Ae = Ca, Sr and Ba - Department of Materials and Environmental ...

AeAlSiH (Ae = Ca, Sr and Ba) Novel 

Semiconducting Zintl Phase Hydrides with 

Tuneable Band Gaps and Strong Metal-H 

Bonds 

Licentiate thesis 

Tomohiro Utsumi 

Division of Structural Chemistry 

Department of Physical, Inorganic and Structural Chemistry 

Stockholm University 

2008

Akademisk avhandling som för avläggandet av licentiatexamen i strukturkemi framlägges till 

offentlig granskning i Rum 5Ö vid Arrhenius laboratoriet, Svante Arrhenius väg 12, 

måndagen den 5:e maj 2008, klockan 1000. 

Extern granskare Dr. Sabrina Sartori, Institute for Energy Technology, Norge 

Intern granskare Prof. Osamu Terasaki, Stockholms universitet 

2

Abstract 

This thesis reports the synthesis and characterisation of two novel semiconductor Zintl 

phase hydrides CaAlSiH and BaAlSiH. CaAlSiH was obtained by direct hydrogenation of 

CaAlSi for 50 minutes at 500 ˚C with 90 bars hydrogen pressure. BaAlSiH was obtained by 

direct hydrogenation of BaAlSi above 600 ˚C using 70 bars of hydrogen pressure for 2 days. 

Similarly deuterated AeAlSiDs (Ae = Ca and Ba) were used to determine the full structure by 

neutron diffraction. Both compounds have the SrAlSiD structure type in space group P3m1. 

The cell parameters were determined to a = 4.133(1) Å c =4.761(2) Å for CaAlSiD and a = 

4.3087 (6) Å c = 5.203(1) Å for BaAlSiD, respectively. 

CaAlSiH and BaAlSiH started to release hydrogen around 420 ˚C and 500 ˚C, 

respectively, when heated in vacuum. These high decomposition temperatures indicate that 

both CaAlSiH and BaAlSiH are fairly stable hydrides. This stability was attributed to a strong 

ionic interaction between Ae (Ca and Ba) and H in addition to the covalent Al-H bond. The 

hydrogenation/dehydrogenation reaction was found to be reversible according to 2AeAlSi + 

H2 ↔ 2AeAlSiH (Ae = Ca and Ba). 

For the barium compound, we could gradually substitute Si - with isoelectric (Al-H) - 

entities into BaAl2-xSixH2-x (0 < x < 2), by direct hydrogenation of BaAl2-xSix. 

The a-axis length is linearly decreasing with increasing x-value, as the smaller Si is 

substituted by Al. The c-axis is linearly increasing with x, as a larger [Si] - -lone pair is 

substitute by an [Al-H] - unity. DFT calculations show that the end compositions BaSi2 (x = 2) 

and BaAl2H2 (x = 0) are metal conductors whereas the intermediate BaAlSiH (x = 1) is a 

semiconductor with a band gap of 0.71 eV. This opens up for a possibility to have tuneable 

band gap by changing the ratio of (Al-H) - to Si - . 

3

List of papers 

The thesis is based on the following papers: 

1. T. Utsumi, T. Björling, D. Moser, U. Häussermann, D. Noréus, “A series of Zintl phase 

hydrides; BaAl2-xSixH2-x (0.4 < x < 1.6) with compositions and structures in between the 

electric conductors BaSi2 and BaAl2H2 ” To be published. 

2. T. Björling, T. Utsumi, D. Moser, B. C. Hauback, U. Häussermann, D. Noréus, 

“Characterisation of two new Zintl phase hydrides BaAlSiH and CaAlSiH” To be published. 

4

Contents 

1 INTRODUCTION.............................................................................................................................................. 6 

2 EXPERIMENTAL SECTION ........................................................................................................................ 10 

2.1 SYNTHESIS .................................................................................................................................................. 10 

2.1.1 BaAl2-xSix (0.4 < x < 1.6) and CaAlSi alloys....................................................................................... 10 

2.1.2 BaAl2-xSixH2-x (0.4 < x < 1.6), CaAlSiH and BaAlSiD........................................................................ 10 

2.2 CHARACTERIZATION ................................................................................................................................... 11 

2.2.1 X-ray powder diffraction .................................................................................................................... 11 

2.2.2 X-ray single crystal diffraction........................................................................................................... 11 

2.2.3 Neutron diffraction ............................................................................................................................. 12 

2.2.4 Thermal analysis................................................................................................................................. 13 

2.3. COMPUTATIONAL DETAILS ......................................................................................................................... 13 

2.3.1 Computational details......................................................................................................................... 13 

3. RESULTS AND DISCUSSION ..................................................................................................................... 14 

3.1 ALLOYS....................................................................................................................................................... 14 

3.1.1 BaAlSi and CaAlSi alloys ................................................................................................................... 14 

3.1.2 BaAl2-xSix (0.4 < x < 1.6) alloys ........................................................................................................ 17 

3.2 HYDRIDES ................................................................................................................................................... 19 

3.2.1 BaAlSiH and CaAlSiH........................................................................................................................ 19 

3.2.2 BaAl2-xSixH2-x (0 < x < 2)....................................................................................................................32 

3.3 COMPUTATIONAL RESULTS ......................................................................................................................... 38 

3.3.1 Computationally relaxed structural parameters................................................................................. 38 

3.3.2 DOS of AeAlSiD ................................................................................................................................ 39 

3.3.3 DOS of BaAl2-xSixH2-x......................................................................................................................... 40 

4. SUMMARY ..................................................................................................................................................... 42 

4.1 ALLOYS OF CAALSI AND BAAL2-XSI X.......................................................................................................... 42 

4.2 ZINTL PHASE HYDRIDES, CAALSIH AND BAALSIH..................................................................................... 42 

4.3 ZINTL PHASE HYDRIDE, BAAL2-XSI XH2-X (0.4 < X < 1.6) ............................................................................... 43 

7. REFERENCES................................................................................................................................................ 46 

5

1 Introduction 

Hydrogen storage in metal hydrides has a number of advantages over gaseous high 

pressure storage or liquid LH2 cryo storage. Gaseous storage even at the highest available 

pressures today, takes a lot of space. This is especially critical when a filling station, with a 

capacity comparable to a normal gas station has to be fitted into urban areas. Compressors are 

also costly and energy consuming. Liquid storage is even more costly as the energy needed to 

liquefy hydrogen at temperatures below -253 o C corresponds to 1/3 of the total energy content 

of the stored hydrogen. Boil off losses adds further to this during practical use. 

Well functioning hydrogen storage is, however, important when trying to realize a 

future “hydrogen economy” where hydrogen produced from renewable energy sources is 

envisioned to become an environmental friendly fuel. This has promoted research and 

development of new materials, and in certain niche markets hydrogen storage in metal 

hydrides has already led to commercial products. The largest is in rechargeable NiMH 

batteries where the metal hydride is both hydrogen storage as well as the hydrogen electrode. 

Interestingly are NiMH batteries the prime choice of batteries for Hybrid Electric Vehicles 

(HEV) such as Toyota Prius and Honda Insight. More models are presently being launched 

and this rapidly growing HEV market will help to reduce the fuel consumption in the 

transport sector. In a sense one can say that fundamental research on hydrogen storage 

hydrides is already helping to reduce green house gas emissions. 

Another coming and probably even more important use of metal hydrides will be in 

hydrogen storage systems for fuel cells. This is so far a niche market mainly developed for 

military submarines, but civilian application such as power back up systems for mobile phone 

base stations and auxiliary power units for airplanes and trucks based on fuel cells are soon 

expected to be commercial. Improved metal hydrides with higher storage capacities at lower 

cost and with better resistance against unintentional oxidising impurities will accelerate this 

development. The long term solution to both climate problems, as well as limited resources of 

fossil fuel, will probably contain fuel cell propelled vehicles using hydrogen produced by 

renewable energy. 

6

The commercial metal hydride production is now several thousand of tons per year. 

The main advantage is the very high volumetric density of hydrogen in the metal hydride. 

This is close to twice of that in liquefied hydrogen and it is also the reason for the doubling of 

the capacity of NiMH batteries compared to the corresponding sizes of NiCd batteries. 

When hydrogen is absorbed in a metal hydride, the bond between the atoms in the H2 

molecule is broken as electrons from the metal lattice is filled into the anti bonding orbitals of 

the H2 molecule and the H atoms are intercalated into the interstitial sites between the metal 

atoms. The bond between the hydrogen and the metal atoms should neither be too strong nor 

too weak, to make it possible for hydrogen to be stored and released at temperatures and 

pressures close to ambient. This is accomplished by alloying electropositive metals from the 

left in the periodic table, which form stable hydrides, with less electropositive metals from the 

right, which usually form unstable or no hydrides. Typical hydrogen storage hydrides are 

FeTiH2, TiMn2H3, LaNi5H6 etc. An H/M ratio around 1 and an average metal atom weight 

around 50 grams per mole leads to a weight storage of around 2 wt%. To make hydrogen 

propelled cars with somewhat comparable performance to conventional cars, car industries 

have globally targeted storage capacities in excess of 6 wt% and assume metals that are more 

available and less costly than those presently used in commercial metal hydrides. This has 

focussed the research on metal hydride for hydrogen storage down to a few possible metals 

such as Li, Na, Mg and Al, which for weight reasons have to make up for most of the content 

in the metal hydride. 

Aluminium based alanates i.e. NaAlH4 and Na3AlH6 have attracted considerable 

interest as possible reversible hydrogen storage systems, since Bogdanović et al. discovered 

that a titanium containing catalyst could significantly improve the usually sluggish 

hydrogen/dehydrogenation reactions [1]. Typically, Al-hydrides (alanates) consists of isolated 

tetrahedral [AlH4] - or octahedral [AlH6] 3- entities [2][3]. The Al-H bond is, however, rather 

weak and the equilibrium pressure is high and not affected by the added Ti-catalyst, the role 

of which also still is a matter of controversy [4]. The high equilibrium pressure necessitates 

impractical high hydrogen absorption pressures, when hydrogen should be stored in the 

system. This limits their practical applicability for regenerative hydrogen storage. To work as 

practical hydrogen storage, alanate needs to reach faster kinetics as well as to strengthen the 

hydrogen bond to the metal atom framework. 

7

In the search for other Al containing hydrides Gingl et al. discovered the Zintl 

phase hydride SrAl2H2 [5]. A Zintl phase is a material class between intermetallic and ionic 

compounds. SrAl2H2 is called a Zintl phase hydride because hydrogen atoms are part of the 

Zintl anion [Al2H2] 2- . The compound has a very interesting structure, because SrAl2H2 

contains both Al-Al and Al-H bonds. In SrAl2H2, Al atoms are formally reduced by 

electropositive Sr. These Al atoms form with hydrogen a polyanionic Zintl anion [Al2H2] 2- . 

Each Al atom is surrounded by four neighbours, one H atom and three Al atoms, in a 

tetrahedral fashion (cf. Figure 10 where structures are further discussed). This coordination 

has not been observed before and is totally different from typical alanates. Moreover, the 

existences of the Zintl phase are also interesting in perspective of alanate synthesis. If the 

SrAl2H2 is continued to be hydrogenated, an alanate phase Sr2AlH7 [6] will appear. Still, the 

Al-H bonds in SrAl2H2 are fairly weak but more moderate hydrogen pressures are sufficient 

during the synthesis. 

A second Zintl phase hydride was discovered, by substituting an [AlH] - unit with an 

isoelectric Si - . Björling et al recently reported on a new semiconducting Zintl phase hydride 

SrAlSiH [7]. This material is structurally closely related with previously discussed SrAl2H2. 

To obtain SrAlSiH, half of the [Al-H] - entities in SrAl2H2 are substituted by isoelectric Si - . 

Interestingly, this substitution result in striking different properties. In contrast to thermally 

labile SrAl2H2, SrAlSiH is more stable, air and moisture insensitive. The thermal 

decomposition temperature of SrAlSiH is 600 ℃ compared to 300 ℃ for SrAl2H2, when the 

hydrides are heated in a DSC. In other words, substitution of [Al-H] - with Si - makes the 

metal-H bond stronger. If this can transferred to alanates, it would be possible to improve 

their properties. Another difference between SrAl2H2 and SrAlSiH is the electric conductivity. 

