Skip to main content
Internet Archive's 25th Anniversary Logo

Full text of "Structure, bonding, and hardness of CrB$_4$: a superhard material?"

See other formats

Structure, bonding, and hardness of CrB 4 : a superhard material? 





Haiyang Niu 1 , Jiaqi Wang 1 , Xing-Qiu Chen 1 !^ Dianzhong Li 1 , Yiyi 

Li 1 , Petr Lazar 2 ' 3 , Raimund Podloucky 2 , and Aleksey N. Kolmogorov 4 

1 Shenyang National Laboratory for Materials Science, Institute of Metal Research, 

Chinese Academy of Sciences, Shenyang 110016, China 

Center for Computational Materials Science, University of Vienna, Sensengasse 8, A-1090 Vienna, Austria 

3 Regional Centre of Advanced Technologies and Materials, 

Department of Physical Chemistry, Faculty of Science, Palacky University Olomouc, 

Tr. 11. Listopadu 12, 111 46 Olomouc, Czech Republic and 

4 Department of Materials, University of Oxford, 

Parks Road, Oxford 0X1 3PH, United Kingdom 

(Dated: March 1, 2013) 

By electron and X-ray diffraction we establish that the CrB4 compound discovered over 50 years 
ago crystallizes in the oPIO (Pnnm) structure, in disagreement with previous experiments but in 
agreement with a recent first-principles prediction. The 3D boron network in the new structure 
is a distorted version of the rigid carbon sp network proposed recently for the high-pressure C4 
allotrope. According to our density functional theory calculations and the analysis of the bonding, 
CrB4 is a potentially superhard material. In fact, the calculated weakest shear and tensile stresses 
exceed 50 GPa and its Vickers hardness is estimated to be 48 GPa. 

PACS numbers: 62.40.-K, 62.20.Qp, 71.20.Be, 










Covalent networks with high atomic densities and 
three-dimensional (3D) morphologies [lMa] are basic fea- 
tures of most of the known superhard materials, in- 
cluding diamond [Fig. 1(a)], 0BC2N, c-BN, and the 
recently found compounds C-BC5 [6j and 7-B28 [2n9(- 
Three new promising superhard allotropes of carbon with 
strong quasi- sp 3 covalent bonding as realized in a mono- 
clinic (M-carbonjlOl]). tetragonal body-centered (bct-C4 
111 Il2|). and orthorhombic (W-carbon 13]) structure 
have been proposed for the interpretation of the X-ray 
diffraction pattern of cold-compressed graphite 1A\. In 
particular, metastable bct-C4 is built up by an unusual 
framework [15 J of interconnected square C4 units [Fig. 
1(b)] and has been predicted to be superhard by several 
first-principles studies 16l4l9l]. Inspired by the search 

for superhard materials which can be fabricated without 
the need of an expensive high pressure [20| or a chemical 
vapor deposition [211 ] methods, we re-examine a known 
stable intermctallic CrB4 compound comprised of simi- 
lar B4 units. We find that, compared to the ReB2 com- 
pound shown recently to have a remarkably high hard- 
ness [2CJ, |22j-l2J], CrB4 holds the promise to have even 
more outstanding mechanical properties. 

First we characterize CrB4 experimentally by means 
of electron diffraction (ED) and X-ray diffraction (XRD) 
techniques confirming a first-principles prediction [25, 26] 
that the orthorhombic structure of CrB4, originally sug- 
gested to have the Immm space group (with the oIlO 
unit cell in Pearson notation, see Fig. 1(c)), has a lower- 
symmetry Pnnm space group (o-PIO, see Fig. 1(d)). On 
the basis of density functional theory (DFT) calculations 

