Single crystal studies on boron-rich τ-borides Ni23-xM xB6 (M=Zn, Ga, In, Sn, Ir)The surprising occurrence of B4-tetraheda as a normal case?

Article abstract of DOI:10.1016/j.jssc.2010.07.048

Single crystals of the cubic τ-borides Ni_23-xM _xB₆ (M=Zn, Ga, In, Sn, Ir) were synthesised from the elements at temperatures between 1200 and 1500 °C. The structure refinements show that the existence of boron-rich phases is quite common. Starting from the idealised composition Ni₂₀M′₃B₆ a part of the metal atoms on site 8c is substituted by B₄ tetrahedra. For M′=Ga a complete exchange seems to be possible leading to the composition Ni₂₀GaB₁₄. For M′=Zn and Sn the formation of solid solutions is less extended. For M′=In no exchange is observed but an unusual pattern of Ni/In distribution is observed. With M=Ir mixed occupations occur for all sites and the boron content varies, too. All compositions were confirmed by EDX measurements.

Full text of DOI:10.1016/j.jssc.2010.07.048

Journal of Solid State Chemistry 183 (2010) 2281–2289
Contents lists available at ScienceDirect
Journal of Solid State Chemistry
journal homepage: www.elsevier.com/locate/jssc
Single crystal studies on boron-rich
In, Sn, Ir)—The surprising occurrence of B -tetraheda as a normal case?
t
-borides Ni2
M B (M¼Zn, Ga,
3ꢀx
x 6
4
a
a
a,b,n
Dominik Kotzott , Martin Ade , Harald Hillebrecht
a
Albert-Ludwigs-Universit a¨ t Freiburg, Institut f u¨ r Anorganische und Analytische Chemie, Albertstr. 21, D-79104 Freiburg, Germany
Freiburger Materialforschungszentrum FMF, Stefan-Maier-Str. 25, D-79104 Freiburg, Germany
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Single crystals of the cubic
t
-borides Ni2
3ꢀx
MxB6 (M¼Zn, Ga, In, Sn, Ir) were synthesised from the
Received 23 May 2010
Received in revised form
elements at temperatures between 1200 and 1500 1C. The structure refinements show that the
existence of boron-rich phases is quite common. Starting from the idealised composition Ni20
0
M
3
B
6
a
2
8 July 2010
0
part of the metal atoms on site 8c is substituted by B
4
tetrahedra. For M ¼Ga a complete exchange
Accepted 31 July 2010
Available online 20 August 2010
0
seems to be possible leading to the composition Ni20GaB14. For M ¼Zn and Sn the formation of solid
0
solutions is less extended. For M ¼In no exchange is observed but an unusual pattern of Ni/In
Keywords:
distribution is observed. With M¼Ir mixed occupations occur for all sites and the boron content varies,
Nickel borides
too. All compositions were confirmed by EDX measurements.
t
B
-borides
&
2010 Elsevier Inc. All rights reserved.
4
-tetrahedra
Single crystal growth
Structure analysis
XRD
EDX
1
. Introduction
Most of the investigations were done on powder samples for
several reasons. Firstly -borides were usually investigated in the
t
Metal rich borides of 3d transition metals with a M:B ratio 42
play an important role in hardening processes of alloys. They are
formed on the surface by boridising (for example with gaseous
course of equilibrium studies. For this the samples were heated up
for a short time near the assumed melting point and then annealed
at temperatures between 800 and 1000 1C [3,5]. The resulting
products were normally characterised only by metallographic
methods and X-ray powder diffraction. Secondly the formation of
single crystals is difficult because of kinetic reasons as melting
points and annealing temperatures were very different. Therefore
only a few detailed investigations of the formation of solid solutions
BCl
The so-called
numerous representatives of metal rich borides. Till date more than
0 representatives are known [1]. Its cubic crystal structure
Fm 3¯ m) belongs to the Cr23
type [2], which contains 4 different
metal sites (48h, 32f, 4a, 8c, see later). As pointed out by
Stadelmeyer [3] -borides are typically ternary borides with a
d-metal (Cr to Ni) on the sites 48h and 32f while the sites 4a and
c are occupied by rare earth metals, 4d and 5d transition metals or
-borides were assumed to
exist only as metastable phases (first well-characterised example
[4]) the minor component was called ‘‘ -stabiliser’’. It was
3
) or between the grain boundaries of the alloys by segregation.
t
-borides with a composition M23 are the most
B
6
8
(
C
6
in
t-borides on the basis of single crystal data are reported [1].
t
A new aspect started up, when it was shown for the -borides
t
3
8
of the system Ni/Al/B [6] that the enlargement of the boron
content is caused by the substitution of aluminium on site 8c
a main group metal. Until 2009 binary
t
4 4
by B -tetrahedra. It was the first example for a B -tetrahedron
in a typical solid state compound. Besides this there is also a
mixed occupation of site 4a by Al and Ni. Therefore the formation
of solid solutions is described by two different mechanisms:
Co23B
6
t
known that mixed occupation occurs for the different metal sites
according to the different size [3–5] but usually a full occupation of
the sites 4a and/or 8c is assumed to reach the ideal compositions
Ni20(Al1–xNi
x
)(Al1–y
B
)
4y 2
B
6
. (x, y between 0 and 1). Representa-
˚
tives are Ni20.5Al2.5
B
6
(a¼10.486 A, x¼0.5; y¼0), Ni20Al
3 6
B
0
0
0
0
˚
˚
M
22M B
6
, M21M
B
2 6
and M20
M
B
3 6
with M as
t-stabiliser.
(a¼10.511 A, x¼y¼0) and Ni20AlB14 (a¼10.617 A, x¼0, y¼1).
The enlargement of the lattice parameter with rising B-content is
4
explained by the higher volume of a B -tetrahedron in compar-
ison to a single Al atom. Besides the use of excess Al played a
central role for the formation and crystallisation of the boron-rich
n
Corresponding author at: Albert-Ludwigs-Universit a¨ t Freiburg, Institut f u¨ r
t
-boride.
Anorganische und Analytische Chemie, Albertstr. 21, D-79104 Freiburg, Germany.
Fax: +49 761 203 6012.
In continuation of this work we have investigated other
0
E-mail address: harald.hillebrecht@ac.uni-freiburg.de (H. Hillebrecht).
t-borides of nickel, i.e. M ¼Zn, Ga, In and Sn concerning the
0022-4596/$ - see front matter & 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.jssc.2010.07.048

2
282
D. Kotzott et al. / Journal of Solid State Chemistry 183 (2010) 2281–2289
formation of boron-rich phases. The special role of Ir-containing
-phases was known from previous work [4,5,7] and therefore the
system Ni/Ir/B was studied, too. Because the differences of the
vary between 5 and 15 at%. Remarkable is the variation of the
boron content between 20 and 35 at%. With increase in the
t
˚
B-content the lattice parameter rises from a¼10.50 to 10.60 A.
diffraction patterns are very small, the use of single crystal data is
essential for unambiguous results. It turned out that in Ni-
2
.2.2. Ni/Zn/B
0
containing
are able to stabilise boron-rich
Furthermore, it was shown that iridium can substitute larger
amounts (i.e. x43) of the 3d-metal (Cr7.9Ir14.1
[5]) and additionally, forms a boron-rich
t
-borides two other main group metals (M ¼Ga, Sn)
Because of the low boiling point of Zn the samples were welded in
t
4
-phases with B -tetrahedra [8,9].
Ta-ampoules. Single phase
t
-borides were obtained for boron-poor
˚
compositions (a¼10.52–10.57 A). With higher boron content the
6
B [7], Fe10.0–15.4
amount of binary Ni-borides increased. Then the
t-phase was only a
Ir13.0–7.6
B
6
t
-boride
by-product and the lattice parameter could not be refined.
(Ni12.3Ir8.7B10.4).
2.2.3. Ni/Sn/B
t
-phases as main products were obtained from boron-poor
2
. Experimental
samples. A nearly single phase sample with small amounts of Ni
˚
resulted from a composition Ni21Sn
2
B
6
(a¼10.585–10.587 A).
2.1. Syntheses and characterisation
With increase in the boron-content binary Ni-borides were found.
The use of an arc melter revealed similar results.