SrAl2H2 is a conductor, but, SrAlSiH is a semiconductor with a narrow indirect band gap. 

Semiconductors with narrow band gap are unusual and notable for metal hydrides. 

Additionally, SrAlSiH has further interesting characteristics. In SrAlSiH, it is possible to 

change the stoichiometric ratio without breaking the basic structure. The formula can be 

expressed as SrAl2-xSixH2-x. This may allow a control of the conductivity by changing the 

composition. I.e. a tuneable band gap can be achieved by substituting the (Si) - lone pair with 

an (Al-H) - unity. SrAlSi which is the precursor material of SrAlSiH is further a super 

conductor with a Tc = 5.1 K [8]. CaAlSi is also a superconductor with a Tc = 7.7 K [8]. Both 

structures are closely related to that of MgB2 with a Tc = 39 K [9]. It would be interesting to 

investigate the superconducting as well as the normal conductive properties of (Ca, Sr, and 

8

Ba)Al2-xSixH2-x and how they depend on the both metal atom composition and hydrogen 

content. 

In this thesis, I want to report on CaAlSiH and BaAlSiH which are homologous 

compounds to SrAlSiH. Both CaAlSi and BaAlSi were synthesized by arcmelting, and the 

corresponding hydrides and deutrides were synthesized and characterized. The refined 

structures were compared with computationally calculated values, including also DOS 

calculations. 

In the barium compound we also varied the Al to Si ratio according to BaAl2-xSix (0.4 

< x < 1.6) and made the corresponding hydrides in order to study how stoichiometry 

influences the properties. By studying these zintl phase hydride system, we hope to 

understand how hydrogen binds to aluminium in the presence of other atoms. This can help us 

to develop better hydrogen storing system based on Al e.g. different alanates, but these new 

Zintl phase hydrides have also interesting electric properties such as possible tunable 

semiconductivity as well as superconductivity. 

9

2 Experimental section 

2.1 Synthesis 

2.1.1 BaAl2-xSix (0.4 < x < 1.6) and CaAlSi alloys 

Commercially pure Si powder, Al powder, Ba and Ca ingots were delivered from 

MERCK, Alfa Aesar, Aldrich and ABCR respectively. The alloys, BaAl2-xSix (0.4 < x < 1.6) 

and CaAlSi were synthesized by arcmelting stoichiometric ratios of the elements in a MAM-1 

(Edmund Buhler) arcmelting furnace. The heating current was 15 A and the ingot cup was 

cooled by water. The melting was done under argon atmosphere. All of the samples were 

analyzed by powder x-ray diffraction. Some samples exhibited diffraction patterns with rather 

broad peaks. The broad peaks were attributed to poor crystallinity. They were especially 

frequent for the more aluminium rich samples. Such samples were subjected to a heat 

treatment at 500 ℃ for 2 days, whereby the shapes of the diffraction peaks improved. All 

samples were handled under argon atmosphere in a glove box with less than 1 ppm of O2 and 

H2O. 

2.1.2 BaAl2-xSixH2-x (0.4 < x < 1.6), CaAlSiH and BaAlSiD 

To hydrogenate the prepared BaAl2-xSix (0.4 < x < 1.6) and CaAlSi alloys, the samples 

were divided into smaller pieces to fit into a corundum tube and reacted with hydrogen in a 

stainless steel autoclave. A stainless steel sealed thermo couple was inserted into the sample 

to record the reaction temperature. The reaction conditions are complied in section 3-2. 

10

2.2 Characterization 

2.2.1 X-ray powder diffraction 

All reactants and products obtained were investigated by powder x-ray diffraction, 

using a Guinier-Hägg focusing camera of diameter 40 mm, with monochromated CuKα1 

radiation (λ = 1.5405980 Å). These samples were mixed with an internal standard of silicon 

before recording the patterns. The films obtained were measured in an LS 18 film scanner 

[10]. The program SCANPI [11] was used to determine d-values and intensities recorded in 

the photographs. The programs TREOR [12] and PIRUM [13] were used to index the patterns 

and refine unit cell parameters. 

2.2.2 X-ray single crystal diffraction 

A single crystal of BaSiAl was used for single crystal X-ray structure determination 

(cf. Table 1). Intensity data were collected by a STOE IPDS image-plate rotating anode 

diffractometer, operated at 50 kV and 90 mA at 291 K using graphite-monochromatized Mo 

Kα radiation (λ = 0.71073 Å). The distance from detector to crystal was 60 mm, and 2 theta 

range was 0 to 200º. The intensities of the reflections were integrated using the STOE 

software supplied by the manufactures of the diffractometer. Numerical absorption correction 

was performed with the programs X-red [14] and X-shape [15]. The structure was solved by 

direct methods SHELXS97 [16] and refined by full matrix least squares on F2 using program 

SHELXL97 [17]. Molecular graphics were prepared with the program DIAMOND [18]. 

11

Table 1: Crystal data for BaAlSi 

Compound BaAlSi 

Formula weight (g/mol) 192.41 

Temperature (K) 291(2) 

Wavelength (Å) 0.71073 

Crystal system Hexagonal 

Space group P6/mmm (191) 

Unit cell dimensions (Å) a =4.311(1) 

b =4.311(1) 

c = 5.155(2) 

Volume (Å 3 ) 82.98 

Formula unit/cell, Z 1 

Density (calculated) (g⋅cm -3 ) 3.850 

Absorption coefficient (mm -1 ) 12.28 

Absorption correction Numerical 

F(000) 83 

Crystal colour Metallic 

Crystal size (mm 3 ) 0.004188 

θ range for data collection (º) 7.9-56.1 

Index ranges -5 ≤ h ≤ 5 

-5≤ k ≤ 5 

-6≤ l ≤ 6 

Reflections collected 807 

Independent reflections 61[R(int) = 0.0185] 

Refinement method Full-matrix least squares on F 2 

Data/restraints/parameters 807/0/5 

Goodness-of-fit on F 2 0.994 

Final R indices [I > 2θ(I)] R1 = 0.0185 

wR2 = 0.0469 

R indices (all data) R1 = 0.0185 

wR2 = 0.0469 

Largest diff. hole and peak (e⋅Å -3 ) -1.39 and 0.52 

2.2.3 Neutron diffraction 

All atomic positions of CaAlSiD and BaAlSiD were determined from Rietveld profile 

refinements of neutron powder diffraction data from deuterided samples measured at room 

temperature at Kjeller, Institute for Energy technology (IFE), Norway (λ =1.5554 Å for 

CaAlSiD, and 1.54675 Å for BaAlSiD). The program FULLPROF [19] was used for the 

refinement of structural parameters of CaAlSiD and BaAlSiD. Neutron diffraction spectra of 

CaAlSiD and BaAlSiD are shown in Figure 5 and Figure 8. 

12

2.2.4 Thermal analysis 

The thermal behaviour of the obtained hydrides was investigated by degassing the 

samples into a tank with known volume while recording temperature and pressure. The 

samples were mounted in a heat resistant corundum tube in a stainless steel autoclave 

connected to the tank, which could be evacuated by a vacuum pump. A pressure gauge and 

thermocouples recorded pressure and temperature to register when the dehydrogenation 

started. 

2.3. Computational details 

2.3.1 Computational details 

Total-energy calculations for BaAlSiH and CaSiAlH were performed in the 

framework of the frozen core all-electron Projected Augmented Wave (PAW) Method [20], as 

implemented in the program VASP [21]. The energy cut-off was set to 500 eV. Exchange and 

correlation effects were treated by the generalised gradient approximation (GGA), usually 

referred to as PW91 [22]. The integration over the Brillouin zone was done on special k-point 

determined according to the Monkhorst-Pack scheme [23]. Total energies were converted to 

at least 1meV/atom. Structural parameters were relaxed until forces had converged to less 

than 0.01 eV/Å. 

13

3. Results and discussion 

3.1 Alloys 

3.1.1 BaAlSi and CaAlSi alloys 

The single X-ray refinement revealed that the BaAlSi alloy crystallize with hexagonal 

symmetry in space group P6/mmm (191) with cell parameters a = 4.311(1) Å and c = 5.155(2) 

Å. The BaAlSi structure is related to that of the superconductor MgB2 with a presumed 

disordered arrangement of Al and Si on the B position. This structure (cf. Figure 1) consists of 

[SiAl] 2- flat hexagonal layers, which are stacked on top of each other. In these layers, Si and 

Al atoms are covalently bonded to each other, forming a hexagonal network. It is assumed 

that Si and Al atoms are more or less randomly distributed over the 2d site. The Ba atoms are 

sandwiched between these layers. The final refinement gave atomic parameters according to 

Table 2 and an R1 value of 1.85 %. Selected interatomic distances in BaAlSi are shown in 

Table 3. It is also possible that BaSiAl crystallizes in space group as P-6/m2 (187), where Si 

and Al atoms are ordered by occupying alternately in the hexagonal network (Figure 2). 

Recently, based on single crystal synchrotron X-ray and powder neutron diffraction, Kuroiwa 

et al. reported that the Si and Al atoms are ordered in this way in the related CaAlSi system 

[24]. However, for BaAlSi, this ordering is difficult to detect with ordinary X-ray diffraction 

experiment as Al and Si are neighbours in the periodic table. 

Fig. 1: Structure of BaAlSi (Space group 191) Fig. 2: Structure of BaAlSi (Space group 187) 

14

Table 2: Atomic positions of BaAlSi 

Atom Site x y z Occupancy U11 U22 U33 U12 

Ba a 0 0 0 1 0.01329 0.01329 0.01136 0.00664 

Al 2d 1/3 2/3 1/2 1/2 0.01027 0.01027 0.02285 0.00513 

Si 2d 1/3 2/3 1/2 1/2 0.01027 0.01027 0.02285 0.00513 

Table 3: Selected interatomic distances (Å) in BaAlSi 

Ba-6 Ba 4.311(1) 

-2 Ba 5.156(2) 

-12 Al/Si 3.583(1) 

Al/Si-3 Al/Si 2.489(1) 

-6 Ba 3.583(1) 

-2 Al/Si 5.156(2) 

15

Powder x-ray diffraction patterns of BaAlSi and CaAlSi were also recorded. On the 

BaAlSi film no impurity phase was observed. However, on the CaAlSi film, small peaks of 

CaAl2Si2 [25] were observed as well as some peaks from some so far unidentified impurities. 

Both BaAlSi and CaAlSi could be indexed by a hexagonal unit cell with cell parameter, a = 

4.2989(6) Å c = 5.1438(7) Å and a = 4.1838(7) Å c = 4.396(1) Å, respectively. Obtained cell 

parameters for CaAlSi, SrAlSi and BaAlSi are shown in Table 4. These cell parameter values 

are in line with previously reported BaAlSi, a = 4.290 Å and c = 5.140 Å, and CaAlSi, a = 

4.189 Å and c = 4.400Å, from Lorenz et al [8]. The unit cell expands when going from 

CaAlSi to BaAlSi following the increase in alkaline earth metal radius. 

Table 4: Cell parameters of AeAlSi (Ae = Ca, Sr and Ba) 

Compound a(Å) c(Å) V(Å 3 ) 

CaAlSi 4.1838(7) 4.396(1) 66.64(2) 

SrAlSi [7] 4.2767(7) 4.7442(9) 73.75(2) 

BaAlSi 4.2989(6) 5.1438(7) 82.32(2) 

16

3.1.2 BaAl2-xSix (0.4 < x < 1.6) alloys 

A wide stability range was reported for CaAl2-xSix (0.6 < x < 1.2) and SrAl2-xSix (0.6 < 

x < 1.2) [8]. Also, recently, Yamanaka et al. revealed that the Al/Si ratio in BaAlSi is also 

changeable according to BaAl2-xSix (1 < x < 1.5) [26]. Additionally, they reported BaAlSi to 

be a superconductor with Tc = 2.8 K. Previous investigation had not observed any 

superconductivity above 2 K. In their report, however, the superconductivity was only found 

for x > 1. 

We also synthesized BaAl2-xSix (0.4 < x < 1.6) alloys as precursors to BaAl2-xSixH2-x. 