27H30]. we establish that CrB4 with the newly claimed 
structure has lowest ideal tensile and shear strengths of 
51 GPa, which are comparable to those of cubic boron 

nitride (c-BN) . Making use of an empirical model [3l|, |32[ 
correlating the elastic moduli and Vickers hardness (H v ), 
we estimate H v « 48 GPa, which exceeds significantly the 
40 GPa threshold of superhardness. To rationalize this 
finding we perform a DFT study of the ten transition 
metal borides, TMB 4 (with TM = Ti, V, Cr, Mn,Fe as 
well as TM = Zr, Nb, Mo, Tc, Ru). Our results indi- 
cate that the atomic size and valence of the TM elements 
play a key role in determining the mechanical properties. 
The hardness reaches a maximum for TM=Cr when all 
bonding B quasi-sp 3 and hybridized Cr-B states become 

In 1968, Andersson and Lundstrom [33J reported the 
synthesis of CrB4 and characterized it as an orthorhom- 
bic o/lO structure. Given the very good fit of the X-ray 
pattern to the o/lO structure and the recently demon- 
strated elastic stability of the compound [34jj there was 
no reason to suspect incompleteness of this structural 
model. However, examination of the ground states of 

(a)Fd3m(cFS) (b) Ijmmm (/.IS) 

(c) Immm (o/lO) 

(d) Pnnm (oPlii) 

FIG. 1: (color online) Isosurfaces of the electron localization 
function (ELF) J35j], corresponding to a value of 0.75. (a,b): 
diamond and bct-C4 carbon; (c,d): orthorhombic structures 
of CrB4. Small and large balls denote B and Cr atoms, re- 

Fc-B [25j and Cr-B systems [26[ revealed a dynamical 
instability of the oIlO structure due to phonon modes 
with imaginary frequency near q = 0. As a consequence, 
the boron framework undergoes a significant distortion 
transforming the orthorhombic body-centered structure 
(oilO) into a primitive one (oPIO). It was observed 26 1 
that this structural transformation leaves the unit cell 
dimensions and the XRD patterns essentially unchanged 
(see Fig. S2 in Ref. |36j) which necessitates the use of an 
alternative characterization technique to finally resolve 
the structure of CrB4. 

A 20g sample with the initial composition of CrB4 
was prepared by repeated arc-melting of electrolytic 
chromium (from Alfa Aesar, claimed purity 99.997%) and 
crystalline boron pieces (from Alfa Aesar, claimed pu- 
rity 99.5%) under argon atmosphere. Cut sample pieces 
were sealed in quartz under argon and annealed in a 
high temperature furnace for 192 hours at 1250 °C. The 
annealed samples were characterized via metallographic 
microscope (LEISS Axiovert 200 MAT), scanning elec- 
tron microscope (SEM, HITACHI S-3400N) in the back- 
scattered electron mode (BSE). Our electron probe mi- 
croanalyser (EPMA, SHIMADZU EPMA-1610) results 
showed 20.374 at.% and 79.626 at.% elemental composi- 
tions of Cr and B, respectively. The presence, distribu- 
tion, and phase characteristics of CrB4 (78.06%), CrB 2 
(7.32%), and amorphous boron (14.62%) were further an- 
alyzed with an electron backscatter diffraction (EBSD) 
micrograph [36| . TEM characterization of finely ground 
samples was carried out with Tecnai G2 F20 S-TWIN 
transmission electron microscope. Finally, we obtained 
X-ray diffraction (XRD) patterns using a Rigaku diffrac- 
tometer and Cu K a irradiation (A=l. 54056 A) and per- 
formed full Rietveld refinement using the FULLPROF 
package [371 ]. 

For DFT calculations we used the Pcrdcw-Burkc- 

Ernzcrhof exchange-correlation functional 27] within the 
generalized gradient approximation and the projector- 
augmented waves method [28| as implemented in 
VASP[2i,|3^|. The energy cutoff was set at 500 eV. We 
allowed spin polarization for all TMB4 but only one com- 
pound of M11B4 showed a small non-zero magnetic mo- 
ment of about 0.7 hb per Mn in the antiferromagnetic 
ordering. A very accurate optimization of structural pa- 
rameters was achieved by minimizing forces (below 0.001 
eV/A) and stress tensors (typically below 0.5 kB). Fur- 
ther simulation details and the procedure for calculating 
the mechanical properties are described in the supple- 
mentary material[36]. 