The elements were mixed in ratios, which covers different
0
amount of M and boron. They were pressed into pellets (mass:
2.2.4. Ni/In/B
ꢁ
300 mg) and put into a corundum crucible. Under an atmo-
The formation of a
samples. Single phase samples were obtained for compositions
between Ni22InB6 and Ni21In B . Obviously boron-rich -phases
t-boride was only observed for boron-poor
sphere of argon the pellets were heated up to a maximum
temperature (1200–1450 1C) followed by a dwell time of several
hours (1–10 h) and cooling down (rate: 10–15 1C/h) to a medium
temperature (600–1000 1C). Then the furnace was turned off.
Further investigations were done with the crushed metallic reguli.
A few syntheses in the system Ni/Sn/B were done with an arc
melter. Table 1 gives an overview on the synthesis of some typical
2
6
t
do not exist.
2.2.5. Ni/Ir/B
t
-phases were only observed for boron-rich samples. Boron-
poor batches contained the binary borides of Ir and Ni. The
variation of the lattice parameter is quite small indicating similar
Ni/Ir-ratios.
samples. Because
t-borides were in the focus of interest
compositions are usually given in a way that the amounts of Ni
and M in the stoichiometric formula sum up to 23 and boron is
between 6 (boron-poor) and 15 (boron-rich).
2.3. Single crystal investigations and refinements
Powdered samples were characterised by XRD (STOE, Darmstadt,
Germany, STADI-P, transmission geometry; MoK
a-radiation and
Single crystals were isolated from the crushed melt and
investigated with a single crystal diffractometer. The indexing
routine revealed a face-centred cubic unit cell in each case. Lattice
linear position sensitive detector (PSD) or CuK -radiation and image
a
plate (IP) detector). Silver metallic single crystals were collected from
the crushed ingots as irregular polyhedral. Most of the single crystal
investigations and data collections were performed by using a single
crystal diffractometer equipped with an image plate detector (STOE,
˚
parameters (10.46–10.95 A) and reflection conditions were char-
acteristic for -borides. The data sets were corrected for absorption
and merged in Laue-class m 3¯ m. The subsequent refinements
Program SHELXL [15]) were started with the known structure
model of a
-boride (cubic, Z¼4, space group Fm 3¯ m—No. 225).
t
Darmstadt, Germany, IPDS II, MoK
mator). Measurements for the Sn-containing crystals were done with
a conventional 4-circle diffractometer (Nonius, CAD4, MoK -radia-
a-radiation, graphite monochro-
(
t
a
Because mixed occupations were expected all site occupation factors
were checked. Mixed occupations were considered, if the deviations
were greater than triple of the (usually very small) standard
deviations. For all systems several single crystals were measured
leading to very similar results. The results of the most reliable
refinements (R-values, standard deviations, residual electron den-
sities) were used for the discussion. Further details are listed in
Tables 2–4, or may be obtained from Fachinformationszentrum
Karlsruhe, D-76344 Eggenstein-Leopoldshafen (Germany) (Fax:
tion, graphite monochromator). All measurements were done at room
temperature. A numerical absorption correction was applied. For the
IPDS data we used the equivalent method (Program XSHAPE [10]), for
the CAD4 data the Psi-scan method.
The metal content was determined by the energy dispersive
X-ray spectroscopy EDX (Electron microscope: Zeiss DSM-960;
EDX-system: Link/LEO 1525). This was necessary for M¼Ga and
Zn because of the similar electron numbers. For M¼Ir it was done
because of the dissipation on all sites. Measurements were usually
made of the crystals which were already used for the structure
determinations (see later). Comparing compounds with known
composition the errors can be estimated to be 71–2 at%,
depending on the quality of the surface and the elements
+
49)724-808-666; e-mail: crysdata@fiz-karlsruhe.de) on quoting
the registry numbers CSD-421820 (Ni20Ga ), CSD-421819 (Ni20.0
10.1), CSD-421823 (Ni21.4Zn1.6 ), CSD-421821 (Ni20.0Zn2.4
7.75), CSD-421825 (Ni22.2In0.8 ), CSD-421826 (Ni22.0In1.0 ), CSD-
21824 (Ni21.6In1.4 ), CSD-421817 (Ni12.3Ir8.7 10.4), CSD-421818
Ni20.9Sn1.9 7.3), CSD-421822 (Ni21.1Sn1.9
Because of the fundamental importance of the refinements
3 6
B
Ga1.6B
B
6
B
4
(
B
6
B
6
B
6
B
involved. In contrast to several other cases (for example Li
11], MgB12 [12], MgB12Si [13]) the unfavourable crystal shape
prevented to obtain reliable data for the boron content.
2 12 2
B C
B
6
B )
[
C
2
2
they will be discussed in more detail.
2
2
.2. Syntheses and experimental results
.2.1. Ni/Ga/B
2.3.1. Ni20Ga3B6
The refinement was done with Ni on the sites of M1 and M2
and Ga on sites M3 and M4 resulting in excellent R-values of
The formation of
compositions. The access to single phase samples is possible. By-
products are the binary Ni-borides, Ni–Ga alloys and the recently
t
-phases occurs in a quite broad field of
R
1
¼0.0171 and wR ¼0.0166. Each occupation factor (sof) was
2
separately checked. The deviations from a complete and ordered
occupation were very small. Furthermore, the displacement
parameters are very similar indicating a correct assignment.
8
described ternary compound Ni12GaB [14]. The Ga-content can

D. Kotzott et al. / Journal of Solid State Chemistry 183 (2010) 2281–2289
2283
Table 1
Typical experimental conditions and results of XRD characterisation of
identified; –, not observed.