The samples with varying stoichometric ratio were investigated by powder XRD. For BaAlSi 

(x = 1), a single phased sample was obtained. However, the BaAl2-xSix (0.4 < x < 1.6) samples 

deviating from x = 1 exhibited secondary phases in the diffraction patterns, that were not 

eliminated by the heat treatment. For x-value above 1, small peaks of BaAl2Si2O8 [27], BaSi2 

(the orthorhombic phase) [28] were observed in addition to peaks from so far unidentified 

phases. For x-values below 1, also some so far unidentified peaks were observed. The amount 

was, however, not enough to allow for an identification of the impurity phases. As x 

increasingly deviates from 1, the intensity of the impurity peaks increased. We expect this 

introduced small systematic error in the targeted BaAl2-xSix compositions, especially at larger 

deviation from x = 1. This indicates that the BaAlSi where x = 1 is a preferred phase, and it 

suggests that Si and Al atoms are distributed in ordered way in BaAlSi. Figure 3 shows the 

cell parameter change of BaAl2-xSix as function of x. It could be confirmed that a solid 

solution is formed between 0.4 < x < 1.6. The a-axis is linearly decreasing with increasing x. 

This is reasonable as bigger Al atoms are replaced by smaller Si atoms. The c-axis is rather 

independent on x. The cell parameters and volumes for each composition are shown in Table 

5. 

17

Figure 3: Cell parameters of BaAl2-xSix as a function of x 

Table 5: Cell parameters of BaAl2-xSix at each x-value 

Composition a-axis[Å] c-axis[Å] V[Å 3 ] 

BaAl1.6Si0.4 4.416(2) 5.142(2) 86.82(7) 









18

3.2 Hydrides 

3.2.1 BaAlSiH and CaAlSiH 

BaAlSiH was obtained by hydrogenation of BaAlSi at temperatures above 600 ℃, as 

described in Table 6. BaAlSiH was found to have the same structure type as the polyanionic 

semiconductor SrAlSiH, and similar to the hydrogenation of SrAlSi. The BaAlSi ingot retains 

the alloy ingot shape after hydrogenation. This is unusual. A typical solid state hydrogenation 

reaction usually yields a finely powdered product, due to the strains induced into the lattice 

when hydrogen enters it. During hydrogenation the color changed from metallic silver to dark 

gray. The cell parameters of the hydride were compared with that of the starting BaAlSi alloy. 

Significant cell parameter change was observed during hydrogenation above 600 o C, 

indicating that the hydride formed above this temperature (cf. Table 6). 

Table 6: BaAlSi + H2 (H2 pressure 70 bar) 

Temperature[ o C] Time[hours] XRD phase analysis. 

200 48 No hydride was present (Only BaAlSi was observed). 

500 48 No hydride was present (Only BaAlSi was observed). 

600 48 Formation of BaAlSiH. Small peaks of an unknown phase 

were observed. 

700 48 Formation of BaAlSiH. 

Figure 4 shows the powder x-ray diffraction pattern of a BaAlSiH sample obtained by 

direct hydrogenation of a BaAlSi at a hydrogen pressure of 70 bar at 700 o C. Using 

TREOR97 and PIRUM, an assumed hexagonal unit cell could be refined with (a = 4.3146(5) 

Å c = 5.2050(7) Å). 

19

Intensity (a.u.) 

♦ 

♦ 

Si 

♦ 

♦ 

15 30 45 60 

20 

♦ 

♦ 

2 theta (degree) 

♦ 

♦: BaAlSiH 

Si: Internal standard 

Figure 4: XRD patten of BaAlSiH 

The neutron diffraction data from a similarly synthesized deutride BaAlSiD (cf. Table 

7) was refined with FULLPROF. Starting parameters were taken from the corresponding 

SrAlSiD [7]. The refinement confirmed that BaAlSiD crystallizes in the same trigonal 

structure type in P3m1 (RB = 5.33 % and RF = 4.44 %) as SrAlSiD. Experimental and 

calculated neutron diffraction pattern from the BaAlSiD is shown in Figure 5, and the 

structure of BaAlSiD is shown in Figure 6. The atomic coordinates are given in Table 8. The 

BaSiAlD is formed by slightly puckered trigonal zintl anion layers [AlSiD] 2- , where Al and Si 

atoms are distributed alternating in ordered way. Sandwiched between these layers are the 

alkaline earth metal atoms of Ba. One D atom is bonded to each Al atoms parallel to the caxis. 

This intercalation of D atom between anionic layers elongates the c-axis. 

Table 7: BaAlSi + D2 (D2 pressure 70 bar) 


700 96 Formation of BaAlSiD. Small peaks of an unknown phase 

were also observed. 

Si 

♦ 

♦ 

Si 

♦ 

♦

Bragg R-factor: 5.43 RF-factor: 4.70 

Some impurity lines are observed in the graph. 

Figure 5: The Rietveld refinement with neutron diffraction data of BaAlSiD 

Figure 6: Structure of BaAlSiD 

21

Table 8: Atomic parameters of BaAlSiD 

Atom Site x y z Biso Occupancy 

Ba a 0 0 0 0.6(1) 1 

Al c 2/3 1/3 0.535(2) 1.1(2) 1 

Si b 1/3 2/3 0.450(2) 0.7(2) 1 

D c 2/3 1/3 0.868(2) 1.4(1) 1 

The cell parameters for BaAlSi and BaAlSiD were determined to a = 4.2989(6) Å c = 

5.1438(7) Å and a = 4.3087(6) Å c = 5.203(1) Å, respectively. The structural difference 

between BaAlSi alloy and BaAlSiD is small. This helps BaAlSiH to retain the ingot shape 

after hydrogenation, instead of being shattered into a powder. 

22

To synthesize CaAlSiH was very difficult. In 2001, H. Tanaka et al. tried to 

hydrogenate CaAlSi to obtain CaAlSiH [29]. However they interpreted their results as a 

decomposition reaction according to 2CaAlSi + H2 → CaH2 + CaSi + 2Al + Si. 

To obtain CaAlSiH, careful control of the hydrogenation condition was required. We 

tried different reaction conditions (cf. Table 9) while recording XRD patterns. After each 

reaction, they were compared with XRD pattern of CaAlSi. At a hydrogen pressure of 90 bar 

with increasing temperature, we found that CaAlSi started to form CaAl2Si2 [25] at 270 ℃. At 

this temperature, no hydride was observed. At temperatures above 310 ℃, new peaks of a 

novel CaAlSiH appeared. However, formation of CaAl2Si2 as a second phase was inevitable. 

To maximize the yield of CaAlSiH with regard to CaAl2Si2, the hydrogenation was performed 

at 500 ℃ for 45 minutes. It took some trial and error to optimize the reaction conditions. 

Table 9: CaAlSi + H2 (H2 pressure 90 bar) 


250 48 No hydride was present (Only CaAlSi was observed). 

260 48 No hydride was present (Only CaAlSi was observed). 

270 48 No hydride was present. Formation of CaAl2Si2. 

310 48 Formation of CaAlSiH but with CaAl2Si2 [25]. 

450 48 Formation of CaAlSiH but with CaAl2Si2 [25]. 

500 0.833 Almost single phased CaAlSiH was formed. 

The XRD pattern of CaAlSiH is shown in Figure 7. The trigonal unit cell parameters 

of CaAlSiH were determined to a = 4.130(1) c = 4.761(2). The trigonal unit cell parameters of 

CaAl2Si2 were reported to a=4.13(1) c=7.15(2) [25]. The similarities between the cell 

parameters leads to peak overlap problems in the XRD pattern, making it difficult to find and 

refine the CaAlSiH phase. 

23

Intensity (a.u) 

♦ 

+ 

Si 

♦ 

+ 

♦ 

20 30 40 50 60 

24 

+ 

♦ 

2 theta (degree) 

Figure 7: XRD pattern of CaAlSiH 

♦ 

Si 

♦ CaAlSiH 

+ CaAl Si 

2 2 

Unknown 

+ 

+ 

♦ 

♦ 

+ 

Si 

+ 

Upon hydrogenation CaAlSi undergoes a unit cell volume expansion of 3.67 [Å/Hatom] 

compared to 2.45 [Å/H-atom] for SrAlSiH and 1.65 [Å/H-atom] for BaAlSiH. The aaxis 

shrinks from 4.1838(7) Å to 4.130(1) Å, and c-axis expands from 4.396(1) to 4.761(2). 

Such a large expansion of cell volume [Å/H-atom] is common for metal hydrides and usually 

causes the hydride alloy to disintegrate into a fine powder upon hydrogenation. But despite 

this larger volume expansion and the fact that the a-axis shrinks and c-axis expands, 

significantly, the shape is retained also for this system during hydrogenationsdehydrogenations. 

The strain introduce, however, problems with the crystallinity, mainly in the cdirection, 

resulting in anisotropic broadening of the diffraction peaks as will be further 

discussed in below.

Similarly deuterated CaAlSiD (cf. Table 10) was measured by neutron diffraction. The 

diffraction pattern was refined by FULLPROF. Starting parameter was taken from the relaxed 

equilibrium structure. The refinement confirmed that also CaAlSiD crystallizes in the same 

trigonal structure type in P3m1 (RB = 4.99 % and RF = 2.51 %). The observed and calculated 

diffraction pattern is shown in Figure 8. The atomic positions of each element in CaAlSiD are 

shown in Table 11. Problem with the crystallinity is manifested by anisotropic broadening of 

the peaks. The (00l) reflections are twice as broad as the (hk0) reflections and (hkl) reflections 

are in between. To get a high yield of CaAlSiD, hydrogenation on CaAlSi had to be 

performed within a short time of about 50 minutes at 500 ℃. This short time of hydrogenation 

probably was insufficient to get a well defined micro structure of CaAlSiD. Intercalation of 

hydrogen along c-axis is probably the reason for the broadening of 00l reflections. If we try to 

hydrogenate CaAlSi at higher temperature and longer time to get better crystallinity, CaAlSi 

start to decompose to CaAl2Si2 and CaD2 reducing the amount of CaAlSiD. Therefore, heat 

treatment and longer time of hydrogenation could not be used to solve the problem. 

During the structural refinement, this crystallinity problem was taken into account by 

refining the crystal size parameter which is implemented in FULLPROF. This procedure was 

not needed for the refinement of the corresponding compound SrAlSiD and BaAlSiD which 

had been hydrogenated for a longer 4 days period. The refinement of the crystal size 

parameters improved the peak fitting, although we still detected a deviation in the (110) 

reflection. 

Table 10: CaAlSi + D2 (D2 pressure 70 bar) 


500 0.833 CaAlSiD was formed. Small peaks of CaAl2Si2 were 

observed. 

25

Bragg R-factor: 4.99 % RF-factor: 2.51 % 

Some impurity lines are observed in the graph. 

Figure 8: The Rietveld refinement with neutron diffraction data of CaAlSiD 

Table 11: Atomic positions of CaAlSiD 

Atom Site x y z Biso Occupancy 

Ca a 0 0 0 1.1(2) 1 

Al c 2/3 1/3 0.548(5) 0.4(1) 1 

Si b 1/3 2/3 0.462(2) 0.4(1) 1 

D c 2/3 1/3 0.915(5) 3.6(2) 1 

26

The structure of CaAlSiD is shown in Figure 9. As in the case of previous SrAlSiD 

and BaAlSiD, Al and Si atoms are alternately distributed in the trigonal Zintl anion. One 

hydrogen atom is attached to each Al atom along the c-axis. Ca atoms are, as for the precursor 

alloy, sandwiched by these Zintl anions when building the CaAlSiD structure. 

Figure 9: Structure of CaAlSiD 

Comparing the structures of CaAlSiD, SrAlSiD, and BaAlSiD suggests some 

interesting things. Going from calcium to barium, the cell volumes become larger as the cell 

parameters become longer reflecting the ionic side difference between Ca 2+ , Sr 2+ and Ba 2+ . 

Cell parameters and cell volumes of the AeAlSiD and AeAlSiH (Ae=Ca, Sr, and Ba) are 

shown in Table 12. 