Figure 2 shows our experimental ED patterns pro- 
jected along the [100], [110], [111] and [101] directions 
revealing that the unit cell has the dimensions |a*|«2.1 
nm" 1 , |b*|«1.8 nm _1 , |c*|«3.5 nm _1 and it can be clas- 
sified as a primitive orthorhombic lattice. The simulated 
ED pattern along [101] for the oPIO structure (Fig. 2f) 
shows additional reflections as compared to o/lO (Fig. 

(g)CrB 4 (oP10, Pnnm) I 



i i 1 1 i 



II II III I II II II I llll Illl 

lu *CrB : 


90 100 110 

FIG. 2: (color online) (a-c,e) Experimental electron diffrac- 
tion (ED) patterns along the [100], [110], [111] and [101] di- 
rections, respectively. (d,f): theoretical [101] ED patterns for 
o/10 and oPIO structures, respectively; (g) Rietveld refine- 
ment (Cu K a i) of the experimental X-ray diffraction data; 
reflections of CrB2 are indicated by vertical bars. Hollow ar- 
rows in (b,e,f) denote the second-order diffraction spots. Inset 
in (g) is a TEM image showing the phase boundary of CrB4 . 

2d), which is expected because the unit cell is doubled 
and the number of symmetry operations is reduced from 
16 to 8. The corresponding extra reflections are clearly 
present in the observed [101] pattern which unambigu- 
ously points at the oPIO structure. The oPIO structural 
model was further used to refine the powder XRD data 
and by that a good agreement between experiment and 
theory is obtained (see Table SI in Ref. |36|). 

The mechanism causing the observed distortion in 
C1T54 is many fold, because it is related to the quasi- 
sp 3 B-B bonding, the hybridization of the B- and Cr- 
like states, the atomic size of Cr relative to the avail- 
able volume in the B cage and the charge transfer be- 
tween B and Cr (i.e. the valencies). It is illustrative to 
look first at the evolution of the B network in the se- 
quence of the related structure types: diamond, bct-C4, 
o/lO, and oPIO (see Fig. 1). In diamond, the tetra- 
hedral arrangement of four nearest neighbors with the 
cos _1 (— 1/3) w 109.5° angles between the bonds is opti- 
mal for the sp 3 hybridization [38| (Fig. 1(a)). In bct-Gi, 
the symmetry of the local atomic environment is broken 
as two bonds form a 90° angle (Fig. 1(b)). In o/lO, 
symmetry is further reduced due to the two bonds now 
having different lengths (namely, 1.73 A and 1.86 A for 
structurally relaxed &B4, see Fig. 1(c)). Finally, in 
oP 10 a further deviation from the ideal sp 3 geometry 

-16 -12 -8 -4 

Energy E-E. (eV) 

FIG. 3: (color online) Calculated electronic structure of 
TMB4 compounds (TM: 3d transition elements), (a): elec- 
tronic densities of states (DOS) with occupied bonding and 
nonbonding states highlighted in grey and black, respectively. 
The corresponding insets of structures show charge density 
isosurfaces (0.1 e/A 3 ) for the energy windows corresponding 
to the nonbonding states, (b-e): variation of bond angles, 
bond lengths, atomic volume, and charge transfer to boron 
obtain by Bader's decomposition. [401] 

occurs as one of the two noncquivalent B sites has two 
B-B bonds at an angle well below 90° and the other is 
no longer four-fold coordinated (a fifth B atom is 2.13 
A away and the electron localization function shown in 
Fig. 1(d) develops a blob along the short diagonal of the 
B parallelogram). 