t-borides Ni/M/B (M¼Ga, Zn, In, Sn, Ir), main part in bold; 4, by-products; c, traces; ?, not
System/sample
Maximum temperature/
holding time
Cooling rate
Lower holding
temperature
t
-phase,
t
-phase single
By-products
composition
powder
crystal
Ni/Ga/B
Ni20Ga
Ni20Ga
Ni20Ga
3
3
1
B
B
B
6
1450 1C/10 h
1450 1C/10 h
1450 1C/10 h
1450 1C/10 h
1450 1C/10 h
1450 1C/10 h
1450 1C/10 h
1450 1C/10 h
10 1C/h
10 1C/h
10 1C/h
10 1C/h
10 1C/h
10 1C/h
10 1C/h
10 1C/h
1000 1C
1000 1C
1000 1C
1000 1C
1000 1C
1000 1C
800 1C
10.5090(9)
10.5886(11)
10.5886(8)
10.5917(7)
–
10.4911(16)
10.5434(18)
10.5528(13)
3
c Ni Ga
–
12
12
Ni
B, Ni12GaB
2 8
Ni20Ga0.5
B
12
Ni
2
B
Ni20Ga
Ni20Ga
Ni20Ga
6
2
2
B
12
B
15
B
12
2
NiGa, Ni B
10.594(3)
10.583(3)
10.5912(11)
Ni
2
Ni
2
Ni
2
B
B
Ni20GaB12
800 1C
B, NiGa
Ni/Zn/B
Ni18Zn
Ni19Zn
Ni21Zn
Ni20Zn
Ni20Zn
Ni20Zn
5
4
2
3
3
3
B
B
B
B
B
B
6
1450 1C/1 h
1450 1C/1 h
1450 1C/1 h
1450 1C/1 h
1450 1C/1 h
1450 1C/1 h
20 1C/h
20 1C/h
20 1C/h
20 1C/h
20 1C/h
20 1C/h
1000 1C
1000 1C
1000 1C
1000 1C
1000 1C
1000 1C
10.5716(5)
10.5642(2)
10.5221(9)
10.54
–
6
–
6
10.5409(17)
10.5521(17)
Ni
Ni
Ni
3
3
2
B
B
B
8
10
14
10.54
–
4 3 3
o-Ni B , NiB, Ni B
Ni/Sn/B
Ni28Sn
Ni25Sn
Ni21Sn
Ni20Sn
Ni20Sn
Ni21Sn
Ni20Sn
2
2
2
3
3
2
3
B
B
B
B
B
B
B
6
1450 1C/1 h
1450 1C/1 h
1450 1C/1 h
1450 1C/1 h
1450 1C/1 h
Arc melting
Arc melting
Arc melting
Arc melting
Arc melting
300 1C/h
300 1C/h
300 1C/h
20 1C/h
800 1C
800 1C
800 1C
1000 1C
1000 1C
10.5853(3)
10.5871(4)
10.6056(8)
10.6129(6)
10.6098(7)
10.5918(11)
10.6045(2)
10.6065(13)
10.5917(4)
–
10.5827(3)
10.5958(3)
Ni
Ni
6
8
Ni
2
B, t-Ni
Sn
B, Ni
3 2
Sn
8
t-Ni
3
2
10
6
20 1C/h
Ni
–
2
3
Sn
2
6
3 2
h-Ni Sn
Ni21SnB10
Ni
Ni
Ni
2
2
2
B
Ni21Sn0.5
Ni20Sn1.5
B
12
B, Ni
3
B
B
12
3 2
B, h-Ni Sn
Ni/In/B
Ni19In
Ni20In
Ni21In
Ni22In
4
3
2
1
B
B
B
B
6
1350 1C/1 h
1350 1C/1 h
1350 1C/1 h
1350 1C/1 h
1350 1C/1 h
1350 1C/1 h
1350 1C/1 h
20 1C/h
20 1C/h
20 1C/h
20 1C/h
20 1C/h
20 1C/h
20 1C/h
900 1C
900 1C
900 1C
900 1C
900 1C
900 1C
900 1C
10.595(2)
10.591(2)
10.578(1)
10.550(4)
10.590(2)
10.585(3)
–
10.5906(25)
3 2
h-In Ni
6
6
6
3 2
4 h-In Ni
–
–
10.5412(18)
10.5951(15)
Ni17.25In5.75
Ni19In
B
6
3 2
h-In Ni2, Ni B
4
B
8
h-In
h-In
3
Ni
Ni
2
, Ni
2
B
Ni18.5In4.5
B
10
3
2
, Ni
2
B, In
Ni/Ir/B
Ni21Ir
Ni17Ir
2
B
B
6
1450 1C/1 h
1450 1C/1 h
1450 1C/1 h
1450 1C/1 h
1450 1C/1 h
1450 1C/1 h
1450 1C/1 h
10 1C/h
10 1C/h
10 1C/h
10 1C/h
10 1C/h
10 1C/h
10 1C/h
1000 1C
1000 1C
1000 1C
1000 1C
1000 1C
1000 1C
1000 1C
–
Ni
Ni
3
B, Ni3ꢀxIr
x
B
6
6
–
3
B, Ni B, IrB1.1
2
Ni13Ir10
B
6
–
x
Ni3ꢀxIr B
Ir
Ni14.5Ir8.5
B
B
15
15
10.9366(13)
10.980(2)
10.982(3)
10.976(5)
Ni13.5Ir9.5
10.9393(13)
NiB, Ni
o-Ni
IrB1.1
2
B, IrB1.1
, IrB1.1
Ni11.5Ir11.5
Ni9.5Ir13.5
B
15
4
B
3
B
15
The final difference Fourier synthesis shows no significant
residual electron density. The electron numbers of Ni and Ga
have a difference of 10%. Accordingly the exchange of Ga by Ni
enlarges the sof-values to 110% (M3) and 113% (M4). A substitu-
tion of Ni1 and Ni2 by Ga is very unlikely because it demands the
formation of short Ga–B bonds. Furthermore, it was not observed
in the system Ni/Al/B.
These findings were confirmed for several single crystals from
˚
boron-rich batches with lattice parameters around 10.55 A.
Unfortunately, we were not able to obtain single crystals with
˚
the largest lattice parameters of 10.60 A as it was refined from the
powder XRD data. Similar to the -borides in Ni/Al/B we expect a
complete substitution of Ga by B -tetrahedra and a composition
Ni20GaB14
t
4
.
2
.3.2. Ni20Ga1.6
With a structure model of a ‘‘regular’’
R-values are R
B
10.1
2.3.3. Ni21.4Zn1.6
According to the EDX results the crystal had a Ni:Zn ratio of
93:7. The refinement of the regular structure of a -boride with Zn
on site 4a (i.e. Ni22ZnB ) resulted in R-values of R
¼0.037 and
wR
B
6
t
3 6
-boride Ni20Ga B the
1
¼0.13 and wR
2
¼0.26. The displacement para-
t
meter of Ga on site 8c is enlarged by a factor 4. With a free site
occupation factor the sof-value is reduced to 27(1)% and R-values
of R
electron density of 3.5 e /A on a site 32f x,x,x with x¼0.194 as it
was known for boron-rich -borides [4,6]. With partial
6
1
2
¼0.091. The displacement parameter of M3 was slightly
1
¼0.031 and wR
2
¼0.056%. After this there remains a residual
enlarged (ꢁ40%) while that of M4 (i.e. Zn) was comparable to Ni1
and Ni2. With a free sof of M3 refined as M¼Ni the occupation is
ꢀ
3
˚
t
a
reduced to 92(2)% and the R-values decrease to R
1
¼0.036 and
occupation of this site by boron an occupation of 51(5)% is
obtained and the R-values are again significantly reduced
wR
2
¼0.089. A share of 7% Zn (EDX) gives the composition
Ni21.4Zn1.6B . Therefore additional Zn is expected which is
6
(
R
1
¼0.0185, wR
2
¼0.0302). The composition determined by EDX
).
probably located on M3 (ca. 25%). On the other hand this was
ignored in the refinement. A clear separation between M-deficit
was in excellent agreement (7.5 at% Ga: Ni21.25Ga1.75
B
x

Table 2
0
0
Data for crystal structure analyses of
t
-borides Ni/M /B (M ¼Zn, Ga, In, Sn, Ir), esd’s in parentheses.
Compound
Ni20Ga
3
B
6
Ni20Ga1.6 Ni20Zn2.4
B
10.1
Ni21.4Zn1.6
B
6
B
7.75
Ni22.2In0.8
B
6
Ni22.0In1.0
Irregular
fragment
B
6
Ni21.6In1.4
B
6
Ni12.3Ir8.7
B
10.4
Ni20.9Sn1.8
B
7.25
Ni21.1Sn1.9
Irregular
fragment
B
6
Crystal shape
Irregular
fragment
Irregular fragment Irregular fragment Irregular fragment Irregular
fragment
Irregular fragment Irregular fragment Irregular
fragment
3
Size in (mm )