Table 12: Cell parameters of AeAlSiD and AeAlSiH (Ae = Ca, Sr and Ba) 

a[Å] c[Å] V[Å 3 ] 

CaAlSiD 4.133(1) 4.761(2) 70.43(1) 

CaAlSiH 4.130(1) 4.761(2) 70.33(4) 

SrAlSiD[7] 4.2113(3) 4.9518(5) 76.05(1) 

SrAlSiH[7] 4.2139(3) 4.9550(6) 76.20(1) 

BaAlSiD 4.3087(6) 5.203(1) 83.65(1) 

BaAlSiH 4.3146(5) 5.2050(7) 83.91(2) 

27

The interatomic distances in CaAlSiD, SrAlSiD [7] and BaAlSiD are shown in Table 

13, 14 and 15, respectively. All the Al-D distances in AeAlSiD are very similar. The Al-D 

distance for CaAlSiD, SrAlSiD and BaAlSiD are 1.75(3), 1.77(1) Å and 1.73(2) Å 

respectively. The Ae-D distances in AeAlSiD are similar to the Ae-D distances in AeD2. The 

Ca-D distance is 2.420(4) Å in CaAlSiD and 2.17(7)-2.8(1) Å in CaD2 [30]. The Sr-D 

distance is 2.476(1) Å in SrAlSiD and 2.36-2.79 Å in SrD2 [30]. The Ba-D distance is 2.581(3) 

Å in BaAlSiD and 2.57-2.98 Å in BaD2 [31]. 

These support a picture that D atoms are covalently bonded to Al atoms. 

Simultaneously, D atoms are also ionically interacting with the Ae atoms explaining the 

increasing Ae-D distance from calcium to barium. 

Because of problem with poor crystallinity along the c-axis in CaAlSiD, the atomic 

positions may include some systematic errors in the z-coordinates. However, the obtained 

values, as given in table 13-15, are well in line with the relaxed values from the related total 

energy calculations. 

28

Table 13: Selected interatomic distances (Å) and angles (˚) in CaAlSiD. The values from the 

total energy calculation are shown in brackets. 

Ca-3D 2.420(4) [2.433] Al-D 1.75(3) [1.746] 

3Si 3.245(6) [3.155] 3Si 2.421(4) [2.451] 

3Al 3.21(2) [3.237] 3Ca 3.21(2) [3.237] 

3Si 3.501(7) [3.610] D 3.01(3) [3.011] 

3Al 3.54(2) [3.519] 3Ca 3.54(2) [3.519] 

6Al 4.133(0) [4.147] 

Si-3Al 2.421(4) [2.451] 

3D 3.22(2) [3.299] 

3Ca 3.245(6) [3.155] 

3Ca 3.501(7) [3.610] 

2Si 4.76(1) [4.757] 

D-Al 1.75(3) [1.753] 

3Ca 2.420(4) [2.497] 3Al-Si-Al 117.20(0) [115.57] 

Al 3.01(3) [3.212] 3Si-Al-Si 117.20(0) [115.57] 

6D 4.133(0) [4.147] 3Si-Al-D 99.74(0) [102.33] 

Table 14: Selected interatomic distances (Å) and angles (˚) in SrAlSiD [7]. The values from 

the total energy calculation are shown in brackets. 

Sr-3D 2.476(1) [2.497] Al-D 1.77(1) [1.753] 

3Si 3.229(7) [3.295] 3Si 2.502(3) [2.484] 

3Al 3.305(7) [3.342] 3Sr 3.305(7) [3.342] 

3Si 3.729(8) [3.677] D 3.18(1) [3.212] 

3Al 3.644(7) [3.625] 3Sr 3.644(7) [3.625] 

6Al 4.211(0) [4.226] 

Si-3Al 2.502(3) [2.484] 

3D 3.386 (8) [3.298] 

3Sr 3.229(7) [3.295] 

3Sr 3.729(8) [3.677] 

2Si 4.95(1) [4.965] 

D-Al 1.77(1) [1.753] 

3Sr 2.476(1) [2.497] 3Al-Si-Al 114.63(0) [116.55] 

Al 3.18(1) [3.212] 3Si-Al-Si 114.63(0) [116.55] 

6D 4.211(0) [4.226] 3Si-Al-D 103.62(0) [100.83] 

29

Table 15: Selected interatomic distances (Å) and angles (˚) in BaAlSiD. The values from the 

total energy calculation are shown in brackets. 

Ba-3D 2.581(3) [2.595] Al-D 1.73(2) [1.741] 

3Si 3.416(7) [3.469] 3Si 2.527(3) [2.538] 

3Al 3.470(7) [3.483] 3Ba 3.470(7) [3.483] 

3Si 3.792(8) [3.778] D 3.47(2) [3.488] 

3Al 3.733(8) [3.763] 3Ba 3.733(8) [3.763] 

6Al 4.309(0) [4.338] 

Si-3Al 2.527(3) [2.538] 

3D 3.30(1) [3.300] 

3Ba 3.416(7) [3.469] 

3Ba 3.792(8) [3.778] 

2Si 5.20(2) [5.229] 

D-Al 1.73(1) [1.741] 

3Ba 2.581(3) [2.595] 3Al-Si-Al 117.0(0) [117.47] 

Al 3.47(1) [3.488] 3Si-Al-Si 117.0(0) [117.47] 

6D 4.309(1) [4.338] 3Si-Al-D 100.08(1) [99.25] 

Upon heating under vacuum, CaAlSiH and BaAlSiH are observed to start to release 

hydrogen at around 420 ℃ and 500 ℃, according to the reactions 2CaAlSiH ↔ CaAlSi + H2 

and 2BaAlSiH ↔ 2BaAlSi + H2. Both reactions are reversible and the CaAlSi and BaAlSi 

alloys were recovered after dehydrogenation. The decomposition temperature of CaAlSiH is 

lower than that of BaAlSiH and SrAlSiH (Decomposition of BaAlSiD and SrAlSiD was 

observed at similar temperatures). The decomposition temperature of BaH2 and SrH2 is 

675 ℃ and 675 ℃ respectively, and CaH2 decompose at lower temperature 600 ℃ [32]. This 

means that the AeAlSiH follow a similar trend to AeH2, confirming the importance of Ae-D 

interaction in AeAlSiD. 

30

As shown above, the decomposition temperatures of AeAlSiD are high and this 

indicate that AeAlSiD are fairly stable hydrides. By comparing the structures of SrAl2D2 [5] 

with SrAlSiD [7], we can understand this. The structures of these compounds are shown in 

Figure 10. D atoms are covalently bonded to Al atoms in both structures. The Al-D distance is 

1.77(1) Å in SrAlSiD compared to 1.706(4) Å in SrAl2D2. Even if SrAlSiD has a longer Al-D 

distance, D atoms are apparently bonded more strongly in SrAlSiD and this was unexpected. 

The decomposition temperature (of atmospheric pressure) is 600 ℃ for SrAlSiH and 300 ℃ 

for SrAl2H2. But on the other hand, the Sr-D bond is significantly shorter in SrAlSiD 

compared to in SrAl2D2. The Sr-D distance in SrAlSiD and SrAl2D2 is 2.476(1) Å and 

2.653(1) Å, respectively. Thus a stronger ionic interaction between Sr and D may bind the D 

atom more firmly in the structure of SrAlSiD, and this compensates for a weaker covalent Al- 

D bond [7]. Also, the relatively short D-D distance in SrAl2D2 is to some extent probably also 

weakening the metal-hydrogen bond strength. The D-D distance is 2.770(1) Å and 4.211(0) Å 

in SrAl2D2 and SrAlSiD, respectively. 

Figure 10: Structural comparison between SrAl2D2 [5] and SrAlSiH [7] 

31

3.2.2 BaAl2-xSixH2-x (0 < x < 2) 

A series of BaAl2-xSix alloys were hydrogenated as described in tables 16-19 below, 

where also the results from subsequent powder x-ray diffraction phase analysis of the reaction 

products are given. A careful control of the hydrogenation conditions was required. Optimum 

hydrogenation reaction temperatures of BaAl2-xSixH2-x are dependent on x. To hydrogenate 

BaAlSi (x = 1), 600 ℃ and 70 bars hydrogen pressure is required. On the other hand, 

BaAl0.4Si1.6H0.4 and BaAl1.6Si0.4H1.6 could be hydrogenated at lower temperature, 300 ℃ 

using 70 bars of hydrogen pressures. At higher temperature, aluminium rich BaAl2-xSixH2-x (0 

< x < 1) tend to decompose to composition close to BaAlSiH (x = 1), BaAl4 [33] and so far 

unknown phases. Also, at higher temperature, silicon rich BaAl2-xSixH2-x (1 < x < 2) tends to 

decompose to a composition close to BaAlSiH(x = 1) and BaSi2 (orthorhombic) [28] and so 

far unknown phases. This indicates that BaAlSiH (x = 1) is a specially stable phase in the 

BaAl2-xSixH2-x system. This was unexpected because we expected that the stability of BaAl2xSixH2-x 

would become higher with increasing Si amount, similar to the homologous system 

SrAl2-xSixH2-x. The decomposition temperature of SrAlSiH is higher than that of SrAl2H2. 

Also, the SrAlSiH (x = 1) is more stable than SrAl2H2 (x = 0) to air and moisture. 

32

Table 16: BaAl1.6Si0.4 + H2 (H2 pressure 70 bar) 


230 24 No hydride was formed. 

300 48 Formation of an assumed BaAl1.6Si0.4H1.6. So far unknown 

phases were also observed. 

700 48 Formation of BaAlSiH instead of BaAl1.6Si0.4H1.6. 

Additionally, strong peaks of BaAl4 and so far unknown 




300 48 Formation of an assumed BaAl1.4Si0.6H1.4. Additionally, so 

far unknown phases were also observed. 

550 3 Formation of an assumed BaAl1.4Si0.6H1.4. Additionally, 

BaAl4, BaAlSiH and so far unknown phases were also 

observed. 


BaAl4 and so far unknown phases were also observed. 


Additionally strong peaks of BaAl4 and so far unknown 





550 3 Formation of an assumed BaAl0.6Si1.4H0.6, but additionally 

BaSi2 and so far unknown phases were also observed. 

600 12 Formation of a phase close to BaAl0.6Si1.4H0.6. 

Additionally BaSi2 and unknown phases were also 

observed. 


Additionally strong peaks of BaSi2 and so far unknown 


33




small peaks of BaSi2 and unknown phases were observed. 

500 48 Formation of a phase closes to BaAl0.6Si1.4H0.6 instead 

BaAl0.4Al1.6H0.4. Strong peaks of BaSi2 and unknown 


700 48 Formation of BaAlSiH instead of BaAl0.4Si1.6H0.4. Strong 

peaks of BaSi2 and so far unknown phases were also 

observed. 

34

The color of the all BaAl2-xSixH2-x (0.4 < x < 1.6) hydrides was dark gray, in contrast 

to the metallic silver of the BaAl2-xSix (0.4 < x < 1.6) alloys. 

The obtained samples were measured by powder x-ray diffraction and the cell 

parameters of each composition are shown in Table 20. Figure 11 shows how the cell 

parameters of BaAl2-xSixH2-x nicely extrapolates with x, between the cell parameters of 

BaAl2H2 (x = 0) and BaSi2 (x = 2) [34]. The structures of BaAl2H2 and BaSi2 are very similar 

to that of BaAlSiH as shown in Figure 13. Also, all of them have the same electron count. The 

BaAl2H2 has so far not been experimentally obtained. The structural parameters were obtained 

from our total energy calculation. Also, in the figure 11, the cell parameters of the 

corresponding SrAl2-xSixH2-x system (SrAl2H2 [5] and SrAlSiH [7]) are plotted, showing a 

similar trend to the Ba compound. Similar to the alloys, the a-axis length decrease with 

increasing x in BaAl2-xSixH2-x as the bigger Al atoms are replaced by smaller Si atoms. The caxis 

increases with increasing x. This trend is different from the alloy, where the c-axis length 

is rather independent of x. By comparing the structures of BaAl2H2 (x = 0), BaAlSiH (x = 1) 

and BaSi2 (x = 2) [34], it can be explained (cf. Figure 12). Three main reasons are responsible 

for the c-axis elongation. Firstly, a [Si] - lone pair needs more space compared to an [Al-H] - 

entity. Secondly, the Al-H bond length in BaAlSiH is apparently longer than in BaAl2H2. The 

computationally obtained Al-H bond length is 1.71 Å in BaAl2H2 compared to 1.74 Å in 

BaAlSiH. The experimentally obtained Al-H bond length is 1.73(2) Å in BaAlSiH. In other 

words, the Al-H bond is elongated when x is increased. The same observation could be 

obtained for the corresponding system SrAl2-xSixH. Experimentally obtained Al-H bond 

length is 1.706(4) Å for SrAl2H2 and 1.77(1) Å for SrAlSiH. Thirdly, the degree of the 

hexagonal network puckering is increased with increasing x-values. BaSi2 [34] has a more 

puckered trigonal network compared to BaAlSiH and BaAl2H2. This more puckered trigonal 

network at higher x-value is also observed in the SrAl2-xSixH2-x system, where SrAlSiH [7] 

has a more puckered trigonal network compared to that of SrAl2H2 [5].Taken together, the 

observed c-axis elongation at higher x-value of BaAl2-xSixH2-x is thus very reasonable. 

