from Mg1)
0.0339/0.0568
0.0000/5.8705
1.25 (1.50 A from Eu),
−1.73 (1.97 A from Mg2)
0.0300/0.0734
0.0361/1.3479
1.48 (0.98 A from Ba),
−1.31 (0.65 A from Ba)
4e
6h (3x)
6h (6x)
2a (3x)
4e (3x)
4e (1x)
2a
3.6313(8)
3.6626(4)
3.6643(4)
382.24(5)
381.55(13)
3.608(3)
3.7474(8)
3.7986(4)
3.7973(5)
395.33(5)
395.83(9)
3.691(2)
3.814(1)
3.6522(2)
3.6522(2)
380.72(4)
380.87(13)
3.7471(5)
3.7399(5)
389.14(6)
391.93(13)
3.9169(3)
3.9033(2)
404.24(3)
409.30(6)
6h (6x)
4e (6x)
6h
3.1262(6)
3.6643(4)
3.120(3)
3.6522(2)
3.2175(7)
3.7973(5)
3.176(3)
3.7399(5)
3.250(1)
3.9033(2)
6h (2x)
2a (2x)
6h (2x)
4e (2x)
4e (4x)
3.086(2)
3.1262(6)
3.166(2)
3.6313(8)
3.6626(4)
3.061(9)
3.120(3)
3.173(9)
3.608(3)
3.6527(4)
3.235(2)
3.2175(7)
3.245(2)
3.7474(8)
3.7986(4)
3.207(7)
3.176(3)
318.1(7)
3.691(2)
3.7471(5)
3.250(4)
3.250(1)
3.429(4)
3.814(1)
3.9169(3)
ꢁ
ꢂ1/2
ꢀ
ꢀ
ꢀ
ꢀ
w(F02 − Fc2)2/ (wF20)2
.
a
R1 = ||F0| − |F0||/ |F0|; wR2 =
The weight factor is defined as: w = 1/[σ2(F02) + (xP)2 + yP] with P = [(F02) + 2Fc2]/3.
b

O. Reckeweg et al. / Journal of Alloys and Compounds 384 (2004) 98–105
101
Table 2
Crystallographic coordinates, refined site occupancy coefficients and atomic displacement parameters of the MgZn2 type compounds
Wyckoff
site
Atom
Refined
occupancy (%)
x/a
y/b
z/c
U11
U22
U33
U12
80(1)
Ueq
CaMg2 [5]
YbMg2 [6]
SrMg2 [7]
EuMg2 [6]
BaMg2 [7]
4e
Ca
100(5)
1/3
2/3
0.06193(6)
0.063
0
160(2)
U11
172(3)
164(2)
2a
6h
Mg1
Mg2
100(6)
101(6)
0
0
187(5)
174(3)
U11
126(4)
120(6)
183(4)
94(2)
63(2)
165(3)
166(2)
0.83119(9)
0.833
0
2x
2x
0
0.25
2b
4e
“N”
Yb
0.02(40)
100(8)
0.25
1/3
2/3
0.06146(6)
0.063
0
162(3)
U11
91(4)
81(1)
61(9)
139(3)
2a
6h
Mg1
Mg2
100(8)
101(8)
0
0
Refined only isotropically
177(14)
122(13)
146(8)
0.8303(5)
0.833
0
2x
2x
0
0.25
123(18)
122(17)
2b
4e
“N”
Sr
6(7)
0.25
100(4)
1/3
2/3
0.06072(3)
160(1)
U11
180(1)
80(1)
166(1)
0.07
0
0.25
2a
6h
Mg1
Mg2
102(4)
101(4)
0
0
186(5)
185(3)
U11
165(3)
142(9)
190(5)
93(3)
83(2)
172(3)
182(2)
0.8330(1)
0.83
0
2x
2x
0
2b
4e
“N”
Eu
7(5)
0.25
100(6)
1/3
2/3
0.06016(6)
0.063
0
164(2)
U11
167(3)
82(1)
165(2)
2a
6h
Mg1
Mg2
99(7)
101(6)
0
0
188(16)
192(11)
U11
153(14)
80(21)
169(13)
94(8)
76(7)
152(10)
176(6)
0.8326(4)
0.833
0
2x
2x
0
0.25
2b
4e
“N”
Ba
0.3(13)
100(5)
0.25
1/3
2/3
0.05716(4)
0.063
0
234(2)
U11
246(3)
117(1)
238(2)
2a
6h
Mg1
Mg2
101(5)
101(5)
0
0
229(8)
264(7)
U11
239(9)
216(13)
233(8)
115(4)
120(5)
225(6)
248(4)
0.8378(2)
0.833
0
2x
2x
0
0.25
2b
“N”
−2.2(7)
0.25
Coordinates in bold are literature data.