0.01 ꢂ 0.02 ꢂ 0.02 0.02 ꢂ 0.03 ꢂ 0.02 0.02 ꢂ 0.02 ꢂ 0.01 0.04 ꢂ 0.04 ꢂ 0.04 0.01 ꢂ 0.01 ꢂ 0.01 0.02 ꢂ 0.02 ꢂ 0.02 0.03 ꢂ 0.03 ꢂ 0.03 0.01 ꢂ 0.01 ꢂ 0.01 0.02 ꢂ 0.02 ꢂ 0.03 0.02 ꢂ 0.03 ꢂ 0.02
Unit cell, a ( A˚ )
Volume, V [ A˚ 3
3
calc [g/cm ]
1
0.4911(16)
10.5434(18)
1172.0(3)
7.879
10.5409(17)
1171.4(3)
8.086
10.5521(17)
1174.9(3)
7.987
10.5951(15)
1189.4(3)
8.142
10.5906(25)
1187.8(3)
8.227
10.5412(18)
1171.3(3)
8.463
10.9323(13)
1306.6(3)
12.777
10.5958(3)
1189.6
10.5827(3)
1185.2
1154.7(2)
8.331
]
d
8.478
8.584
0
1r
o
r1801
¼01,111;
¼21
01r
o
r1801
¼01, 111;
¼21
01r
o
r1801
¼01, 111;
¼21
01r
o
r1801
¼01, 111;
¼21
01r
o
¼01;
r1801
¼21
01r
o
¼01,
r901
¼21
01r
o
r1801
¼01, 111;
¼21
01r
o
r1801
¼01, 111;
o¼21
O
/2
Y
-scan
O/2Y-scan
0.70170.40 tan Y
c
D
300
c
c
D
c
c
D
o
c
D
o
c
c
0.70170.40 tan
Y
o
Do
o
Do
D
o
D
Exposure time (s)
Theta range
300
300
180
900
6.71o2
–16oho16
–16oko16
–16olo16
35.64
180
o67.61 6.71o2
180
300
6.71o2
–16oho16
–16oko16
–16olo16
105.7
90
90
6.71o2
y
o651
31o2
y
o69.81
31o2
y
o651
6.71o2
y
o65.81
y
y
o651
6.71o2
y
o651
y
o64.81
41o2
y
o801
41o2yo801
–
–
–
15oho15
–16oho16
–16oko16
–16olo16
34.67
–14oho14
–14oko14
–14olo14
36.62
–16oho16
–16oko16
–16olo16
35.86
–16oho16
–16oko16
–16olo16
35.77
–16oho16
–16oko16
–16olo16
36.41
–19oho0
–19oko0
–19olo0
36.39
–19oho0
–19oko0
–19olo0
36.43
15oko15
15olo15
–1
m
(mm
)
38.54
R
int/Rsigma
0.053/0.068
0.282/0.524
0.117/0.109
0.200/0.395
6613, 160
133
0.130/0.054
0.191/0.398
5622, 118
97
0.0685/0.046
0.015/0.056
6046, 135
94
0.129/0.054
0.017/0.061
5595, 151
114
0.0774/0.0273
0.027/0.078
1274, 134
102
0.094/0.048
0.017/0.064
6412, 135
86
0.174/0.077
0.0045/0.048
6389, 144
108
0.0856/0.0498
0.115/0.213
1040, 223
193
0.0615/0.0371
0.056/0.084
1040, 224
182
T
min/Tmax
N (hkl) meas.; unique 6576, 137
0
N (hkl) (I42
s(I))
87
Param. refined
14
18
15
16
15
15
14
16
16
16
2
R
R
1
(F)/wR
2
(F )
0.0171/0.0166
0.0380
0.0182/0.0297
0.0247
0.0357/0.0889
0.0411
0.0119/0.0125
0.0249
0.0489/0.0770
0.066
0.0201/0.0295
0.0300
0.0283/0.0401
0.0466
0.0206/0.0224
0.0366
0.0193/0.0365
0.0273
0.0165/0.0294
0.0281
1
(F), all data
Weighting scheme
15]
0.0/0.0
0.0146/0.0
0.0618/3.19
0.0/0.0
0.021/630.0
0.0088/0.0
0.013/0.0
0.0/0.0
0.0/0.0
0.0046/0.0
[
Extinction [15]
Goodness of fit
0.00021(1)
0.446
0.0018(1)
0.449
0.0005(2)
1.056
0.00024(1)
0.559
0.00025(7)
1.200
0.00128(6)
1.066
0.00000(3)
0.852
0.00076(3)
0.588
0.0019(1)
0.978
0.0014(1)
1.022
Residual electron
+0.75/–0.70
0.16
+0.86/–1.06
0.19
+0.95/–0.99
0.28
+0.36/–0.36
0.10
+1.38/–3.25
0.36
+1.11/–0.77
0.18
+1.05/–1.52
0.26
+2.36/–2.11
0.33
+1.06/–1.03
0.25
+0.90/–1.11
0.23
ꢀ
˚
3
density in [e /A ]
(max, min, s)

D. Kotzott et al. / Journal of Solid State Chemistry 183 (2010) 2281–2289
2285
Table 3
0
0
Coordinates, displacements parameters and occupation factors of
t
-borides Ni/M /B (M ¼Zn, Ga, In, Sn, Ir).
Compound
3
Ni20Ga B
6
Ni20.9Ga1.6B10.1 Ni21.4Zn1.6
B
6
Ni20Zn2.4
B
7.75 Ni22.2In0.8
B
6
Ni22.0In1.0
B
6
Ni21.6In1.4
B
6
Ni12.3Ir8.7B10.4 Ni20.9Sn1.8B7.25 Ni21.1Sn1.9B
6
Ni1, site 32f (x,x,x)
x
0.38228(4) 0.38027(4)
0.0071(2) 0.0095(2)
0.38287(11)
0.0168(6)
0.38242(3) 0.38418(14)
0.38467(6) 0.38486(9)
0.37617(4)
0.38457(2)
0.0066(1)
0.38503(2)
U
eq.
0.0151(1)
0.0206(6)
0.0112(3)
0.0161(4)
0.0048(2)
0.0069(1)
a
a
a
a
a
a
a
b
a
a
sof Ni, ratio Ni/M
98.4(5)
100.6(5)
98.7(13)
99.7(6)
100.0(15)
101.0(7)
98.4(11) 15.8(5)/84.2
100.2(4)
100.0(3)
U
U
11¼U22¼U33
12¼U13¼U23
0.0095(2) 0.0095(2)
0.0008(2) 0.0017(2)
0.0168(6)
0.0012(4)
0.0151(1)
0.0022(1)
0.0206(6)
0.0067(6)
0.0117(3)
0.0044(3)
0.0161(4)
0.0015(4)
0.0047(2)
0.0008(1)
0.0060(1)
0.0023(1)
0.0069(1)
0.0007(1)
Ni2, site 48h (0,y,y)
y
0.16923(4) 0.16924(4)
0.0072(1) 0.0093(1)
0.16979(9)
0.0167(5)
0.16994(3) 0.16958(11)
0.16923(5) 0.16994(3)
0.16539(7)
0.16895(2)
0.0050(2)
0.16873(2)
0.0069(1)
99.5(4)
U
eq
0.0134(1)
0.0156(5)
0.0076(2)
0.0149(3)
0.0047(4)
a
a
a
a
a
a
a
b
a
sof Ni, ratio Ni/M
101.6(4)
100.1(6)
101.1(13)
100.3(4)
100.3(15)
99.2(7)
101.8(11) 84.0(8)/18.0
100.3(2)
U
U
U
11
0.0087(3) 0.0102(3)
0.0065(2) 0.0089(2)
0.0007(2) 0.0003(2)
0.0168(8)
0.0166(6)
0.0007(5)
0.0146(2)
0.0127(1)
0.0009(1)
0.0172(8)
0.0148(5)
0.0014(6)
0.0088(3)
0.0070(4)
0.0140(4)
0.0165(4)
0.0047(4)
0.0049(5)
0.0000(2)
0.0045(1)
0.0060(1)
0.0004(1)
0.0062(1)
0.0084(1)
0.0003(1)
22¼U33
23
0.0004(2) ꢀ0.0007(4)
M3, site 8c (1/4,1/4,1/4)
0.0123(3) 0.0100(15)
U
iso
0.0156(10)
0.0166(5)
0.0142(14)
0.0072(6)
0.0174(8) 0.0055(29)#
0.0065(1)#
0.0082(1)#
7.0(7)/93.0
a
a
b
b
sof M, ratio Ni/M
97.0(8)
26.8(8)
92.3(13) (Ni) 68.5(6) (Zn) 85(4)/15(4) 80(3)/ 20(3)b 31(3)/69(3)
65.8(4) (Ni)
85.8(4) ###
M4, site 4a (0,0,0)
U
iso
0.0090(5) 0.0119(4)
0.0156(10)
0.0136(3)
0.0161(19)
0.0074(10)
0.0150(10) 0.0055(29)#
0.0050(3)#
0.0085(4)#
a
a
a
a
a
b
a
b
b
sof M, ratio Ni/M
99.9(17)
96.7(13)
97.5(25)
99.2(11)
54(6)/46(6)
60(2)/40(2)
98(3) 17.8(17) (Ni) 14.4(11)/85.6 95.6(10)/4.4
B1, site 24e (0,0,z)
z
0.2765(8) 0.2782(8)
0.0060(14) 0.0120(11)
0.2744(19)
0.0205(37)
0.2763(6)
0.0174(9)
0.2770(23)
0.0189(38)
0.2739(9) 0.2755(16)
0.2793(20)
0.2714(4)
0.2714(4)
U
eq/Uiso
0.0062(15) 0.0147(27) 0.01 fixed
0.0060(5)## 0.0080(6)##
a
a
a
a
a
a
a
a
a
a
sof
103(5)
88(4)
114(7)
107(2)
89(11)
90(5)
96(8)
102(7)
101(2)
101(2)
U
U
11¼U22
33
0.0067(18) 0.0146(16)
0.0047(36) 0.0068(25)
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
B2, site 32f (x,x,x)
x
–
–
–
0.1939(8)
0.0206(48)
51(5)
–
–
–
0.1933(15)
–
–
–
–
–
–
–
–
–
0.1954(17)
0.010 fixed
55(6)
0.1923(14)
0.0060(5)##
14.2(4)###
–
–
–
U
eq.