35

Figure 11: Cell parameters of BaAl2-xSixH2-x as a function x. (■: Computationally obtained 

BaAl2H2, ★: SrAl2H2 [5], ▲: SrAlSiH [7], ●: BaSi2 [34]) 

BaAl2H2 BaAlSiH BaSi2 [34] 

Figure 12: Structures of BaAl2-xSixH2-x (x=0, 1 and 2) 

36

Table 20: Cell parameters of BaAl2-xSixH2-x. The cell parameters before hydrogenation are 

shown in brackets. 


BaAl2H2 (calculated) 4.6524 4.9768 93.29 

BaAl1.6Si0.4H1.6 4.451(3) [4.416(2)] 5.125(4) [5.141(2)] 87.9(1) [86.82(7)] 


BaAlSiH 4.3146(5) [4.2989(6)] 5.2050(7) [5.1438(7)] 83.97(4) [82.32(2)] 

BaAl0.6Si1.4H 4.283(3) [4.2342(4)] 5.239(5) [5.1345(6)] 83.2(1) [79.72(1)] 


BaSi2 [35] 4.047(3) 5.330(5) 75.6(1) 

We could observe an interesting stability problem when hydriding the BaAl2-xSix 

alloys with compositions deviating from x = 1. If an aluminium rich BaAl2-xSix (0 < x < 1) or 

a silicon rich BaAl2-xSix (1 < x < 2) are hydrogenated, at the lowest temperatures when a 

reaction is observed to proceed, (as given in Tables from 16 to 19), then a hydride with the 

same metal atom composition will be produced. But if the temperature is increased, the 

hydride composition will start to move towards the stable BaAlSiH (x = 1). 

Also, from the point of synthesis condition, comparison of BaSi2 [34] and BaAl2xSixH2-x 

system is very interesting. The BaSi2 has three phases, orthorhombic, cubic and 

trigonal structure [28] [34] [35]. The orthorhombic phase is the stable phase and the other 

phases are metastable high pressure phases. The metastable cubic and trigonal BaSi2 are 

obtained when the orthorhombic phase is subjected with the high-pressure and high- 

temperature conditions (4GPa, 600-800 ℃ for the cubic phase and 4GPa and 1000 ℃ for the 

trigonal phase) [34] [35]. In other words, the trigonal BaSi2 is difficult to form. This is in 

contrast to BaAl2-xSixH2-x. Because, by substituting Si - in BaSi2 with [Al-H] - , we could obtain 

BaAl0.4Si1.6H0.4 which is structurally and compositionally very close to trigonal BaSi2 at 

moderate conditions (300 ℃ at 70 bar hydrogen of pressure). 

37

3.3 Computational results 

3.3.1 Computationally relaxed structural parameters 

Experimentally obtained structural parameters of AeAlSi (Ae = Ca and Ba) and 

computationally relaxed structural parameters of AeAlSiH are compared. Structural 

parameters of CaAlSiH and BaAlSiH from both calculation and experiment are shown in 

Table 21 and Table 22, respectively. The structural parameters for BaAlSiH agreed well with 

each other. Also for CaAlSiH, where we had problem with poor crystallinity, the 

experimental and calculated values are in good agreement. 

Table 21: Experimentally obtained and computationally relaxed structural parameters for 

CaAlSiH 

CaAlSIH (calculated) CaAlSiD (experimental) 

V[Å 3 ] 71.72 70.43(1) 

a[Å] 4.1363 4.133(1) 

c[Å] 4.7673 4.761(2) 

c/a 1.1526 1.152 

Ca position 0,0,0 0,0,0 

Si position 1/3, 2/3, 0.446 1/3, 2/3, 0.462(2) 

Al position 2/3, 1/3, 0.541 2/3, 1/3, 0.548(5) 

H position 2/3, 1/3, 0.893 2/3, 1/3, 0.915(5) 

Table 22: Experimentally obtained and computationally relaxed structural parameters for 

BaAlSiH 

BaAlSiH (calculated) BaAlSiD (experimental) 

V[Å 3 ] 85.22 82.25(1) 

a[Å] 4.3381 4.3087(6) 

c[Å] 5.2289 5.203(1) 

c/a 1.2053 1.208 

Ba position 0,0,0 0,0,0 

Si position 1/3, 2/3, 0.458 1/3, 2/3, 0.450(2) 

Al position 2/3, 1/3, 0.537 2/3, 1/3, 0.535(2) 

H position 2/3, 1/3, 0.869 2/3, 1/3, 0.868(2) 

38

3.3.2 DOS of AeAlSiD 

We calculated the density of states (DOS) for the CaAlSiH, SrAlSiH and BaAlSiH as 

compared in Figure 13. All hydrides have a narrow band gap. From this, we can expect that 

CaAlSiH and BaAlSiH compounds are semi-conductors, as in the case of previously reported 

corresponding SrAlSiH system. The band gap is 0.71 eV and 0.42 eV in BaAlSiH and 

CaAlSiH respectively, compared to 0.65 in SrAlSiH [7]. The DOS of AeAlSiD (Ae = Ca, Sr 

and Ba) is characterized by a pronounced singularity just below Fermi level. 

DOS 

10 

8 

6 

4 

2 

DOS 

8 

6 

4 

2 

0 

-1.0 -0.5 0 

E-E [eV] 

F 

0.5 1.0 

0 

-10 -8 -6 -4 -2 0 2 

E-E F [eV] 

39 

Ca Al Si H 

Sr Al Si H 

Ba Al Si H 

Figure 13: DOS comparison between AeAlSiH (Ae = Ca, Sr and Ba)

3.3.3 DOS of BaAl2-xSixH2-x 

DOS of BaAl2H2 (x = 0), BaAlSiH (x = 1) and BaSi2 (x = 2) [34] are compared in 

Figure 15. Trigonal BaSi2 (x = 2) is reported to have metallic conductivity [36] which could 

also be confirmed by our DOS calculation. The DOS calculation for BaAl2H2 also indicated it 

to have metallic conductivity. This is interesting, because by substituting Si - with [AlH] - in 

BaAl2-xSixH2-x, the electric property of BaAl2-xSixH2-x can be changed. From this, we expect 

that at some x-value between 2 and 1, a transformation from a metallic conductor to a 

semiconductor will occur and again at some x value between 1 and 0, the transition will be 

reversed. In other word, the band gap of BaAl2-xSixH2-x may be tuneable by substituting Si - 

with [AlH] - . 

Moreover, to investigate a possible superconductivity of BaAl2-xSixH2-x would be very 

interesting. The DOS of BaAlSiH (x = 1) is characterized by a pronounced singularity below 

the Fermi level. If we could move the singularity onto the Fermi level, we may able to 

increase the Tc. This could be attempted by tuning the Al/Si ratio in the compound, but also 

by varying the hydrogen content. BaAlSi, which is precursor material of BaAlSiH, was 

reported to exhibit superconductivity [37]. Thus to investigate the effect of hydrogen doping 

to BaAlSi would be interesting. 

Also, CaAlSi the precursor material of CaAlSiH, shows superconductivity below 7.7 

K, as does SrAlSi with a Tc of 5.1 K [8]. The trigonal BaSi2 has a Tc of 6.8 K and the related 

CaSi2 has an unusual high Tc of 14 K. Thus to investigate superconductivity of all these 

systems, while varying the Al/Si ratio as well as the hydrogen content, would be interesting. 

40

DOS 

10 

8 

6 

4 

2 

0 

-4 -2 0 2 

E-E F [eV] 

41 

BaSi 2 


BaAl 2 H 2 

Figure 14: DOS comparison of BaAl2-xSixH2-x (x =0, 1 and 2)

4. Summary 

4.1 Alloys of CaAlSi and BaAl2-xSix 

As a precursor material of CaAlSiH and BaAlSiH, we synthesized CaAlSi and 

BaAlSi alloys by arcmelting the stoichiometric ratio of elements. For these compounds, the 

cell parameters we obtained from powder XRD, correspond well to previously reported values. 

Single XRD experiment was performed for BaAlSi and the structural parameters also agreed 

with previously reported values. 

For the barium system, we could vary the Al/Si composition according to BaAl2xSix 

(0.4

4.3 Zintl phase hydride, BaAl2-xSixH2-x (0.4 < x < 1.6) 

A series of BaAl2-xSix (0.4 < x < 1.6) were hydrogenated to synthesize the 

corresponding BaAl2-xSixH2-x. 

Careful control of hydrogenation condition was required to synthesize BaAl2xSixH2-x, 

as high temperatures tended to shift the composition toward the x = 1 phase of 

BaAlSiH. 

The a-axis length of BaAl2-xSixH2-x decreased with increasing x-value. This is due 

to Al atoms being replaced by smaller Si with increasing x. On the other hand, the length of caxis 

is increased with increasing x. This may be explained by an increasing volume as the 

[AlH] - units are replaced by larger Si - lone pairs, a more puckered Zintl anion and longer Al- 

H bond length at higher x. 

Our computations show that BaAlSiH is expected to be a semiconductor with a 

narrow band gap. BaSi2 is reported to have metallic conductivity. Also from our computations, 

BaAl2H2 is expected to have metallic conductivity. Therefore, we expect transitions from 

semiconductors to conductors to occur at some x-values of BaAl2-xSixH2-x. In other words, the 

band gap of BaAl2-xSixH2-x may be tuneable by changing x. 

The DOS of AeAlSiH is characterized by a pronounced singularity below the 

Fermi level. Especially for BaAlSiH, the singularity is very high and sharp. Therefore, it 

might be rewarding to investigate the low temperatures conductivity for the BaAl2-xSixH2x(0.4

6. Acknowledgement 

I would like to express my deepest gratitude to my supervisor Prof. Dag Noréus for giving me 

a wonderful experience here at Stockholm University. Through his help, I could come here 

for studying and learning many things. I will never forget his bottomless kindness and 

supports that he gave me. 

I would like to thank Mr. Lars Göthe for measuring powder XRD of my samples. He also 

shared office with me and created a nice atmosphere in the room. I really appreciated that. 

I would like to thank Dr. Toyoto Sato (He received his Ph. D. in our laboratory in 2006) for 

help with powder x-ray analysis. Also, I really appreciate his great support, helpful advice and 

fruitful discussion through this work. 

I would like to thank Mr. Thomas Björling for his collaboration, advice, fruitful discussion 

and help throughout this work. Also, I really enjoyed sharing office with him. 

I would like to thank Mr. David Moser for his computational calculation in this thesis. Also, I 

really appreciate his helpful advice, fruitful discussion and help throughout this work. 

I would like to thank Dr. Lars Eriksson for help with single crystal XRD measurements of 

BaAlSi. I also appreciate your supportive and interesting lectures. 

I would like to thank Kjell Jansson for help with SEM, FTIR and EDX measurements of my 

samples. I also appreciate you for giving me a lot of help. 

I would like to thank Dr Bjorn Hauback, IFE, Norway, for help with collecting the powder 

neutron data of BaAlSiD and CaAlSiD. I also appreciate your hospitality during my stay in 

Norway. 

I would like to acknowledge Juan Rodriguez-Carvajal, Diffraction group, Institut Laue- 

Langevin for his helpful advice with the how anisotropic broadening of the diffraction peaks 

in CaAlSiD could be refined by crystal shape parameters using Fullprof. 