Within a semiempirical extended Hiickel approach 
Burdett et al. [39| studied the relative stability of carbon- 
and boron-based structures by analyzing the moments 
(/!„) of the electronic density of states (DOS). They ar- 
gued, that bct-Cj-carbon is less stable than diamond- 
carbon because (i) the non-optimal 90° angle gives rise 
to a strain energy and (ii) the four-membered rings re- 
sult in a higher ^4 which translates into a more uni- 
modal DOS and a lower stability for elements with a 
half-filled shell. Their conclusion that the second contri- 
bution could stabilize boron-based materials with a lower 
number of electrons is supported by our DFT calcula- 
tions: for the group-IV carbon the diamond structure is 
favored by 0.20 eV/atom while for the group-Ill boron 
the bct-C4 structure is favored by 0.08 eV/atom. The 
structural differences within the bct-Ci, o/lO, and oPIO 
family are less pronounced. However, the presence of 
three-membered rings in oPIO may significantly influence 
the structure's important third and forth DOS moments. 

To elucidate the stabilization role of the TM atom we 
carried out a series of DFT calculations for ten TMB4 
compounds. The DOS and formation enthalpies are pre- 
sented for the more stable of the o/lO or oPIO structure. 
For the 3d series Fig. 3(a) shows rather similar DOS 

profiles with the Fermi level (Ep) moving upwards as the 
electron count increases. In the exemplary CrB4 case, the 
DOS in the range of (-14, -5) eV is mostly of B-s — p-like 
character. The weight of the Cr DOS increases gradually 
and in the range of (-5, -2) eV a strong hybridization be- 
tween Cr- d xy (-d xz ) and B-p y (-p z ) states is observed. 
In the region from -2 to eV the Cr-d( x 2_ y 2) and Cr- 
d yz nonbonding states become dominant. The position 
of Ep in the pseudogap along with the lowest forma- 
tion enthalpy achieved for TM = Cr (see Fig. 0h) is 
consistent with the prediction [39( of maximum stability 
occurring in the middle of the 3d series. Further DFT 
calculations as detailed in the supplementary material 
[36| shed light on the stability competition between the 
oPIO and the o/lO structures. We find that Ti and V 
as well as Zr, Nb, and Mo tetraborides prefer the oIlQ 
type. Furthermore, the energy gained by the o/lO to 
oPIO transformation is larger for the 3d compounds and 
increases within both series from left to right. It is also 
noticeable, that the structural transformation is accom- 
panied by a volume reduction whereby over 80% of the 
energy gain comes just from the distortion of the B net- 
work. Finally, by artificially decreasing (increasing) the 
volume one can induce (disfavor) the distortion for all 
the considered TMB4 compounds. Figure 3(b-e) sum- 
marizes the structural trends and shows the variation of 
the average Bader's charge [40j. We employ the Bader's 
charge decomposition to illustrate that the charge trans- 
fer from the TM element to B (which is around 1 e/TM) 
decreases in the sequence from Ti to Fe. Hence, the dis- 
tortion could be explained by the decreasing number of 
electrons transferred to B. The derived geometrical re- 
sult, that three out of four B-B bonds have a minimum 
bond length for TM = Cr (Fig. 3c) demonstrates fur- 
ther why CrB4 is particularly stable. Considering that 
Ep in FeB4 moves from a deep valley in o/lO into the 
shoulder of the antibonding B-p-Fe-d peak in oPIO [25|, 
the optimality of the p-d bonding appears to be of less 
importance for the compound's stability (note that the 
unexpectedly high DOS at Ep in oP10-FeB4 makes the 
compound a good candidate to be a vhonon-mediated su- 
perconductor with a T c of 15-20 K 25]). 

The mechanical properties of &B4 are examined and 
rationalized via DFT calculation of the elastic proper- 
ties for the mentioned ten TMB4 compounds (see Ref. 
[36| ) . All of them are found to exhibit ultra incompress- 
ibility along the 6-axis and high bulk (B) and sheer (G) 
moduli. CrB4 is found to have the highest shear modu- 
lus (G = 261 GPa) and Pugh's ratio HH (k = G/B = 
261/265 = 0.985), which are two important elastic prop- 
erties thought to be strongly correlated to hardness [31| . 
The compound's low Poisson ratio of v = 0.12 is typ- 
ical for materials with strong covalent bonding. [20( 
Strikingly, the calculated lowest ideal shear and tensile 
strengths of 51 GPa are remarkably high and compara- 
ble to the lowest tensile strength of 55 GPa for super- 