0.025 (fixed)
21.7(16)
sof
#, ##, ### coupled refinement.
a
In order to check for mixed occupations and/or vacancies site occupation factors were treated by turns as free variables at the end of the refinement. For the
discussion and the final refinement a full occupation is assumed.
b
Refined as mixed occupation Ni/M (M¼Ga, Zn, Sn, In, Ir), in sum a full occupation is assumed.
and Zn/Ni ratio is impossible. The residual electron density was
featureless, especially of the 32f site.
and wR
2
¼0.036. The displacement parameters of Ni4 (4a) and
Sn3 (8c) were slightly reduced and enlarged, respectively.
Accordingly, the refinement with mixed occupations resulted in
a small improvement (R
electron density was featureless.
1
¼0.0165, wR ¼0.030). The residual
2
2
.3.4. Ni20.0Zn2.4
The refinement with the ordered model of Ni20Zn
completely occupied sites yielded R-factors of R
¼0.043 and
wR
B
7.75
3 6
B and
1
2
¼0.060. The displacement parameter of the Zn atom on
2.3.6. Ni20.0Sn2.4B
7.75
the 8c site was three-times larger than the others. With a
The refinement was started with the model of Ni21Sn B . It
2 6
free site occupation factor the occupation decreased to 69% and the
turned out that the electron density is about 10% enlarged on site
R-values to R
1
¼0.016 and wR
2
¼0.020. The maximum of the
4a and 14% reduced on site 8c. The introduction of mixed Ni/Sn
ꢀ
3
˚
residual electron density was +1.2 e /A on a 32f-site with
xE0.196. With boron on this site (free sof, displacement parameter
occupations resulted in R-values of R
1
¼0.019 and wR ¼0.037.
2
ꢀ
3
˚
The highest residual electron density of +1.4 e /A was found for
a 32f-site with xE0.195, slightly above the background. Because
the crystal was from a boron-rich sample with enlarged lattice
parameter we assumed a partial exchange of Sn by boron
(7.6(14)%). Now the electron density map was ‘‘empty’’. An
additional substitution of Sn by Ni is possible but was not taken
into account. According to the small exchange and the low
electron number of boron the R-values are nearly unchanged.
fixed to 0.02) the refinement converged at R
wR
¼0.0125 and the residual electron density decreased signifi-
cantly to +0.36 e /A . The Ni/Zn ratio from the refinement is
9.4%/10.6%. This is close to the data obtained by EDX (88%/12%).
1
¼0.0119 and
2
ꢀ
3
˚
8
The small deviation may result from a partial Ni-occupation of the
sites 4a and 8c, but the difference between Ni and Zn is not
sufficient for a separation.
2
.3.5. Ni21.1Sn1.9
The refinement with a completely ordered model, i.e. Sn on 8c
and the composition Ni21Sn , resulted in R-values of R
¼0.018
B
6
2.3.7. Ni21.4In1.6
The refinement was started with the model of Adelsberger and
Jansen [16], i.e. the composition Ni21In and In on site 8c
B
6N
2
B
6
1
2 6
B

2
286
D. Kotzott et al. / Journal of Solid State Chemistry 183 (2010) 2281–2289
Table 4
Selected distances in
t-borides Co/M/B (M¼Al, Ga, Sn, Ir), esd’s in parentheses.
Compound Ni20Ga
B
3 6
Ni20.9Ga1.6
B
10.1 Ni21.4Zn1.6
B
6
Ni20Zn2.4
B7.75 Ni22.2In0.8
B
6
Ni22.0In1.0
B
6
Ni21.6In1.4
B
6
Ni12.3Ir8.7B10.4 Ni20.9Sn1.8B7.25 Ni21.1Sn1.9B
6
Ni1–B1 3x
Ni1–M3
2.069(5)
2.4036(8)
2.4701(10)
2.6101(6)
–
2.085(4)
2.3790(8)
2.5247(9)
2.6109(6)
2.255(2)
2.087(11)
2.426(2)
2.469(2)
2.623(1)
–
2.082(3)
2.4203(6)
2.4814(7)
2.6214(5)
2.293(3)
2.074(13)
2.462(3)
2.454(2)
2.6458(13)
–
2.088(5)
2.4703(11)
2.4428(13)
2.6502(6)
–
2.068(9)
2.4621(17)
2.4275(19)
2.6410(10)
–
2.188(11)
2.3871(7)
2.7074(8)
2.7074(8)
2.265(3)
2.105(2)
2.4697(5)
2.4462(5)
2.6528(2)
2.340(3)
2.099(2)
2.4751(5)
2.4334(5)
2.6540(2)
–
Ni1–M1 3x
Ni1–M2 6x
Ni1–B2 3x
Ni2–B1 2x
Ni2–M2
2.102(5)
2.122(4)
2.102(11)
2.391(3)
2.531(1)
2.531(1)
2.623(1)
2.8938(7)
–
2.115(3)
2.3895(8)
2.5360(6)
2.5360(6)
2.6214(5)
2.8959(5)
2.069(19)
2.127(13)
2.410(3)
2.5409(17)
2.5409(17)
2.6458(13)
2.9100(8)
–
2.107(5)
2.4194(15)
2.5346(8)
2.5346(8)
2.6502(6)
2.9109(5)
–
2.105(9)
2.417(2)
2.5184(13)
2.5184(13)
2.6410(10)
2.8991(7)
–
2.195(13)
2.5470(11)
2.5570(11)
2.5570(11)
2.7071(5)
3.0282(6)
2.186(24)
2.093(2)
2.4290(6)
2.5317(3)
2.5317(3)
2.6528(2)
2.9141(2)
2.067(18)
2.090(2)
2.4327(6)
2.5252(3)
2.5252(3)
2.6540(2)
2.9119(2)
–
2.3966(14) 2.4086(11)
Ni2–M2 4x
Ni2–M4
2.5109(8)
2.5109(8)
2.6102(6)
2.8835(5)
2.5234(7)
2.5234(7)
2.6109(6)
2.8979(5)
2.076(12)
Ni2–M1 4x
Ni2–M3 2x
Ni2–B2 2x
–
Ni3–M1 4x
Ni3–M2 12x
Ni3–B2 4x
2.4036(8)
2.8835(5)
2.3790(8)
2.8979(5)
1.027(16)
2.426(2)
2.8938(7)
–
2.4203(6)
2.8959(5)
1.04(3)
2.462(3)
2.9100(8)
–
2.4703(11)
2.9109(5)
–
2.4622(17)
2.8991(7)
–
2.3891(7)
3.0282(6)
1.04(3)
2.4697(5)
2.9141(2)
1.06(2)
2.4751(5)
2.9119(2)
–
–
Ni4–M2 12x
2.5109(8)
2.5234(7)
2.531(1)
2.5360(6)
2.5409(17)
2.5346(8)
2.5184(13)
2.5570(11)
2.5317(3)
2.5252(3)
B1–M1 4x
B1–M2 4x
2.069(5)
2.102(5)
2.085(4)
2.122(4)
2.087(11)
2.102(11)
2.082(3)
2.115(3)
2.074(13)
2.127(13)
2.088(5)
2.107(5)
2.068(9)
2.105(9)
2.188(11)
2.195(13)
2.093(2)
2.105(2)
2.090(2)
2.099(2)
B2–Ni3
–
–
–
–
1.025(14)
1.68(3)
–
–
–
–
1.04(3)
1.69(4)
–
–
–
–
–
–
–
–
–
–
–
–
1.04(3)
1.69(5)
1.06(2)
1.73(4)
–
–
–
–
B2–B2 3x
B2–M2 3x
B2–M1 3x
2.076(12)
2.255(2)
2.069(19)
2.26(2)
2.186(24)
2.265(3)
2.067(18)
2.340(3)
(
R
1
¼0.042, wR
2
¼0.059). The subsequent refinement with free
and M4 can be occupied by Ir, too. But according to the high
symmetry (and the low multiplicity) this approximation has only
a small influence on the overall composition, so we assumed an
occupation solely by Ni. The boron site 24e can be derived from
sof’s resulted in small deviations as indicated by the displacement
parameters and an amelioration of the R-values. While site 4a was
only occupied by Ni, site 8c showed a mixed occupation with
ꢀ
3
˚
6
9(3)% In (R
1
¼0.028, wR
2
¼0.040).
difference Fourier synthesis with a value of +8.1 e /A . At that
stage of the refinement there is a residual electron density of +4.4
–
3
˚
e /A on a 32f site with x¼0.192, which is clearly higher than the
2
.3.8. Ni22.0In1.0
B
63
ꢀ
3
˚
background of 72 e /A . The occupation of this site by boron
results in a sof of 55(6)%. To obtain a stable refinement the
According to the increased In-content of the melt and the
enlarged lattice parameter we expected a higher In-content on
sites 4a and/or 8c. This was true for the 4a-site with 40(2)% In, but
surprisingly the In-occupation on site 8c sank to 20(2)%.