I would like to acknowledge Ms Rie Takagi for help with single XRD structural investigation 

of BaAlSi. I also appreciate your kindness and support during my stay in Stockholm. 

I would like to thank Dr. Karim Kadir for help with structural investigation of CaAlSiD. 

I am very grateful to Mr. Keiichi Miyasaka for his great support and helpful advice during my 

stay in here. I really appreciate you for teaching me a lot of things. 

I would like to acknowledge Dr Daniel Fredrickson for his helpful advice and fruitful 

discussion throughout this work. I really enjoyed your lectures and learned many things from 

you. 

I would like to acknowledge Mr. Norihiro Muroyama for your great support and helpful 

advice during my stay in here. I really appreciate your kindness. 

44

I would like to acknowledge Metal hydride group: Prof. Dag Noréus, Dr. Karim Kadir, Dr. 

Toyoto Sato, Mr. Thomas Björling, Mr. David Moser and Dr. Weikang Hu (moved to Norway 

in 2005). Thanks for supporting my study and life in here. Thanks to everybody in Metal 

hydride group, I could learn a lot and enjoy my study in here. I am really happy that I could 

stayed here as a member of Metal hydride group!! 

I would like to express my gratitude to Prof. Osamu Terasaki for giving me a lot of advice 

with kindness. 

I would like to thank every friend in our department. I enjoyed studying and life in Stockholm 

thanks to every friend. All of the times we spend will be really good memory to me. I really 

appreciate your kindness and I will never forget the wonderful and meaningful time that I 

spend in here with you guys. 

I would like to thank The Scandinavia-Japan Sasakawa Foundation for financial support 

during my stay in Stockholm University. 

I would like to express my sincere gratitude to my previous supervisor Seijrau Suda, professor 

of Kogakuin University, Japan. Without his support, my studies here would not have been 

possible. I really appreciate you for giving me a great opportunity and supports for my studies 

here. 

Partial funding by the European Commission DG Research (contract SES6-2006-518271/ 

NESSHY) is also gratefully acknowledged. 

これまで、両親の支えのお陰で、長い間学生として自由に勉強をさせて頂くことが 

出来ました。勉強が楽しいと思える様になるまで、自由に勉強させて頂いた事、両 

親には大変感謝申し上げます。 

45

7. References 

1. B. Bogdanovic, M. Schwickardi, J. Alloys Compd. 1997, 253, 1. 

2. J W. Lauher, D. Dougherty and P. J. Herey, Acta Cryst. 1979. B35, 1454. 

3. D. Noréus, K. Kadir, A Reiser and B. Bogdanovic, J. Alloys Compd. 2000, 299, 101. 

4. Y. Nakamura, A. Fossdal, H. W. Brinks and B. C. Hauback, J. Alloys Compd. 2006, 

416, 274. 

5. F. Gingl, T. Vogt and E. Akiba, J. Alloys compd. 2000, 306, 127. 

6. Q. Zhang, Y. Nakamura, K. Oikawa, T. Kamiyama, E. Akiba, Inorg. Chem. 2002, 41, 

6547. 

7. T. Björling, D. Noréus, K. Jansson, M. Andersson, E. Leonova, M. Edén, U. Hålenius, 

and U. Häussermann, Angew. Chem. Int. Ed. 2005, 44, 7269. 

8. B. Lorenz, J. Lenzi, J. Cmaidalka, R.L. Meng, Y.Y. Sun, Y.Y. Xue and C.W. Chu, 

Physica C 2002, 383, 191. 

9. J. Nagamatsu, N. Nakagawa, T. Muranaka, Y. Zenitani, and J. Akimitsu, Nature 2001, 

410, 63. 

10. K. E. Johansson, T. Palm, P. -E. Werner, J. Phys. E 1980, 13, 1289. 

11. P. E. Werner, L. Eriksson, S. Salome, SCANPI, A Program for Evaluating Guinier 

Photographs, Stockholm University, Stockholm, 1980. 

12. P. E. Werner, L. Eriksson, M. Westerdahl, J. Appl. Crystallogr. 1985, 18, 367. 

13. P. E. Werner, Ark. Kemi 1969, 31, 513. 

14. STOE & Cie GmkbH. (1999). X-RED Version 1.22. Stoe & Cie GmbH, Darmstadt, 

Germary. 

15. STOE & Cie GmkbH. (2001). X-SHAPE Revision 1.06. Stoe &Cie GmbH, 

Darmstadt . Germary. 

16. Sheldrick, G.M. (1997). SHELXS97. University of Göttingen, Germany. 

17. Sheldrick, G.M. (1997). SHELXL97. University of Göttingen, Germany. 

18. Brandenburg, K. (2001). DIAMOND Release 2.1e. Crystal Impact GbR, Bonn, 

Germany. 

19. J. Rodriguez-Carvajal, FULLPROF: A Program for Rietveld Refinement and Pattern 

Matching Analysis, Abstracts of the Satellite Meeting on Powder Diffraction of the 

XV Congress of the IUCr, Toulouse, France, 1990, 127. 

46

20. a) P. E. Blöchl, Phys. Rev. B 1994, 50, 17953 b) G. Kresse, J. Joubert, Phys. Rev. B 

1999, 59, 1758. 

21. a) G. Kresse, J. Hafner, Phys. Rev. B 1993, 47, 558. b) G. Kresse, J. Furthmüller, 

Phys. Rev. B 1996, 54, 11169. 

22. J. P. Perdew, Y. Wang, Phys. Rev. B 1992, 45, 13244. 

23. H. J. Monkhorst, J. D. Pack, Phys. Rev. B 1972, 13, 5188. 

24. S. Kuroiwa, H Sagayama, T. Kakiuchi, H. Sawa, Y. Noda, and J. Akimitsu, Phys. Rev. 

B 2006, 74 , 014517. 

25. E. I. Gladyshevskii, P. I. Krypyakevich, O. I. Bodak, E. I. Hladyshevskii, Ukrains'kii 

Fizichnii Zhurnal (Ukrainian Edition) 1967, 12, 445. 

26. S. Yamanaka, T. Otsuki, T. Ide, H. Fukuoka, R. Kumashiro, T. Rachi, K. Tanigaki, F. 

Guo, K. Kobayashi, Physica C 2007, 451, 19. 

27. N. N. Roy, Nature 1965, 206, 501. 

28. H. Schäfer, K. H. Janzon, A. Weiss, Angew. Chem. Int. Ed. 1963, 2, 393. 

29. H. Tanaka, H. Thakeshita, N. Kuriyama, T. Sakai, and I. Uehara, IEA Task 12 report: 

Metal Hydrides and Carbon for hydrogen storage 2001. 

30. T. Sichla, H. Jacobs, European Journal of Solid State Inorganic Chemistry 1996, 33, 

453. 

31. W. Bronger, C. S. Chi, P. Mueller, Zeitschrift fuer Anorganische und Allgemeine 

Chemie 1987, 545, 69. 

32. W. Grochala, P. P. Edwards, Chem. Rev. 2004, 104, 1283-1315. 

33. K. R. Andress, E. Alberti, Zeitschrift fuer Metallkunde 1935, 27(6), 126. 

34. J. Evers, G. Oehlinger, A. Weiss, Angew. Chemi. (German Edition) 1977, 89, 673. 

35. J. Evers, G. Oehlinger, A. Weiss, Angew. Chem. Int. Ed. Engl. 1978,17, 538. 

36. M. Imai, T. Hirano, J. Alloys Compd. 1995, 224, 111. 

37. M. Imai, K. Hirata, T. Hirano, Physica C 1995, 245, 12. 

47

Paper I

A series of Zintl phase hydrides; BaAl2-xSixH2-x (0.4 < x < 1.6) with compositions and 

structures in between the electric conductors BaSi2 and BaAl2H2 

T. Utsumi*, T. Björling*, D. Moser*, U. Häussermann**, D. Noréus*. 

*Structural Chemistry, Stockholm University SE-10691 Stockholm, Sweden. 

** Department of Chemistry and Biochemistry, Arizona State University. 

Abstract 

By substituting Si - with (Al-H) - in trigonal BaSi2, a series of semiconducting Zintl phase 

hydrides has been made with compositions BaAl2-xSixH2-x (0.4 < x < 1.6). DFT calculations 

show that the end compositions BaSi2 and BaAl2H2 are conductors whereas the intermediate 

BaAlSiH is a semiconductor with a bandgap of 0.71 eV. The cell parameters for the trigonal 

cell vary linearly as a function of the Si - /(Al-H) - substitution in this MgB2 related structure 

type, opening up for a possibility of continuously tuneable electric properties. 

1. Introduction 

BaSi2 is an interesting compound having so far three observed structures with different 

properties; an orthorhombic phase which is stable at ambient condition and two metastable 

high pressure phases, a cubic and a trigonal [1] [2] [3] [4]. Both can be retained at ambient 

condition, if the samples are rapidly quenched from the high pressure synthesis. The cubic 

and trigonal phase are obtained by subjecting the orthorhombic phase to pressure of 4 GPa at 

temperature 600-800 o C and 1000 o C, respectively [2] [3]. Trigonal BaSi2 has electric 

conductivity and has been reported to be a superconductor with an onset temperature of 6.8 K 

[5] [6]. A similar trigonal CaSi2 phase has been reported with an unusual high Tc of 14 K [7] 

The cubic and orthorhombic phases are semiconductors [8]. The latter with a bandgap of 1.3 

eV, has been considered to be a good candidate for solar cell materials [9]. This paper will 

focus on modifications in the trigonal system. 

Recently, Björling et al reported on the Zintl phase semiconducting hydrides AeAlSiH 

(Ae = Ca, Sr and Ba). [10] [11] In BaAlSiH, every second Si - ion in BaSi2 have been 

substituted with an isoelectronic [AlH] - entity. The structure of BaAlSiH is closely related to 

the trigonal BaSi2, but the substitution modifies the electric properties. The computationally 

obtained DOS of BaAlSiH show a narrow band gap of 0.71 eV, compared to semi metallic 

trigonal BaSi2. 

BaAlSiH can be obtained by direct hydrogenation of BaAlSi. The BaAlSi, precursor 

material of BaAlSiH, itself is also interesting. Recently, Yamanaka et al. reported that BaAlSi 

is a superconductor with Tc = 2.8 K [12]. The group revealed that Al/Si ratio in BaAlSi is 

changeable according to BaAl2-xSix (1 < x < 1.5). However, superconductivity was only 

observed for x > 1. Also the corresponding CaAlSi and SrAlSi are superconductors with Tc´s 

at 7.8 K and 5.1 K, respectively [13]. 

The aim of this work was to investigate the hydrogenation of the compositionally 

different BaAl2-xSix alloys and study how structure and properties of the trigonal system 

change when varying the composition between BaAl2H2 and BaSi2 (x = 0 and 2 in BaAl2xSixH2-x). 

Attempts to extend the investigation for the Ca- and Sr-analogues have been carried 

out without satisfactory results due to the larger amounts of impurity phases in the latter 

systems. The Ba-system seemed to be more forgiving when manipulating the composition. In 

the Ba-system we thus synthesized and characterized a series of hydrides according to BaAl2xSixH2-x 

(0 < x < 2). In figure 1, the trigonal structures of BaSi2 (x = 2), BaAlSiH (x = 1) and a 

hypothetical as calculated BaAl2H2 (x = 0) are depicted. They are structurally close to each 

other. All of them have a trigonal network of a Zintl anion [Al2-xSixH2-x] - sandwiched by Ba 2+

ions. A simple electron count would suggest straight forward sp 3 hybridization for the Zintl 

anion. Non metallic BaAlSiH and electric conducting BaSi2 and BaAl2H2 indicate a more 

complex situation as the Si - (Si-lone pair) is substituted by an isoelectronic (AlH) - unit, when 

going from BaSi2 to BaAl2H2. In this paper we report how cell parameters of the precursor 

BaAl2-xSix (0.4 < x < 1.6) as well as the corresponding hydrides correlate with physical 

properties during this substitution. 