TABLE I: Calculated elastic properties (in GPa) for CrEU and known (super)hard materials. The calculated bulk (B) and 
shear moduli (G) are Reuss-Voigt-Hill averages. The Vickers hardness estimates {H„ lc ) were obtaned with our proposed 
formula [3U , |33 | using the calculated elastic moduli. Finally, the experimental Vickers hardness values (H^ xp ) for diamond, 
BC2N, bct-C4, c-BN and B4C were taken from Refs. [3U 14a. |47| . Additional information is given in Table S5 [3q |. 




c 4 


C 6 








11 », 


BC 2 N 

bct-C 4 


B 4 C 








CrB 4 (oP10) 554 880 473 254 282 250 
CrB 4 (oI10) 591 931 467 252 280 225 





442 536 95.7 



















95 265 261 
97 275 259 


hard c-BN |42j. These values exceed considerably the 
lowest ideal shear strength of 34 GPa 42] in ReB2 which 
structure is comprised of buckled 2D boron nets[43j. As 
a corroboration for a possibly outstanding hardness, by 
breaking Cr-B bonds along the [00 ll direction we found 
the lowest critical cleavage stress [44j of 53 GPa, which 
matches the lowest ideal strengths. Finally, for esti- 
mation of the Vickers hardness (H v ) in terms of elas- 
tic properties we employ a recently proposed empirical 
modelJ3ll SI, H v = 2.0(/c 2 G) 0585 - 3.0 (H v and G in 

GPa) , which performs well across a large class of mate- 
rials and hardness values (Fig. HI panel (a)). Figure 0] 
reveals that the predicted behavior of H v for the TMB4 
compounds (panel b) mirrors the trend in their enthalpy 
of formation (panel c). The largest hardness value of 

Strain (%) 

Strain (%) 

4 5 6 7 
Valence electrons 

FIG. 4: (color online) (a): Vickers hardness, H v , as a func- 
tion of k 2 G, with k being the ratio of the shear (G) to bulk 
(B) modulus; the experimental data are discussed in the sup- 
plementary material [3y]; (b): H v for the most stable TMB4 
structures derived from the calculated B and G values (see 
text and Ref. |31| ); (c): formation enthalpy; (d,e): ideal tensile 
and shear strengths of oP10-CrB 4 . 

H v = 48 GPa for CrB4 is well above the superhardness 
threshold of 40 GPa and decreases rapidly for the con- 
sidered TMs. In particular, the isoelectronic but larger 
Mo atom streches the B network beyond its optimal size, 
leading to a 25% reduction in hardness. When compared 
against the known B4C material which can also be syn- 
thesized under ambient pressure, CrB 4 displays (see Ta- 
ble I) superior elastic properties and estimated H v (note 
that according to recent measurements B4C is not super- 
hard in its crystalline form [45|). 

Our findings make oP10-CrB4 a prime candidate to be 
an (up-to-now overlooked) affordable ambient-pressure 
superhard material. Measurement of the compound's 
Vickers hardness will be a challenge as pure CrB4 samples 
are difficult to produce with standard methods due to 
the particular behaviour of the Cr-B system in the high- 
temperature - B-rich part of the binary phase diagram 
[46|. Namely, the cooling of an arc- melted 1:4 elemental 
mixture leads unavoidably to a two-phase coexistence of 
CrB2 and B in a wide high-temperature region from 1830 
°C to 1500 °C. Formation of CrB 4 occurs below 1500 
°C but significant fractions of CrB2 and B can still be 
present after week(s) of sample annealing, as happened 
in the original [33] and present [36| studies. Our cur- 
rently best samples with 78% content of CrB 4 allowed us 
to reliably characterize the compound's crystal structure 
but were not suitable for investigation of its mechani- 
cal properties. Therefore, alternative approaches, such 
as the powder metallurgical process or the single-crystal 
growth method, may need to be employed to obtain sam- 
ples of desired quality. 