According to the residual electron density +1.1 e /A on the
well-known 32f-site a small amount of additional boron might be
possible, but the quality of the data allows no clear statement, so
displacement parameters of boron were fixed. Finally, R-values of
2
R
1
(F)¼0.014 and wR
2
(F )¼0.023 were obtained.
ꢀ
3
The general findings (mixed occupation of M1 and M2 with
˚
preference of M1 by Ir, partial occupations of M3 and M4, residual
electron density on site 32f) were confirmed for several crystals
from boron-rich samples.
this is discarded. Final R-values are R
The decrease in the In-content was also confirmed by EDX
: 8%, Ni22.0In1.0 : 5%).
1
¼0.020 and wR ¼0.0295.
2
(Ni21.4In1.6
B
6
B
6
3. Results and discussion
2
.3.9. Ni22.2In0.8
The results of the refinement were very similar to the previous
one. In comparison to Ni22.0In1.0 the In-content on 8c was
B
6
The crystal structure of
2]. Boron or carbon has a quadratic antiprismatic coordination.
The same coordination is found for the binary borides of those
transition metals (Cr B, Mn B, Fe B, Co B, Ni B [1]), which are also
the most frequent metals for -borides. Fig. 1 shows the crystal
structure of a -phase with a special emphasis on the four different
t-phases was first determined for Cr23C
6
[
B
6
further decreased and that of 4a increased. The resulting
occupations by In were 54(6)% on site 4a and 15(4)% on site 8c.
The higher R-values are caused by the crystal quality.
2
2
2
2
2
t
t
metal positions. It is built up by a cubic face-centred packing of
cuboctahedra which are formed by the metal atoms M2 on site 48h
and centred by the metals M4 on site 4a. The octahedral ‘‘voids’’ are
filled by cubes of the metal atoms M1 on site 32f (M1) and the
atoms M3 on site 8c occupy the tetrahedral ‘‘voids’’. The quadratic
antiprismatic coordination of the boron atoms is formed by the
square planes of the cuboctahedra and the cubes. Coordination
polyhedra are shown in Fig. 2.
2
.3.10. Ni12.3Ir8.7B
10.4
According to the composition for the synthesis (58.3% Ni,
1.7% Ir) and the value of the lattice parameter (a¼10.932 A) we
˚
4
expected a partial occupation of the metal sites M1 and M2 by Ir.
It turned out very soon that the electron density is by far the
highest for M1, while it is reduced for M3 and M4. Therefore the
refinement was done in a way that the ratio Ni:Ir on site M1 was
fixed and the ratio on M2 refined as a free parameter. The sites M3
and M4 were assumed to be occupied by Ni and refined with free
occupation factors. The ratio on M1 was changed stepwise until
the total Ni–Ir ratio corresponded to the value which was
determined for that crystal by EDX (58.4% Ni, 41.6% Ir). This
procedure is not unambiguous, because in principle the sites M3
In general the distances M–M and M–B correspond to the sum
˚
˚
˚
˚
of the covalent radii (Ni: 1.25 A, Zn: 1.34 A, Ga: 1.35 A, In 1.67 A,
˚
˚
Sn: 1.54 A, Ir: 1.36 A) [17] and are therefore not discussed in
detail. The larger radius of Ir compared to Ni is clearly seen in the
distances M–B and the corresponding M–M distances (M1–M1,
M1–M2, M2–M2). An interesting feature is the surrounding M3

D. Kotzott et al. / Journal of Solid State Chemistry 183 (2010) 2281–2289
2287
on site 8c which is usually occupied by the large metal atoms.
2 atoms M2 form a truncated tetrahedron (Friauf-polyhedron)
with quite long distances, while the four hexagons of the
Friauf-polyhedron are capped by M1 with significantly shorter
distances. This seems to be important for the substitution
1
4
M32B -tetrahedron (see later).
3.1. Ni/Ga/B
The system Ni/Ga/B was already investigated by Stadelmaier
and Fiedler [18] and Chaban and Kuzma [19]. Both confirm the
existence of a -boride. According to Stadelmaier it melts
congruently. At 800 1C the stability extends to 21 at% boron and
–12 at% gallium. The lattice parameters vary between 10.481
t
6
˚
and 10.519 A, Ga prefers the site 8c (i.e. Ni21Ga
2 6
B with 8.7% Ga)
and there is no boron-rich phase. Kuzma confirms composition
˚
6
Ni19.5Ga3.5B ) and lattice parameter (10.512 A) but assumes a
(
homogenous distribution of Ga.
Our results confirm the existence of boron-poor t-boride with
˚
a lattice parameter around 10.50 A and a composition Ni20Ga
3 6
B .
The variation of the lattice parameter is clearly larger with an
˚
upper limit of 10.60 A.
The refinement of the single crystal data with the shortest
3 6
lattice parameter suggests a composition Ni20Ga B with Ga on
both sites 4a and 8c, but small amount of Ni cannot be excluded.
There is no significant occupation of M1 and M2 by Ga, in
agreement to the very rare examples of Ga–B bonds [14] and the
non-existence of binary Ga borides [1].
With increase in the boron content the lattice parameter rises
˚
up to a maximum of 10.60 A (ꢁ33 at% B). The single crystal data
clearly show a decrease in the electron density on site 8c and the
additional density on site 32f (xE0.194). The refinement
6 8
Fig. 1. Crystal structure of t-Borides M23B ; M1: M -cubes, M2: M12-cuboctahedra
centred by M3, isolated M-atoms: M4; grey circles: boron, black circles: metal
indicates a partial substitution of Ga by B
4
-tetrahedra. The
atoms.
˚
resulting Ni–B distances 2.08 A are comparable to other binary
˚
Ni borides. The B–B distances of 1.68(3) A within the tetrahedron
˚
are similar to molecules with B
˚
4
-tetrahedra (B
4 4
Cl : 1.65 A [20,21];
(tBu)
4
B
4
: 1.71 A [22]) and other boron-rich
t
-borides (Ni20AlB14:
B10.5: 1.63 A [4]). The perfect fitting of
M2
˚
˚
1
.68 A, [6], Co12.3Ir8.9
M3
the coordination around the 8c-site with symmetry 4¯ 3m for a
M1
M1
M3
M1
M1
M3
B
4
-tetrahedron is shown in Fig. 3.
In general, there is a strong analogy to the t-borides of the
system Ni/Al/B. Here, the substitution increases the lattice
M2
M2
M2
M2
M2
M2
M2
B1
B1
˚
˚
B1
parameter from 10.511 A for Ni20Al
3 6
B to 10.617 A for Ni20AlB14.
M1
We did not obtain suitable single crystals with lattice parameters
M2
M2
M2
M2
M2
M2
˚
above 10.56 A. This might be due to the conditions of crystallisation.
B1
B1
We expect the interaction with a liquid phase (solution/flux) is
essential to form single crystals of sufficient size. According to the
results of the powder investigations the maximum of the lattice
M1
M1
M1
M4
M1
˚
parameter is around 10.60 A. Estimated from the increase in the
volume by the substitution Ga-B the corresponding composition
of the boron-rich phase is quite close to Ni20GaB14
M2
M2
M2
M2
M2
4
M2
M2
.
M2
M2
M3
M2
M2
M4
M2
M1
M2
M2
Ni2
Ni2
M2
M1
M2
M1
Ni2
M2
M2
M2
M1
M2
M2
M2
Ni1
Ni1
B2
Ni2
B2
Ni1
Ni2
M1
Ni2
M1
M1
B2
B2
Ni2
B1
Ni2
Ni2
Ni2
Ni2
Ni2
Ni1
M2
M2
M2
M2
Fig. 2. Coordination polyhedra in
t-borides.