For the second series of reactions with the nominal com-
position ‘AeMg2N2’ the respective metals and NaN3 as a
nitrogen source were used as starting materials. The gen-
eral tendency found in the first series of reactions (reactiv-
ity Sr, Ba ꢀ Ca > Mg, Yb, Eu) could be observed here
as well. The quite stable ternary nitrides AeMg2N2 (Ae =
Ca, Sr) [2,25] were the first products that formed as trans-
parent, pale yellow hexagonal plates. At higher tempera-
ture it appears that nitrogen reacts with the Nb container
walls and AeMg2 phases were obtained. The ternary nitrides
AeMg2N2 (Ae = Ca, Sr) were stable to nitrogen loss to
somewhat higher temperatures than Ae3N2. A few crystals
of Ba(Mg3.33Nb0.67)N4 [1] were found in one reaction with
Ba, Mg and NaN3 at T = 1000 ◦C, but as mentioned in
the literature, the crystallinity was poor and the compound
could not be detected in the respective powder X-ray diffrac-
togram.
properties of Ae, AeMg2 and Ae–Mg–N compounds for
comparison. The compounds found in the X-ray powder
diffractogram are listed along with the fraction of each, as
estimated from the relative diffraction intensities.
If only one metal was employed with NaN3 and the nom-
inal starting composition was ‘AeN’, almost every reaction
turned out as expected. Sr and Ba reacted readily even at low
temperatures to form the subnitrides with concomitant attack
of the container walls. The reaction at T = 600 ◦C yielded
the first crystal of Ba5[NbN4]N but with very poor quality;
however, we never found any evidence for this compound
in the powder X-ray diffractograms showing that this is a
minority phase (≤5%). At T = 1200 ◦C orange-red pow-
ders of Ae2NbN3 (Ae = Sr or Ba) [24,27] were found in
the Sr and Ba reactions. Mg, Yb and Eu were less reactive,
but formed the quite stable binary phases Mg3N2, YbN or
EuN, respectively, above 800 ◦C. Ca showed higher reactiv-
ity than the previous three elements, at temperatures below
approximately T = 650 ◦C black ‘␤-Ca3N2’ formed, above
this temperature the dark brown-red ␣-Ca3N2 was recov-
ered. With increasing temperature and nitrogen loss to the
container walls, increasing amounts of the subnitride Ca2N
could be found along with ␣-Ca3N2.
A third series of reactions was carried out using binary
nitrides as starting materials. The reactivity of the binary
nitrides with Mg3N2 was poor with the exception of Sr2N,
which formed SrMg2N2 at T = 800 ◦C. Mixtures of Mg3N2
with YbN and EuN remained inert up to 1000 ◦C, at which
temperature formation of AeMg2 alloys started.

102
O. Reckeweg et al. / Journal of Alloys and Compounds 384 (2004) 98–105
Table 3
Table 5
Parameters of the X-ray single crystal structure determination on
Ba5[NbN4]N
Ranges for the a and c lattice parameters of Ca2N, Sr2N and Ba2N
(literature and own results)
Compound
Ba5[NbN4]N
Compound
a (pm)
c (pm)
Reference
Refined stoichiometry
Space group (no.), Z
CSD-number
Ba5[NbN4]N0.7(1)
C2/m (12), 4
412072
Ca2N
Sr2N
Ba2N
361.6–363.8
384.8–386.3
401.3–404.6
1878.0–1913.8
2061.0–2072.5
2241.0–2269.0
[3,10–12]
[3,13,14]
[3,8,15,16]
Lattice constants
a, b, c (pm)
¯
Crystallographic data for Ae2N are: R3m (#166), Ae on 6c with z ≈
1231.7(3), 1094.6(2), 853.8(2)
90, 113.65(3), 90
5.332
Transparent orange
α, β, γ (◦)
0.268 2 and N on 3a.
Calculated density (g cm−3
Crystal colour
Crystal form
Crystal size (mm3)
Diffractometer
)
also moderately air-stable and was believed to be the alloy
Ag8Ca3 [29] for many years.