The confirmation of the new oPIO crystal structure of 
CrB 4 makes the prospect of synthesizing the FeB 4 phase 
with the same structure -predicted to be a viable high- 
temperature and high-pressure ground state of the Fe-B 
system 25ll26|- more exciting. Our detailed experimental 
and theoretical study of the presumably superhard com- 
pound CrB 4 demonstrates that materials with appealing 
properties may still be found in reportedly well-known 
binary systems. 

Acknowledgement We thank Xiaobing Hu and Prof. 
Shaobo Mi in the IMR for their valuable helps in per- 

forming and analyzing ED experiments and Prof. Shi 
Liu for his help in synthesizing experimental samples. 
We are grateful for supports from the "Hundred Talents 
Project" of Chinese Academy of Sciences and from NSFC 
of China (Grand Numbers: 51074151, 51174188) as well 
as Beijing Supercomputing Center of CAS (including its 
Shenyang branch in the IMR). A.N.K acknowledges the 
support from EPSRC CAF EP/G004072/1 in the UK. 













Corresponding author: xingqiu.chen@im[ 

R.B. Kaner, J.J. Oilman, and S.H. Tolbert, Science, 208, 

1268 (2005). 

V. V. Brazhkin, A. G. Lyapin, and R, J. Hemley, Philo. 

Mag. A, 82, 231 (2002). 

V. V. Brazhkin, N. Dubrovinskaia, M. Nicol, N. Novikov, 

R. Riedel, V. Solozhenko, and Y. Zhao, Nat Mater, 3, 576 


J. J. Oilman. Science 261, 143 (1993). 

S. Veprek, J. Nanosci. Nanotechnol., 11, 14 (2011). 

V. L. Solozhenko, O. O. Kurakevych, D. Andrault, Y. L. 

Godec, and M. Mezouar, Phys. Rev. Lett., 102, 015506 


V. L. Solozhenko, O. O. Kurakevych and A. R. Oganov, 

J. Superhard Mater., 30, 428 (2008). 

A. R. Oganov, J. H. Chen, C. Gatti, Y. Z. Ma, Y. M. 

Ma, C. W. Glass, Z. X. Liu, T. Yu, O. O. Kurakevych, 

and V. L. Solozhenko, Nature, 457, 863 (2009). 

E. Y. Zarechnaya, L. Dubrovinsky, N. Dubrovinskaia, Y. 

Filinchuk, D. Chernyshov, V. Dmitriev, N. Miyajima, 

A. El Goresy, H. F. Braun, S. Van Smaalen, I. Kantor, 

A. Kantor, V. Prakapenka, M. Hanfland, A. S. Mikhay- 

lushkin, I. A. Abrikosov, and S. I. Simak, Phys. Rev. 

Lett., 102, 185501 (2009). 

Q. Li, Y. Ma, A. R. Oganov, H. Wang, H. Wang, Y. Xu, 

T. Cui, H. K. Mao, and G. Zou, Phys. Rev. Lett. 102, 

175506 (2009). 

K. Umemoto, R. M. Wentzcovitch, S. Saito, and T. 

Miyake, Phys. Rev. Lett. 104, 125504 (2010). 

R. H. Baughman and D. S. Galvao, Phys. Rev. Lett. 104, 

125504 (2010). 

J.-T. Wang, C. Chen, and Y. Kawazoe, Phys. Rev. Lett. 

106, 075501 (2011). 

W.L. Mao, H.-k. Mao, P. J. Eng, T.P. Trainor, M. 

Newville, C.-c. Kao, D. L. Heinz, J. F. Shu, Y. Meng 

and R. J. Hemley, Science, 302, 425 (2003). 

H. Y. Niu, X.-Q. Chen, S. B. Wang, D. Z. Li, W. L. Mao, 

Y. Y. Li, Phys. Rev. Lett., 108, 135501 (2012). 

P. Y. Wei, Y. Sun, X.-Q. Chen, D. Z. Li, and Y. Y. Li, 

Appl. Phys. Lett., 97, 061910 (2010). 

F.M. Gao and X.F. Hao, Phys. Stat. Sol. (RRL) - Rapid 

Res. Let., 4, 200 (2010). 