4 10.1
Fig. 3. Coordination of the B -tetrahedron in Ni20Ga1.6B .

2
288
D. Kotzott et al. / Journal of Solid State Chemistry 183 (2010) 2281–2289
3
.2. Ni20.0Zn2.4
B
7.75
decrease on site 8c. The reason why indium now prefers the 4a
site is not yet clear but the refinement indicates this without doubt.
The unexpected behaviour might be explained by a phase
separation. The existence of two liquid phases was already
described by Stadelmaier, who had investigated the solidified
melts by metallography. The binary diagram Ni/In performs
The existence of a
Stadelmaier et al. [23,24]. In the isothermal section at 800 1C they
t-boride with Zn was already described by
found an appreciable phase width of the
˚
t-phase with lattice
parameters between a¼10.498 A (72% Ni, 8% Zn, 20% B: Ni20.3Zn2.7
˚
several different binary alloys between the compositions In3Ni2
B
6
) and a¼10.555 A (63% Ni, 12% Zn, 25% B: Ni19.3Zn3.7
B7.25).
Our results are in excellent agreement with the results of
˚
and Ni
in our experiments were chosen in a way that there is an excess of
In with respect to the stable -phase. Therefore the coexistence of
two liquid phases around the temperature of the crystal growth of
the -phase is possible. This might explain, too, why we observe
In Ni as an In-rich intermetallic phase in the solidified regulus.
The increase in the cell volume with the increase in Ni-content
2
In, which all solidify around 900 1C [32]. The compositions
Stadelmaier. The boron-poor single crystal (a¼10.541 A) shows
nearly completely occupied metal sites on 4a and 8c and no
significant residual electron density. In this case the reliability is
limited because of the poorer crystal quality (see R-factors) and
the unfortunate electronic situation (separation Ni/Zn, extent of
additional boron).
t
t
3
2
˚
˚
For the single crystal from the boron-rich sample the occurrence
of an additional residual electron density on a 32f-site is connected
to a decrease in the electron density on site 8c and an enlargement
of the lattice parameter. For the investigated single crystal the
is puzzling because of the bigger size of In (rIn: 1.67 A; rNi: 1.28 A
[17]). The calculation of the valence sum according to the method
of Brown [33] shows that the efficient size of the coordination
sphere for M3 is larger than for M4. With the parameters of
occupation by the B
˚
4
-tetrahedra is about 20%. The distances Ni–B
˚
6
Ni22InB one obtains a valence sum of 7.84 for M3 and of 5.94 for
(
2.07 A) and B–B (1.69 A) fit very well to the other
t
-borides with
M4. Therefore one would expect that large metal atoms like Al, Ga,
Sn and In are preferentially located on site 8c. This is true for most
4
B -tetrahedra. But again the small difference of the electron
numbers between Ni and Zn and the limited exchange of Zn by
-tetrahedra reduce the reliability of the X-ray investigation.
of the examples, but not for the t-borides grown from an In-rich
melt. Nevertheless, is explains, why the lattice parameter is
B
4
enlarged despite the lower In content. Additionally, it corresponds
6 6
to the longer distances M4–Ni2 in Ni22.2In0.8B and Ni22.0In1.0B
3.3. Ni/Sn/B
and the shorter ones in Ni21.6In1.4B6.
It should be mentioned that the
t-borides of the system
The system Ni/Sn/B was investigated by Stadelmaier and
Co/In/B play a special role, too. The binary Co23B
6
was grown from
co-workers in the 1960s [24–26]. They found the
only ternary compound with a small variation of the Sn-content
6.0–6.8 at%) and larger one for boron (17.5–27.5 at%,
i.e. Ni21Sn –Ni21Sn ). Lattice parameters were between
0.584 and 10.598 A. Our results are in very good agreement.
The variation of the boron content is explained by the substitution
t
-boride as the
an In-melt and there is a miscibility gap of the liquid phases [29–32].
3.5. Ni12.3Ir8.7B10.4
(
a
2
B
5
2 9
B
˚
1
Till date there were no investigations on the ternary system
Ni/Ir/B. The tendency of Ir to form
Ir-content was well-known [4,5,7,34]. Furthermore, it was known
that -borides of Ni form extended solid solutions with platinum
metals (Ni23ꢀxRe , x¼10–13 [35]). Finally, the substitution
of B -tetrahedra was shown for the Ir-containing boron-rich
-boride Co12.3Ir8.9 10.5 [4]. Our results for Ni12.3Ir8.7
perfectly in line with these observations.
t-borides with a very high
˚
of Sn on site 8c by B
4
-tetrahedra. The distances (Ni–B: 2.07 A, B–B:
.73 A) are comparable to the other systems with B -tetrahedra.
The substitution has only a little influence on the lattice
parameter, because the volumes of B -tetrahedra and Sn are
comparable (see -borides Ni/Al/B). Therefore a statement on the
˚
1
4
t
x 6
B
4
4
t
t
B
B10.4 are
upper limit of the substitution is difficult. According to the results
of Stadelmaier it is about 20%.
According to the refinement iridium shows a clear preference of
the site M1 (84%), but M2 contains significant amounts of Ir, too
(
(
18%). This leads to an enlargement of the M–B distances
3.4. Ni/In/B
˚
2.19–2.20 A) in comparison to other binary and ternary Ni borides.
Furthermore, the sites M3 and M4 are only partially occupied. This
was already observed for Cr7.9Ir14.1 10.5. As a result
the distances M3–M1 are quite short despite the enlarged lattice
The system Ni/In/B was investigated in detail by Sch o¨ bel and
6
B and Co12.3Ir8.9B
Stadelmaier [27] Stadelmaier and Yun [28]. Besides the perows-
kit-related compound Ni InB0.5 they found at 800 1C a -boride
3
t
parameter. Finally, the uptake of additional boron is evident. The
with 20–24 at% boron. The In-content varied between 4 at%
˚
˚
˚
distances M–B (2.186 A) and B–B (1.69 A) are comparable to the
(
Ni21.8In1.2
B
5.8, a¼10.581 A) and 8 at%, (Ni20.6In2.4
B
7
, a¼10.609
values in related compounds. In total, the results are very similar to
˚
A). The
2 6
t-phase with the composition Ni21In B melts congru-
the boron-rich
t
-boride of the system Co/Ir/B.
ently. Additionally, they observed at higher In-content the
occurrence of two liquid phases, similar to the system Co/In/B
[
29–31]. In a recent contribution Adelsberger and Jansen [16]
4. Conclusions
The cubic -borides Ni23ꢀx
Ir were synthesised from the elements. On the basis of single
crystal data detailed statements on the existence of boron rich
communicated a completely ordered structure of Ni21In
2
B
6
(i.e. In
˚
0
0
on site 8c) by refinement of single crystal data (a¼10.5911 A).
Our results for the refinements of the single crystal from an In-
poor melt are in agreement with the work of Stadelmaier and
Jansen. The lower limit of the In-content corresponds to a lattice
t
M
xꢀy
B
6+4y with M ¼Zn, Ga, In, Sn,
0
phases and the arrangement of M were possible. It could be
˚
parameter of about 10.55 A (powder and single crystal), the
4
shown the substitution of metal atoms by B -tetrahedra on one of
smaller lattice parameter is explained by the partial occupation of
site 8c by In.
the metal sites (8c) is a common feature for Ni-containing
0
t
-borides. For M ¼Zn and Sn the extent of substitution is limited
˚
0
The enlargement of the lattice parameter to a ꢁ10.60 A by
(up to 20%), while for M ¼Ga there is a complete substitution
increase in the In-content was expected, but surprisingly we
observed a reduction of the single crystal’s In-content because the
increase of indium on site 4a is more than compensated by the
Ga2B
4
similar to the system Ni/Al/B. No additional boron was
0
observed for M ¼In, but there is an unexpected pattern of
substitution. This might be explained by a miscibility gap of the

D. Kotzott et al. / Journal of Solid State Chemistry 183 (2010) 2281–2289
2289
metallic melt. Iridium as a
exists only a boron-rich boride and Ni and Ir are in nearly equal
ratios. Detailed investigations were done for a single crystal with
t
-stabiliser plays a special role. There
P. Villars, K. Cenzual, in: Pearson’s Crystal Data: Crystal Structure Database
for Inorganic Compounds, ASM International, Materials Park, Ohio, USA,
2009/2010 Release.