Plate
0.15 × 0.11 × 0.05
Previously reported crystallographic data for most of the
AeMg2 phases [5–7,21] has been derived from the refine-
ment of Guinier powder diffractograms. The compounds
AeMg2 (Ae = Ca, Yb, Sr, Eu and Ba) are isostructural with
the hexagonal Laves phase MgZn2. In AeMg2, Mg forms
columns of corner sharing trigonal bipyramids parallel to the
c axis and the Ae atoms are located in the channels formed
by the Mg columns (Fig. 1). The Mg atoms have 6 Mg and 6
Ae nearest neighbors, while Ae have a 16-fold coordination
sphere consisting of 12 Mg and a nearly perfect tetrahedron
of 4 Ae closest neighbors. Emphasizing the Ae coordina-
tion, the crystal structure might also be described as hexag-
onal diamond analogue Ae sublattice in which the columns
of corner sharing Mg5 bipyramids are embedded. The mo-
lar volumes of the AeMg2 compounds are slightly smaller
than the sum of the atomic volumes VAe + 2VMg (Table 8).
As the differences in electronegativity between the partici-
pating elements are small, the phases AeMg2 should follow
the Pauling-Simon law [31,32]. The validity of this law was
demonstrated earlier for CaMg2, SrMg2 and BaMg2 [33].
Fig. 2 shows a strain parameter diagram of MX2. Laves
phases formed from only electropositive metals. Their dis-
tribution over a large range of radius ratio RM/RN is well
represented by the line [31] y = 0.645x − 0.767 (x =
RM/RN, y = SP = [2RM − d(MM)]/2RN with d(MM) rep-
resenting the experimentally determined mean distance be-
tween M atoms). Both EuMg2 and YbMg2 follow the rule
Bruker Smart CCD
Radiation, monochromator,
temperature (K)
Ranges [2θmax]; h, k, l
Mo K␣ (λ = 71,073 pm),
graphite, 293(2)
50.04◦; 14, −10 → 13, 10
Distance detector − crystal (mm)
50
0.3
30
Increment, ꢀϕ (◦)
Exposure time (s)
Data and absorption corrections
Min./Max. transmission
LP and SADABS
0.315997/0.603229
19.38
6318
977
889
0.0490
64
µ (mm−1
)
Measured reflections
Unique reflections
Unique reflections, F0 > 4σ(F0)
Rint
Refined parameter
R1, wR2, GooF (all Refl.)
Max. shift/esd, last ref. cycle
Res. electron density: max., min.
0.0447, 0.0774, 1.193
<0.0005
2.14, −1.23 e− (Å3), 99,
110 pm from Ba
ꢁ
ꢂ1/2
ꢀ
ꢀ
ꢀ
ꢀ
R1 =
||F0| − |F0||/ |F0|; wR2 =
w(F02 − Fc2)2/ (wF20)2
;
w = 1/[σ2(F02) + (0.0082P)2 + 24.23P] with P = [(F02) + 2Fc2]/3.
4.3. The binary alloys of the MgZn2 type and their crystal
structure
At this stage we decided to establish whether AeMg2 com-
pounds really were pure intermetallics or if they also con-
tained nitrogen interstitial impurities. This possibility was
considered by comparison with Ag16Ca6N [28] which is
Table 4
Comparison of the crystallographic data for ‘Ba5[TaN4]’ [8] and Ba5[NbN4]N
Stoichiometry (X-ray data)
‘Ba5[TaN4]’
Ba5[NbN4]N
Uiso (pm2)
Space group (no.), Z
Lattice constants
a (pm)
b (pm), ␤ [◦]
c (pm)
C2/m (12), 4