BaAl2H2 (x = 0) BaAlSiH (x = 1) BaSi2 (x = 2) 

Figure 1: Structures of BaAl2-xSixH2-x (x=0, 1 and 2) 

2. Experimental details 

2.1 Synthesis 

All sample handling were carried out under argon atmosphere. Commercially pure Si 

powder, Al powder, Ba ingot were delivered from MERCK, Alfa Aesar and Aldrich 

respectively. The alloys of BaAl2-xSix (0.4 < x < 1.6) were synthesized by arcmelting 

stoichiometric ratios of the elements in a MAM-1 (Edmund Buhler) arcmelting furnace. The 

heating current was 15 A and the ingot cup was cooled by water. All samples were analyzed 

by X-ray powder diffraction as described below. Some samples exhibited diffraction patterns 

with rather broad peaks. The broad peaks were attributed to poor crystallinity. There were 

especially frequent for the more aluminium rich samples. Such samples were subjected to a 

heat treatment at 500 o C for 2 days, whereby the shape of the diffraction peaks improved. The 

prepared BaAl2-xSix (0.4

2.3 Computational details 

Total-energy calculations for BaAl2H2, BaAlSiH and BaSi2 were performed in the 

framework of the frozen core all-electron Projected Augmented Wave (PAW) Method [18], as 

implemented in the program VASP [19]. The energy cut-off was set to 500 eV. Exchange and 

correlation effects were treated by the generalised gradient approximation (GGA), usually 

referred to as PW91 [20]. The integration over the Brillouin zone was done on special gamma 

centred k-point mesh determined according to the Monkhorst-Pack scheme [21]. Total 

energies were converted to at least 1meV/atom. Structural parameters were relaxed until 

forces had converged to less than 0.01 eV/Å. The equilibrium structure of SrAl2H2 [28] was 

used as starting configuration for BaAl2H2. 

3. Results and discussion 

In the BaAl2-xSix (0.4 < x < 1.6) sample series, only BaAlSi (with x = 1) showed a 

completely single phased XRD pattern. BaAl2-xSix (0.4 < x < 1.6) samples deviating from x = 

1 exhibited secondary phases in the diffraction patterns, which were not eliminated by heat 

treatments. For x-value above 1, small peaks of BaAl2Si2O8 [22], BaSi2 (orthorhombic) [1] 

and also some so far unknown impurity peaks were observed. For x-value below 1, some 

other unidentified peaks were also observed. The amount was, however, not enough to allow 

for an identification of the impurity phases. As x increasingly deviates from 1, the intensity of 

the impurity peaks increased. This introduces a small systematic error in the targeted BaAl2xSix 

compositions, especially at higher deviation from x = 1. But as discussed below we 

estimated the deviations to be small and not significantly affecting the general conclusions. 

The BaAl2-xSix crystallizes with hexagonal symmetry in space group P6/mmm (191) 

(cf. Figure 2) [23]. The BaAl2-xSix structure consists of hexagonal [Al2-xSix] 2- layers, which 

are stacked on top of each other. The Ba 2+ ions are sandwiched between these layers. 

The detailed ordering of Al/Si in BaAl2-xSix is not clear, as Al and Si are the neighbours in 

the periodic table and making them difficult to distinguish by XRD. It has been assumed that 

Si and Al atoms are more or less randomly distributed in the hexagonal network, but research 

done by Akimitsu et al [24] on CaAlSi indicate that the atoms in the poly anionic network are 

ordered and therefore the symmetry description of CaAlSi is better described by space group 

P-6m2 instead of P6/mmm. 

Figure 3 and Table 1 show the cell parameter and cell volume change of BaAl2-xSix 

(0.4 < x < 1.6). The a-axis and cell volume is linearly decreasing with increasing x-value. This 

is reasonable as bigger Al atoms are replaced by smaller Si atoms, in the hexagonal network, 

as x increases. The c-axis is rather independent on x. These trends were also reported by 

Yamanaka et al. [25]. For the BaAlSi (x = 1) phase, the cell parameter could be indexed by a 

= 4.2989(6) Å and c = 5.1437(7) Å close to the previously reported BaAlSi, a = 4.290 Å and c 

= 5.140 from Lorenz et al. [13].

Figure 2: Structure of BaAl2-xSix (Space group 191) 

Cell parameter [Å] 

5.2 

5.0 

4.8 

4.6 

4.4 

4.2 

c-axis 

a-axis 

0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 

x-value 

Figure 3: Cell parameters of BaAl2-xSix as a function of x. 

Table 1: Cell parameters of BaAl2-xSix for each x-value 










BaAl0.4Si1.6 4.1948(3) 5.1230(6) 78.07(1)

The corresponding hydrides BaAl2-xSixH2-x (0.4 < x < 1.6) were synthesized by direct 

hydrogenation of the BaAl2-xSix alloys, as described in Table 2 to Table 5, where also the 

results from subsequent XRD phase analysis are given. During hydrogenation, the colour 

changed from metallic silver to dark gray. Careful control of the hydrogenation conditions 

was necessary. To hydrogenate BaAlSi (x = 1), 600 o C and 70 bars of hydrogen pressure was 

required. On the other hand, BaAl1.6Si0.4 (x = 0.4) and BaAl0.4Si1.6 (x = 1.6) start to take up 

hydrogen at a lower temperature, 300 o C at 70 bars of hydrogen pressures. But at higher 

temperature, aluminium rich BaAl2-xSixH2-x (0 < x < 1) tend to decompose to a composition 

close to BaAlSiH (x = 1) and BaAl4 [26] as well as other impurity phases. Also, at higher 

temperature, silicon rich BaAl2-xSixH2-x (1 < x < 2) tend to decompose to composition close to 

BaAlSiH (x = 1) and BaSi2 [1] and other impurity phases. This indicates that BaAlSiH (x = 

close to 1) is a specially stable phase in the system. A similar situation was also observed for 

SrAl2H2 and SrAlSiH, where SrAl2H2 starts to release hydrogen upon heating at 300 o C but 

SrAlSiH is stable to above 600 o C. Probably the stability of the ternary hydrides SrAlSiH and 

BaAlSiH correspond to the stable lattice of there precursor AeAlSi. For SrAl2H2 the precursor SrAl2 

is thermodynamically unfavoured towards SrAl4. When hydrogen is present SrH2 also forms. 

Interestingly the Sr to H bond distance in the ternary hydrides is similar to that in SrH2 [27]. 

SrAlSi also needed to be heated to 600 o C at 50 bars to form the hydride whereas SrAl2 could 

be hydrogenated already at about 190 o C [28]. The 1:1:1 composition is observed to be 

especially stable both as alloy and hydride. 

The obtained cell parameters are shown in Figure 4 and Table 6. The unit cell 

parameters extrapolate nicely between to those of the computationally obtained BaAl2H2 (x = 

0) and the previously reported BaSi2 (x = 2 )[1] [2]. Also the cell parameters of the 

corresponding SrAl2-xSixH2-x system (SrAl2H2 and SrAlSiH) are plotted showing a similar 

trend. When going from BaAl2H2 to BaSi2 by substituting (Al-H) - entities for (Si) - -lone pairs, 

the ab-plane shrinks in good agreement with the difference in covalent radius between Al and 

Si. The a-axis decrease in hydrides is almost identical to that in the alloys, indicating that the 

direct covalent bonding between Al/Si or Si/Si in the net work is not much affected by the 

substitution. On the other hand the c-axis expands as the (Al-H) - entities are substituted by an 

increasing number of repulsive (Si) - -lone pairs indicating that these need more space than the 

(Al-H) - entities. Interestingly is the puckering angle in the Si/Al net work also increasing to 

become 111 o , close to the ideal sp 3 angle, as BaSi2 is reached. The increase in c-axis is also 

followed by an increase in Al-H bond distance. The computational obtained Al-H is 1.71 Å in 

BaAl2H2 increases to 1.74 in BaAlSiH. The difference of Al-H distance is slightly less than 

that in the Sr-system. In SrAl2H2 the Al-H bond is 1.706 Å, which increased to 1.77 Å in 

SrAlSiH. 

It is interesting to compare the synthesis condition of BaSi2 and BaAl2-xSixH2-x. 

Trigonal BaSi2 require a high isostatic pressure to be formed. However, by substituting Si - 

with [Al-H] - a trigonal BaAl0.4Si1.6H0.4 structurally close to BaSi2 could be obtained at 300 ℃ 

and 70 bars of hydrogen pressure. 



300 48 Formation of an assumed BaAl1.6Si0.4H1.6. Unknown 


700 48 Formation of BaAlSiH instead BaAl0.4Si1.6H0.4. Unknown 

phases were also observed.




small peaks of unknown phases are observed. 


BaAlSiH and unknown phases were also observed. 


unknown phases are also observed as a second phase. 




550 3 Formation of an assumed BaAl0.6Si1.4H0.6, but additionally 

BaSi2 and unknown phases were also observed. 

600 12 Formation of a phase close to BaAl0.6Si1.4H0.6. 

Additionally BaSi2 and unknown phases were also 

observed. 


Additionally BaSi2 and unknown phases are also observed. 




small peaks of BaSi2 and unknown phases were observed. 

500 48 Formation of a phase closes to BaAl0.6Si1.4H0.6 instead 

BaAl0.4Al1.6H0.4. Strong peaks of BaSi2, BaSi and 

unknown phases were also observed. 

700 48 Formation of BaAlSiH instead of BaAl0.4Si1.6H0.4. Strong 

peaks of BaSi2, BaSi and unknown phases were also 

observed. 

We observed an interesting stability problem when hydrogenating BaAl2-xSix alloys 

with compositions different from an 1:1:1 stochiometry. If an aluminium rich BaAl2-xSix (0 < 

x < 1) or a silicon rich BaAl2-xSix (1 < x < 0) are hydrogenated at the lowest temperatures 

when a reaction is observed to proceed, as given in Tables 2-5, then a hydride with the same 

metal atom composition will be produced. But if the temperature is increased, the hydride 

composition will start to move toward the stable 1:1:1 ratio. At this composition all the [Al- 

H] - or silicon lone pairs point away from each side of the network. This probably leads to a 

more relaxed structure and with the electrons being localized, a more simple description in 

terms of a sp 3 hybridization can be justified, but as the composition deviates from the 1:1:1 

ratio either hydrogen or electron lone pairs appear on the both side with increasing overlap 

problems, between the layers.

Table 6: Cell parameters for BaAl2-xSixH2-x and BaAl2-xSix as function of x. The cell 

parameters of the alloys are shown in brackets. 


BaAl2H2 (calculated) 4.6524 4.9768 93.29 



BaAlSiH 4.3145(5) [4.2989(6)] 5.2049(7) [5.1438(7)] 83.91(2) [82.32(2)] 



BaSi2 [2] 4.047(3) 5.330(5) 75.6(1) 

Figure 4: Cell parameters of BaAl2-xSixH2-x 

(Computationally obtained BaAl2H2, ★: BaSi2 [2])

DOS 

10 

8 

6 

4 

2 

0 

-4 -2 0 2 

E-E F [eV] 

BaSi 2 


BaAl 2 H 2 

Figure 5: DOS of BaAl2-xSixH2-x (x=0,1 and 2) 

DOS calculation shows BaAlSiH to be a semiconductor with narrow band gap 0.71 eV 

Trigonal BaSi2 (x=2) is reported to be an electric conductor [3] which could also be confirmed 

by our calculations above. The DOS calculation for BaAl2H2 indicates metallic conductivity. 

These is interesting, because by substituting Si - with [AlH] - in BaAl2-xSixH2-x, the electric 

property of BaAl2-xSixH2-x can be changed. We expect that at some x-value between 1 and 2, a 

transformation from a semiconductor to a metallic conductor will occur and again at some x 

value between 1 and 0, the transition will be reversed. In other word, the band gap will be 

tuneable by substituting Si - with [AlH] - . 

Moreover, it would be interesting to investigate a possible superconductivity transition 

in theses systems. The DOS of all the 1:1:1hydrides is characterized by a pronounced 

singularity below the Fermi level. By tuning the Al/Si ratio and the hydrogen content it might 

be possible to find even higher Tc in this type of compound. 