[
2] A. Westgren, Jernkontorets Ann. 117 (1930) 501–512.
a composition Ni12.3Ir8.7
B
10.4. Iridium is mainly localised on site
[3] H.H. Stadelmeyer, Metal-rich metall-metalloid phases, in: B.C. Giessen (Ed.),
Developments in the Structural Chemistry of Alloy Phases, first ed., Plenum
Press, New York1969, pp. 141–180.
3
2f. The 8c-site is occupied by 60% with B -tetrahedra.
4
[
[
4] D. Kotzott, M. Ade, H. Hillebrecht, J. Solid State Chem. 182 (2009) 538–546.
5] P. Rogl, H. Nowotny, Monatsh. Chem. 104 (1973) 1325–1332.
5
. Note in addition
[6] H. Hillebrecht, M. Ade, Angew. Chem. 110 (1998) 981–983 Angew. Chem. Int.
Ed. 37 (1998) 935–938.
[
[
7] D. Kotzott, M. Ade, H. Hillebrecht, Solid State Sci. 10 (2008) 2921–302.
8] D. Kotzott, Diploma Thesis, Universit a¨ t Freiburg, Germany, 2006.
During the preparation of this article Sologub et al. [36] reported
on single crystal investigations on the -borides (Co0.64Ir0.36
(i.e. Co13.6Ir7.6 20Fe (i.e. Fe13.8Ir9.2
For the boron-poor phase Fe13.8Ir9.2 they confirmed the mixed Fe/Ir
occupation of the sites 48h and 32f with a preference of 32f as we
have found it for Cr7.9Ir14.1 [7] and Co16.2Ir6.8 [4]. For the boron-
rich phase Co13.6Ir7.6 10 describes the structure with a model in space
t
)
)
21
[9] M. Ade, Ph.D. Thesis, Universit a¨ t Freiburg, Germany, 1997.
10] Fa. STOE, programme XSHAPE, part of programme XAREA, Darmstadt,
Germany, 2004.
11] N. Vojteer, H. Hillebrecht, Angew. Chem. 118 (2006) 172–175 Angew. Chem.
Int. Ed. 45 (2006) 165–168.
[
[
Co0.16
B
4
B
6
B
10 and (Fe0.54Ir0.46
3
B
6
6
B ).
B
6
[
[
12] T. Ludwig, H. Hillebrecht, J. Solid State Chem. 179 (2006) 1623–1629.
13] V. Adasch, K.-U. Hess, T. Ludwig, N. Vojteer, H. Hillebrecht, Chem. Eur. J. 13
B
6
B
6
B
(2007) 3450–3458.
group F 4¯ 3m. The essential difference between the two models is a
splitting of the site 8c in Fm 3¯ m into two independent fourfold sites
in F 4¯ 3m. One of these two positions is completely occupied by
[14] M. Ade, D. Kotzott, H. Hillebrecht, J. Solid State Chem, in press,
doi:j.jssc.2010.05.009.
[
15] (a) G.M. Sheldrick, in: Programme SHELXL, University of G o¨ ttingen, Germany,
997;
b) G.M. Sheldrick, Acta Crystallogr. A 64 (2008) 112–122.
[16] T. Adelsberger, M. Jansen, Z. Anorg. Allg. Chem. 625 (1999) 438–442.
1
4
B -tetrahedra, the other one partially by Co (16%).
According to our experience this model has some short-
(
[
[
[
17] U. M u¨ ller, in: Anorganische Strukturchemie, Teubner-Verlag, 2006.
18] H.H. Stadelmaier, M.-L. Fiedler, Z. Metallkd. 60 (1969) 447–448.
19] N.F. Chaban, Y.B. Kuzma, Dopov Akad. Nauk Ukr. RSR (Ser. A) (1973) 550 cited
from [1].
comings. The displace parameter of the two independent boron
atoms differ by a factor of 20 (Uiso B1: 0.0039(12), B2: 0.0002(15)).
In comparison to our results, an early report on Co21Ta
the already mentioned Ni21In [16], and to recent findings of
Gumeniuk et al. [38] on Ni21Sc and Ni21Zr and the rare-
earth representatives RE2ꢀxNi21 (RE¼Er, Yb, Lu) [39] the value
2 6
B [37],
[
[
[
20] M. Atoji, W. Libscomb, Acta Crystallogr. 6 (1953) 547–550.
21] D. Tierry, Ph.D. Thesis, Univerity of Stuttgart 1992.
22] T. Menncke, P. Paetzold, R. Boese, D. Blaser, Angew. Chem. 103 (1991)
199–201 Angew. Chem. Int. Ed. Engl. 30 (1991) 172–174.
2 6
B
2
B
6
2 6
B
B
6
[
[
[
23] H.H. Stadelmaier, J.-D. Sch o¨ bel, L.T. Jordan, Z. Metallkd. 16 (1962) 752–754.
24] H.H. Stadelmaier, R.A. Draughn, G. Hofer, Z. Metallkd. 54 (1963) 640–644.
25] H.H. Stadelmaier, L.T. Jordan, Z. Metallkd. 53 (1962) 719–721.
for B2 is unreasonably small. Additionally, there is no significant
shift of the metal atom on site 48h x,x,0, which is severely
involved in the surrounding of the site 8c in Fm 3¯ m. Therefore
[26] H.H. Stadelmaier, J.B. Balance, Z. Metallkd. 58 (1967) 449–451.
[
[
[
27] J.-D. Sch o¨ bel, H.H. Stadelmaier, Z. Metallkd. 55 (1964) 378–382.
28] H.H. Stadelmaier, T.S. Yun, Z. Metallkd. 53 (1962) 754–756.
29] J.-D. Sch o¨ bel, H.H. Stadelmaier, T.R. Smith, Metall. (Berlin) 25 (1971) 10–12.
there are no significant differences between the structure models
except the splitting of the 8c-site (Fm 3¯ m) in 4c and 4d (F 4¯ 3m). The
only partial occupation of 4c by Co and the unusually small
displacement parameter of B2 on 4d suggest that the occupation
factor of B2 is higher and the description of structure in F 4¯ 3m
instead of Fm 3¯ m is not really validated by the X-ray data. The
refinement of our data of boron-rich crystals with the model of
Rogl did not result in a significant improvement. Nevertheless,
the data of Rogl et al. clearly confirm the existence of a boron-rich
[30] L.T. Jordan, R.K. Mathur, K.R. Brose, H.H. Stadelmaier, Z. Metallkd. 72 (1981)
782–784.
[
31] E. Ganglberger, H. Nowotny, F. Benesovsky, Monatsh. Chem. 96 (1965)
144–1146.
1
[32] H. Okamoto, in: Phase Diagrams for Binary Alloys, ASM International,
Materials Park, Ohio, 2000.
[
[
33] I.D. Brown, J. Appl. Cryst. 29 (1996) 479–480.
34] M. Ade, D. Kotzott, M. Robertson, H. Hillebrecht, Z. Kristallogr. Suppl. 20
(
2006) 151.
[35] E. Ganglberger, H. Nowotny, F. Benesovsky, Monatsh. Chem. 97 (1966)
01–102.
t
B
-boride Co13.6Ir7.6
B10 and the substitution of M3 on site 8c by
1
4
-tetrahedra as it was earlier proposed [4].
[
[
36] O. Sologub, P. Rogl, G. Giester, Intermetallics 18 (2010) 694–701.
37] H.H. Stadelmaier, H.H. Davis, H.K. Manatakla, E.T. Henig, Z. Metallkd. 80
(
1989) 370–373.
38] R. Gumeniuk, Y. Prots, W. Schnelle, A. Leithe-Jasper, Z. Kristallogr. NCS 223
2008) 327–328.
References
[
(
[
1] P. Villars, in: Pearson’s Handbook, Crystallographic Data for Intermetallic
Phases, ASM International, Materials Park, OH, 1997;
[39] J. Veremchuk, R. Gumeniuk, Y. Prots, W. schnelle, U. Burkhardt, H. Rosner,
Y. Kuzma, A. Leithe-Jasper, Solid State Sci. 11 (2009) 507–512.