C2/m (12), 4
1231.9(2)
1098.5(2), 113.64(1)
855.4(1)
1231.7(3)
1094.6(2), 113.65(3)
853.8(2)
M (Nb or Ta), 4i
Ba1, 4i
Ba2, 4i
Ba3, 4g
Ba4, 8j
N1, 4i
N2, 4i
N3, 8j
N4, 8j, SOF: 35(5)%
0.7536(1)/0/0.0335(2)
0.1100(2)/0/0.4059(3)
0.5509(2)/0/0.2182(3)
0/0.1782(2)/0
0.8219(2)/0.2723(2)/0.2763(2)
0.1865(36)/0/0.1517(56)
0.8751(41)/0/0.2622(64)
0.1479(20)/0.3629(19)/0.9991(32)
–
0.75313(11)/0/0.03353(16)
0.10874(9)/0/0.40506(13)
0.55173(7)/0/0.21960(12)
0/0.17841(9)/0
0.82238(6)/0.27315(7)/0.27828(8)
0.1888(12)/0/0.1496(17)
0.8790(13)/0/0.2618(19)
0.1486(8)/0.3615(9)/0.0011(12)
0.0919(29)/0.2447(30)/0.3471(37)
61(3)
204(3)
152(3)
124(3)
160(2)
178(32)
309(40)
157(22)
324(85)

O. Reckeweg et al. / Journal of Alloys and Compounds 384 (2004) 98–105
103
Table 6
Summary of the products from all reactions
Starting composition: ‘AeN’ or
Products for T = 600 ◦C
Products for T = 800 ◦C
Products for T = 1000 ◦C
Products for T = 1200 ◦C
ꢀ
‘AeMg2N2’,
m ≈ 500 mg
Mg:NaN3
Ca:NaN3
Sr:NaN3
Ba:NaN3
Yb:NaN3
Eu:NaN3
Mg, Mg3N2
Ca, ␤-Ca3N2
Sr2N
Ba2N
Yb
Mg3N2
Ca3N2
Sr2N
Ba2N
Yb, YbN
Eu, EuN
Mg3N2
Ca3N2, Ca2N
Sr2N
Ba2N
YbN, Yb
EuN
Mg3N2
Ca2N, Ca3N2
Sr2NbN3, Sr
Ba2NbN3, Ba
YbN
Eu
EuN
Ca/Mg/NaN3
Sr/Mg/NaN3
Ba/Mg/NaN3
Yb/Mg/NaN3
Eu/Mg/NaN3
Mg, Ca, ␤-Ca3N2
Sr2N, Mg, Mg3N2
BaMg2, Ba2N, Mg
Mg, Yb
CaMg2N2
SrMg2N2
Ba2N, Mg3N2, Mg
Mg3N2, Yb, YbN
Mg3N2, Eu, EuN
CaMg2N2
SrMg2N2
Mg3N2, BaMg2
Mg3N2, YbN
EuN, Mg3N2, EuMg2
CaMg2N2, CaMg2
SrMg2, SrMg2N2
BaMg2
YbMg2, Mg3N2, YbN
EuMg2, Mg3N2, EuN
Mg, Eu
Ca3N2/Mg3N2
Mg3N2, Ca3N2
Mg3N2, Sr2N
Mg3N2, Ba2N
Mg3N2, YbN
Mg3N2, EuN
Ca3N2, Mg3N2, CaMg2N2
SrMg2N2, Sr2N, Mg3N2
Ba2N, Mg3N2
YbN, Mg3N2
EuN, Mg3N2
CaMg2N2
CaMg2N2, ∗CaMg2
∗SrMg2, SrMg2N2
∗BaMg2, Mg3N2
Mg3N2, ∗YbMg2, YbN
Mg3N2, ∗EuMg2, EuN
Sr2N/Mg3N2/NaN3
Ba2N:Mg3N2:NaN3
YbN:Mg3N2:NaN3
EuN:Mg3N2:NaN3
SrMg2N2, SrMg2
Mg3N2, BaMg2
Mg3N2, YbN
EuN, Mg3N2, EuMg2
Products were identified by their respective powder X-ray diffractograms. Metallic sodium was observed in all reactions with NaN3. If not mentioned
oth∗erwise, the cubic ␣-Ca3N2 is referred to as ‘␤-Ca3N2’.
Single crystals of AeMg2 were picked from the reactions marked with an asterisk.
Table 7
Selected properties of Ae, AeMg2, Ae–Mg–N and Ae–Nb–N compounds
Element
Tmelt (◦C) [20]
Compound
Tmelt (◦C) [21]
Compound
Treact. (◦C)
Mg
Ca
Sr
Ba
649
839
769
725
CaMg2
SrMg2
BaMg2
YbMg2
EuMg2
714
680
607
718
719
Mg2NbN3
Ba5NbN5
Ba16Nb5N19
Ba2NbN3
BaMg3.33Nb0.67N4
1100 [22]
625
950 [23]
1050 [24]
1025 [1]
Yb
824
Eu
822
Compound
Treact (◦C)
Compound
Treact (◦C)
Compound
Sr5NbN5
Treact (◦C)
Mg3N2
Ca3N2
Ca2N
Sr2N
700–900
700–900
800–1000
550–700
550–700
␤-Ca3N2 [2]
CaMg2N2
CaMg2N2
SrMg2N2
SrMg2N2
300 ≤ T ≤ 650
725 [2]
1000 [25]
725 [2]
900 [26]
Sr2NbN3
1200 [27]
Ba2N
950
closely provided the radii for the rare earth elements in their
lower oxidation states, 2.1 and 2.0, respectively [34], are
used.