HY. Niu, PY. Wei, Y. Sun, X.-Q. Chen, C. Franchini., 

D.Z. Li, and YY. Li, Appl. Phys. Lett., 99, 031901 























Q. Zhu, A. R. Oganov, M. A. Salvado, P. Pertierra, and 

A. O. Lyakhov, Phys. Rev. B, 83, 193410 (2011). 

J. B. Levine, S. H. Tolbert, and R. B. Kaner, Adv. Funct. 

Mater, 19, 3519 (2009). 

P. F. Mcmillan, Nat. Mat., 1, 19 (2002). 

J.B. Levine, J.B. Betts, JD. Garrett, S.Q. Guo, J.T. 

Eng, A. Migliori, and R.B. Kaner, Acta Mater, 58, 1530 


H.-Y. Chung, M.B. Weinberger, J.B. Levine, A. Kavner, 

J.-M. Yang and S.H. Tolbert, Science 316, 436 (2007). 

J.B. Levine, S.L. Nguyen, H.I. Rasool, J.A. Wright, S.E. 

Brown and R.B. Kaner, J. Am. Chem. Soc. 130, 16953 


A. N. Kolmogorov, S. Shah, E. R. Margine, A. F. Bialon, 

T. Hammerschmidt, and R. Drautz, Phys. Rev. Lett., 

105, 217003 (2010). 

A. F. Bialon, T. Hammerschmidt, R. Drautz, S. Shah, E. 
R. Margine, and A. N. Kolmogorov, Appl. Phys. Lett., 
98, 081901 (2011). 

J. P. Perdew, K. Burke, and M. Ernzerhof, Phys. Rev. 

Lett. 77, 3865 (1996). 

P. E. Blochl, Phys. Rev. B, 50, 17953 (1994). 

G. Kresse and J. Hafner, Phys. Rev. B 47, 558 (1993). 

G. Kresse and J. Furthmuller, Phys. Rev. B 54, 11169 


X.-Q. Chen, H. Y. Niu, D. Z. Li and Y. Y. Li, Inter- 

metallics, 19, 1275 (2011). 

X.-Q. Chen, H. Y. Niu, C. Franchini, D. Z. Li and Y. Y. 

Li, Phys. Rev. B, 84, 121405(R) (2011). 

S. Andersson, and T. Lundstrom, Acta Chem. Scand., 

22, 3103 (1968). 

H. B. Xu, Y. X. Wang and V. C. Lo, Phys. Stat. Sol. 

(RRL) - Rapid Res. Let., 5, 13 (2011). 

B. Silvi and A. Savin, Nature (London), 371, 683 (1994). 
See supplementary material at 
http://link.aps. orgsupplementary/ 

J. Rodriiez-Carvajal, Physica B, 192, 55 (1993). 
D. G. Pettifor, Bonding and Structure of Molecules and 
Solids, (Clarendon Press) 1995. 

J. K. Burdett and S. Lee, J. Am. Chem. Soc. 107, 3063 

W. Tang, E. Sanville and G. Henkelman, J. Phys.: Con- 
dens. Matter 21, 084204 (2009). 
S. F. Pugh, Philos. Mag. Ser. 7, 45, 823 (1954). 
R. F. Zhang, S. Veprek, and A. S. Argon, Appl. Phys. 
Lett., 91, 201914 (2007). 

X.-Q. Chen, C. L. Fu, M. Krcmar, and G. S. Painter, 
Phys. Rev. Lett., 100, 196403 (2008). 
P. Lazar and R. Podloucky, Phys. Rev. B, 78, 104114 

S. Grasso, C. Hu, O. Vasylkiv, T. S. Suzuki, S. Guo, 
T. Nishimura and Y. Sakka, Scripta Materialia, 64, 256 

T. B. Massalski and H. Okamoto, Binary Alloy Phase 
Diagrams, (ASM Intl; 2nd ed., ASM International 1990). 
P. S. Kislyi, Superhard and Refractory Materials, (Kiev: 
Institute of Superhard Materials), p. 86, 1985