The most likely nitrogen site would be located inside the
Mg bipyramids (Wyckoff position 2b), a structural motif
which is also found in Mg3BN3 [30]. Table 8 shows that
the Mg–N bond lengths in the Mg bipyramid are plausi-
ble when compared to those found in Mg3BN3. The overall
largest residual electron density peaks occur for the Yb, Eu
and Ba compounds, but at unphysically close distances to
those metal atoms. These are likely due to truncation errors
of the intensity data or non-perfect absorption corrections.
Table 8
Properties of AeMg2 compounds according to our results
ꢀ
Compound
(Vmol
/
VMe) (%)
d(2b − 2a) (pm), 2x
d(2b − 6h), 3x
(2b) Residual electron density
(electrons/106 pm3)
YbMg2
EuMg2
CaMg2
SrMg2
BaMg2
97.1
96.7
95.0
92.1
93.2
253.0
258.1
253.6
261.4
265.3
184.2
185.2
182.6
187.6
187.7
1.07
1.20
0.21
<0.10
<0.10
Mg3[BN2]N
–
204.6
204.0
–
Selected data for Mg3BN3 [30] are given for comparison.

104
O. Reckeweg et al. / Journal of Alloys and Compounds 384 (2004) 98–105
Fig. 3. View down the b axis of Ba5[NbN4]N. [(N4)Ba6] octahedra are
displayed white hatched, [NbN4] tetrahedra are shown black hatched.
4.4. Results of the crystal structure refinement of the new
compound Ba5[NbN4]N
The few crystals of the new compound Ba5[NbN4]N were
all transparent orange and only of moderate crystallinity. De-
spite the relatively low reaction temperature employed and
the sluggishness of nitrogen to react under these conditions,
the Na–Ba–N mixture attacked the Nb container material
and incorporated the dissolved transition metal into a ternary
phase.
The crystal structure of Ba5[NbN4]N may be described
as one-dimensional chains of edge sharing [(N4)Ba6] octa-
hedra (white hatched). The chains run parallel to the a axis
(Fig. 3). The Ba–N bond distances are in the expected range
(259 pm < d < 340 pm, Table 9). The chains are farther in-
terconnected via corner sharing to form sheets. These sheets
are bound to each other by corner sharing the [NbN4] anions
that lie between these [(N4)Ba6] cationic layers. A projec-
tion of the unit cell of Ba5[NbN4]N along (0 1 0) is shown
in Fig. 3.
Fig. 1. View of the unit cell of AeMg2 (Ae = Ca, Yb, Sr, Eu and Ba)
perpendicular to the c axis. [Mg5] bipyramids are drawn shaded and
cross-hatched. Ae atoms (drawn as black circles) are connected with lines
to emphasize the connectivity pattern resembling the hexagonal diamond
structure.
For the most likely candidates SrMg2 and BaMg2 (which
deviate most from the ideal volume and have the largest Mg5
polyhedra) our refinement results do not exhibit significant
electron densities on (2b). According to our X-ray data re-
finements, the maximum possible nitrogen content for all
AeMg2 compounds studied would be SrMg2N0.07(5), that is,
the nitrogen content is zero within error (Table 2).
However, the refinement suggests a nitrogen defi-
ciency on the (8j) position resulting in the formula
(Ba2+)5[Nb5+(N3−)4](N3−
)0.7(1). This would imply that
the electric charges are not balanced and there is an excess
of approximately one electron per formula unit that is unac-
counted for. Since the obtained crystals are transparent and
since the refined formula Ba5[NbN4]N0.7(1) and the ideal,
SP
0.15
0.10
0.05
0.00
-0.05
Table 9
Average Ba–N (all d ≤ 350) and Nb–N (d ≤ 210 pm) bond length of
Ba–Nb–N compounds
¯
Compound
d (Nb–N)
Range of
Reference
(pm)
d(Ba–N) (pm)
Ba5[NbN4]N
Ba19[NbN4]2[Nb2N7]
Ba2[NbN3]
Ba9[NbN4]2N[N3]
Ba9[NbN4]2O[CN2]
194.5
196.0
194.2
197.0
196.7
259.3–314.6
259.1–337.8
279.2–335.4
261.0–336.6
261.0–336.9
–
1.1
1.2
1.3
1.4
RM
/RN
[23]
[24]
[35]
[36]
Fig. 2. Strain parameter diagram for selected Laves phases formed from
only electropositive metals (filled circles refer to phases investigated in
the present work). For details see text.

O. Reckeweg et al. / Journal of Alloys and Compounds 384 (2004) 98–105
105
ionic composition Ba5[NbN4]N are still within the range
of three standard deviations of each, it seems reasonable to
assume the ideal ionic composition and to believe that the
nitrogen deficiency is an artifact. There are several possible
explanations for the nitrogen deficiency occurring in the
refinement, e.g., the presence of heavy atoms and less than
perfect absorption corrections (Rint = 0.049) and/or of the
moderate crystallinity of the crystals or partial reduction
of Nb. Regarding the latter point, we find that the average
Nb–N bond distance in the [NbN4] tetrahedron is 194.6 pm
(Table 9), which is comparable to those found in Nb(V)
compounds (Table 9). Nb(IV) seems improbable since it has
never been observed in a tetrahedral nitrogen environment,
and one would expect the Nb–N bond length to be signif-
icantly larger in such a case. Unidentified impurities such
as H, B or C must as well be considered when compounds
with surprising compositions are obtained in low yield.
However, neither the yield nor the quality of the crystals ob-
tained was improved when metal powder, wire or foil were
used instead of employing the container wall as the niobium
source. Longer reaction times, higher reaction temperatures
and/or intentionally adding reagents such as C, BaH2 or
NaH into the starting mixture resulted in the formation of
BaCN2, BaC2, BaNH, Ba2NH, Ba19[NbN4]2[Nb2N7] or
Ba2[NbN3] as identified by powder X-ray diffraction in the
respective product mixture.
beyond any doubt that the composition of Ba5[NbN4]N is ex-
act; there is a small possibility that the N site connected only
to Ba has partial occupation. However, we argue based on the
transparency of the crystal that partial occupancy is unlikely.
References
[1] O. Reckeweg, J.C. Molstad, F.J. DiSalvo, J. Alloys Compd. 315
(2001) 134–142.
[2] O. Reckeweg, F.J. DiSalvo, Z. Anorg. Allg. Chem. 627 (2001) 371–
377, and references therein.
[3] O. Reckeweg, F.J. DiSalvo, Sol. State Sci. 5 (2002) 575–584, and
references therein.
[4] SAINT and SADABS: Bruker Analytical X-ray Instruments Software
Package, Version 5.03., G. M. Sheldrick, Göttingen, 1997.
[5] H. Witte, Naturwiss. 25 (1937) 795–800.
[6] E. Hellmer, A. Laves, Z. Kristallogr. A 105 (1943) 134–143.
[7] P.I. Krypyakevich, V.I. Evdakimenko, Dopov. Akad. Nauk. Ukr. SSR,
1964, pp. 766–769.
[8] H.-T. Künzel, Doctoral Thesis, Univ. Stuttgart, Germany, 1980.
[9] W. Klemm, G. Winkelmann, Z. Anorg. Allg. Chem. 288 (1956) 87–
90.
[10] E.T. Keve, A.C. Skapski, Inorg. Chem. 7 (1968) 1757–1761.
[11] D.H. Gregory, A. Bowman, C.F. Baker, D.P. Weston, J. Mater. Chem.
10 (2000) 1635–1641.
[12] C.F. Baker, M.G. Barker, A.J. Blake, Acta Crystallogr. E 57 (2001)
i6–i7.
[13] J. Gaude, P. L’Haridon, Y. Laurent, J. Lang, Bull. Soc. Fr. Min.
Crystallogr. 95 (1972) 56–60.
Künzel reported that the compound ‘Ba5[TaN4]’ formed
black crystals, obtained by heating ‘Ba3N2’ to 1400 K for
1 h in Ta ampoules [8]. A comparison with the crystallo-
graphic data reported for ‘Ba5[TaN4]’ and those found by
us for Ba5[NbN4]N (Table 4) suggests that these two com-
pounds would be isotypic if additional nitrogen on the po-
sition (8j) in the Ta compound was overlooked. However,
the reported black color of ‘Ba5[TaN4]’ suggests partial (or
zero) occupancy of the (8j) site in the Ta compound.
[14] N.E. Brese, M. O’Keeffe, J. Sol. State Chem. 87 (1990) 134–140.
[15] O. Seeger, Doctoral Thesis, Univ. Tübingen, Germany, 1994.
[16] V. Schultz-Coulon, Doctoral Thesis, Univ. Bayreuth, Germany, 1998.
[17] G. Auffermann, Y. Prots, R. Kniep, Angew. Chem. Int. Ed. 40 (2001)
547–549.
[18] Y. Prots, G. Auffermann, M. Tovar, R. Kniep, Angew. Chem. Int.
Ed. 41 (2002) 2288–2290.
[19] G.V. Vajenine, G. Auffermann, Y. Prots, W. Schnelle, R.K. Kremer,
A. Simon, R. Kniep, Inorg. Chem. 40 (2001) 4866–4870.
[20] H.W. King, in: D.R. Lide (Ed.), CRC Handbook of Chemistry and
Physics, 74th ed., Chemical Rubber Publ. Co., 1993–1994, Chap. 12,
pp. 10–12.
[21] M. Hansen (Ed.), Constitutions of Binary Alloys, Genium Publ.
Corporation in Arrangement with McGraw-Hill, 1985.
[22] C. Lind, F.J. DiSalvo, unpublished results.
5. Conclusions
[23] P. Höhn, R. Kniep, Z. Naturforsch. 56b (2002) 604–610.
[24] O. Seeger, J. Strähle, J.P. Laval, B. Frit, Z. Anorg. Allg. Chem. 620
(1994) 2008–2013.
[25] V. Schultz-Coulon, W. Schnick, Z. Naturforsch. 50b (1995) 619–622.
[26] P. Höhn, R. Kniep, Z. Anorg. Allg. Chem. 628 (2002) 463–467.
[27] Ch. Wachsmann, Doctoral Thesis, Univ. Dortmund, 1995;
X.Z. Chen, H.A. Eick, W. Lasocha, J. Sol. State Chem. 138 (1998)
297–301.
[28] J. Snyder, A. Simon, Angew. Chem. Int. Ed. Engl. 33 (1994) 689–
690.
[29] L.D. Calvert, C. Rand, Acta Crystallogr. 17 (1964) 1175–1176.
[30] H. Hiraguchi, H. Hashizume, O. Fukunaga, A. Takenaka, M. Sakata,
J. Appl. Cryst. 24 (1991) 286–292.
[31] A. Simon, Angew. Chem. Int. Ed. Engl. 22 (1983) 95–113.
[32] C.D. Churcher, V. Heine, Acta Crystallogr. A 40 (1984) 291–296.
[33] G.J. Snyder, A. Simon, Z. Naturforsch. 49b (1994) 189–192.
[34] W.B. Pearson, The Crystal Chemistry and Physics of Metals and
Alloys, Wiley-Interscience, New York, 1972, p. 151.
[35] S.J. Clarke, F.J. DiSalvo, Z. Kristallogr. 212 (1997) 309–310.
[36] O. Reckeweg, F.J. DiSalvo, Z. Naturforsch. 58b (2003) 201–204.
A series of reactions with alkaline earth metals and ni-
trogen in Nb containers at different temperatures were per-
formed. Known binary nitrides and subnitrides, the ternary
nitrides AeMg2N2 (Ae = Ca, Sr) and ‘␤-Ca3N2’ were ob-
tained, of which only the latter is not fully characterized.
Under some conditions Nb and Ta containers are relatively
inert. At higher temperatures and especially with a limited
nitrogen content, the Nb walls start to participate in the re-
actions first by leaching out a little Nb into the reaction mix-
ture, and then at higher temperatures by absorbing most or
all of the available nitrogen leaving behind the binary alloys
AeMg2. Single crystal structure analyses do not indicate
the presence of nitrogen on (2b) and confirm earlier results.
Ba[Mg3.33M0.67]N4 (M = Nb, Ta), ‘Ba5[TaN4]’ and, as
reported here, Ba5[NbN4]N were obtained first by serendip-
ity as a by-product of reactions of alkaline earth metals with
nitrogen in Nb or Ta containers. We are not yet able to prove