4. Conclusion 

A series of Zintl phase hydrides with closely related structures (BaAl2-xSixH2-x (0.4 < x 

< 1.6)) has been synthesised in between BaSi2 and BaAl2H2, by gradual substituting Si - (a 

silicon-lone pairs) with isoelectric [AlH] - entities. The structure in this trigonal system can be 

described by a two dimensional network of interconnected [Al2-xSixH2-x] 2- ions sandwiching 

Ba 2- ions. A simple electron count would suggest straight forward sp 3 hybridization in the 

network. Metallic BaSi2 and BaAl2H2 encompassing semiconducting BaAlSiH indicate a 

more complex electron structure, at least at each end of the compositional chain. This opens 

up for interesting electric phenomena related to this tunable substitution. Hydrides, close to 

the 1:1:1 metal atom ratio as in BaAlSiH, were more stable. At this composition all the [Al- 

H] - or silicon lone pairs point away from each side of the network. As the composition 

deviates from the 1:1:1 ratio either H or electron lone pairs appear on the both side with 

increasing overlap problems.

5. References 

[1] H. Schäfer, K. H. Janzon, and A. Weiss, Angew. Chem. Int. Ed. Engl. 2, 393 (1963). 

[2] J. Evers, G. Oehlinger, and A. Weiss, Angew. Chem. Int. Ed. Engl. 16, 659 (1977). 

[3] J. Evers, G. Oehlinger, and A. Weiss, Angew. Chem. Int. Ed. Engl. 17, 538 (1978). 

[4] J. Evers, J. Solid State Chem. 32, 77 (1980). 

[5] M. Imai, T. Hirano, J. Alloys Compd. 55, 132 (1997). 

[6] M. Imai, K. Hirata, T. Hirano, Physica C 245, 12 (1995). 

[7] M. Affronte, S. Sanfilippo, M. Nunez-Regueiro, O. Laborde, S. Lefloch, P. Bordet, M. 

Hanfland, D. Levi, A. Palenzona, G.L. Olcese, Physica B 284-288, 1117 (2000). 

[8] M. Imai, T. Hirano, J. Alloys Compd. 224, 111 (1995). 

[9] J. Evers and A. Weiss, Mat. Res. Bull. 9, 549 (1974). 

[10] T. Björling, D. Noréus, K. Jansson, M. Andersson, E. Lenova, M. Edén, U. Hålenius, and 

U. Häussermann, Angew. Chem. Int. Ed. 44, 7269 (2005). 

[11] To be published. 

[12] S. Yamanaka, T. Otsuki, T. Ide, H. Fukuoka, R. Kumashiro, T. Rachi, K. Tanigaki, F. 

Guo, K. Kobayashi, Physica C 451, 19 (2007). 

[13] B. Lorenz, J. Lenzi, J. Cmaidalka, R.L. Meng, Y.Y. Sun, Y.Y. Xue and C.W. Chu, 

Physica C 383, 191 (2002). 

[14] K. E. Johansson, T. Palm, P. -E. Werner, J. Phys. E 13, 1289 (1980). 

[15] P. E. Werner, L. Eriksson, S. Salome, SCANPI, A Program for Evaluating Guinier 

Photographs, Stockholm University, Stockholm, 1980. 

[16] P. E. Werner, L. Eriksson, M. Westerdahl, J. Appl. Crystallogr. 18, 367 (1985). 

[17] P. E. Werner, Ark. Kemi 31, 513 (1969). 

[18] a) P. E. Blöchl, Phys. Rev. B 50, 17953 (1994). b) G. Kresse, J. Joubert, Phys. Rev. B 59, 

1758 (1999). 

[19] a) G. Kresse, J. Hafner, Phys. Rev. B, 47, 558 (1993). b) G. Kresse, J. Furthmüller, Phys. 

Rev. B 54, 11169 (1996). 

[20] J. P. Perdew, Y. Wang, Phys. Rev. B 45, 13244 (1992). 

[21] H. J. Monkhorst, J. D. Pack, Phys. Rev. B 13, 5188 (1972). 

[22] N. N. Roy, Nature 206, 501 (1965). 

[23] M. Imai, K. Nishida, T. Kimura, H. Kitazawa, H. Abe, H. Kito, K. Yoshii, Physica C 382, 

361 (2002). 

[24] 14. S. Kuroiwa, T. Kakiuchi, H. Sagayama, H. Sawa and J. Akimitsu. Physica C: 

Superconductivity, In Press, Corrected Proof. (2007) 

[25] Shoji Yamanaka, Teruyoshi Otsuki, Takayuki Ide, Hiroshi Fukuoka, Ryotaro Kumashiro, 

Takeshi Rachi, Katsumi Tanigaki, FangZhun Guo and Keisuke Kobayashi, Physica C 451, 19 

(2007) 

[26] Andress, K.R.;Alberti, E, Zeitschrift fuer Metallkunde, 27(6), 126-125, (1935) 

[27] Zintl, E.;Harder, A, Strukturbericht, 3, 306-307 (1937) 

[28] F. Gingl, T. Vogt, E. Akiba, J. Alloys Compd. 306, 127 (2000)

Paper II

Characterisation of two new Zintl phase hydrides BaAlSiH and CaAlSiH 

T. Björling*, B. C. Hauback**, T. Utsumi*, D. Moser*, D. Noréus*, U. Häussermann*** 

*Structural Chemistry, Stockholm University SE-10691 Stockholm, Sweden. 

**Institute for Energy Technology, P.O. Box 40, NO-2027 Kjeller, Norway. 

***Department of Chemistry and Biochemistry, Arizona State University. 

Abstract 

Two new Zintl phase hydrides CaAlSiH and BaAlSiH were synthesised by reacting hydrogen 

with the ternary alloys. Both structures were determined from powder neutron diffraction on 

the corresponding deuterides and were found to be isostructural to SrAlSiH. (Björling et al. 

Angew. Chem. Int. Ed. 44(2005)7269) The unit cell dimensions for the trigonal space group 

P3m1 were a = 4.1278(7) and c = 4.7618(9) Å for CaAlSiH, and a = 4.3186(4), c =5.2080(9) 

Å for BaAlSiH. DFT-calculations showed CaAlSiH and BaAlSiH to be semiconductors with 

narrow indirect bandgaps of 0.42 eV and 0.71 eV, respectively. The refined Al-D distance 

was within one standard deviation the same, 1.73 Å, for both hydrides and in close agreement 

with the DFT calculated of 1.75 Å. The alkaline earth-hydrogen distances were also in line 

with that in the corresponding binary alkali earth hydrides, showing again a significant alkali 

earth-hydrogen contribution to the bonding situation in the new hydrides. 

Introduction 

Light weight aluminium is tempting to use in metal hydrides for hydrogen storage to reduce 

weight. Aluminium can form alane, AlH3, with a hydrogen storage capacity of 10 %. The 

direct reaction to form AlH3 is, however, difficult as an extreme; mega bar hydrogen pressure 

is needed since the polar covalent aluminium-hydrogen bond is weak and therefore difficult to 

form. [1] 

Alkaline or alkaline earth metal alanates are somewhat more reversible i e NaAlH4, Na3AlH6 

and have been subject to a lot of research. Especially since Bogdanović et al found that certain 

titanium containing catalysts could improve the usually sluggish reaction kinetics. [2] Later 

also other transition metal additives have been found to work even better. [3] A detailed 

understanding, upon how the transition metal additives improve the reaction kinetics, is, 

however, still lacking. Alanates are systems which can be seen as hydride ions added to an 

AlH3 –unit, decreasing the polarization of the covalent bond and hence increasing its 

covalence and strength. Alanate synthesis requires 100-200 bars of hydrogen pressure at 

temperatures slightly over 100 o C, which is still somewhat impractical for hydrogen storage, 

and a search for further strengthening the Al-H bond to the metal atom framework is needed. 

Attempts to produce alanates for hydrogen storage materials, based on alkaline earth metals, 

resulted in the discovery of a new hydride, SrAl2H2 [4]. This Zintl phase hydride differs from 

ordinary complex hydrides in that aluminium is bonded to both aluminium and hydrogen. 

Sensitivity to moisture and oxygen is lower compared to saline hydrides as SrH2. This is 

expected since the negative charge is distributed on the poly anionic part in SrAl2H2 compared 

to only on hydrogen, as in SrH2, resulting in less basic characteristics for the poly anionic 

hydrides. SrAl2H2 synthesis starts from SrAl2 where aluminium is connected to four other 

aluminium atoms; in a network where aluminium has a formal charge of -I. One of the Al-Al 

distances is longer than the other three, and probably breaks first under hydrogenation, 

resulting in the hydride. The reaction mechanism when synthesising SrAl2H2 is different from 

that of alane and alanate. An electron input to the aluminium environment puts aluminium in 

a reduced state and decrease the extremely difficult synthesis conditions to rather mild 

conditions, for SrAl2H2 at 50 bar and 190 o C. The Al-Al bond distances are within 2.76 to 

3.00 Å in SrAl2 compared to 2.87 Å in ordinary cubic fcc aluminium. 

Further, substitution of one of the aluminium atoms in SrAl2, with an equivalent amount of 

silicon, gives poly anionic SrAlSi, where the p-block elements are arranged in a hexagonal 

network sandwiched by strontium atoms. The silicon addition leads to SrAlSiH, which is also 

less reactive to water and air in comparison with SrAl2H2. The structures of SrAl2H2 and 

SrAlSiH are similar. Networks of (Al2H2) 2- or (AlSiH) 2- , Zintl ions are sandwiched between

Sr ions (fig 1.) In SrAlSiH a Si-electron lone pair has substituted every second (Al-H - ). A 

simple electron count would suggest a straight forward sp 3 hybridization for both (Al2H2) 2- or 

(AlSiH) 2- . Metallic character of SrAl2H2 in combination with a flatter puckering of the 

(Al2H2) 2- ion compared to the (AlSiH) 2- ion and a non metallic SrAlSiH, suggest that this is an 

oversimplification, but that the latter has a less problematic electronic structure. This could be 

explained by the fact that strontium in SrAlSiH interacts with three hydrogen atoms, instead 

of six as in SrAl2H2. Hydrogen acts as a bridge for the electron density from the network 

towards the polarizing positive ion. This leads to a shorter bond distance between strontium 

and hydrogen in SrAlSiH and increased covalent character of the strontium-hydrogen 

interaction in SrAlSiH compared to SrAl2H2. 

This shows that changes in the aluminium environment affect the electronic structure and 

related properties. The discovery of SrAl2H2 and SrAlSiH gave an opportunity to manipulate 

the anionic network environment to study how a substitution in the SrAl2 influences the Al-H 

bond. Hopefully this can help us to understand more about the alane and alanate compounds. 

The aim is to clarify if it is possible to design aluminium based hydrides with a higher 

stability and lower absorption pressures for hydrogen. In a first study we used silicon as a 

substituent in SrAl2 to increase the amount of electrons in the poly anionic network to obtain a 

more stabile hydride. In the present paper we want to report the homologs BaAlSi and CaAlSi 

and their properties. These systems are also of interest for their electric properties. CaAlSi, 

SrAlSi and BaAlSi have been investigated for superconductivity as their structure is closely 

related to MgB2, with a Tc of 39 K [5a]. The Tc of CaAlSi and SrAlSi - 7.9 K and 5 K, 

respectively, were obtained from a previous study [5b]. No Tc above 2 K has so far been 

found for BaAlSi. SrAlSiH is a semiconductor but DOS (density of states) calculations show 

a narrow gap and flat band just below the Fermi level. A trigonal high pressure phase of CaSi2 

with a similar electron structure was also found to be superconducting up to at 14 K, which is 

among the highest for silicides so far [6]. 

Experimental 

The synthesis of starting alloys, (Ba, Ca)AlSi were done by arc melting stoichiometric ratios 

of the elements in a MAM-1 (Edmund Buhler) arc melting furnace under argon. The heating 

current was 15 A and the ingot cup was cooled by water. To obtain homogeneity, the sample 

was turned over and melted several times. 

During the heating the tablet melts to a mercury looking droplet. 

High temperature can cause evaporation of the alkaline earth metal and make the 1:1:1 phase 

to decompose to more stable 1:2:2 phase, e.g. a 1:4 compound. Traces of an amorphous black 

powder could be observed covering the cold cupper surface, if the sample was overheated 

during the synthesis. All powders were handled in an inert atmosphere inside a glove box with 

less than 1 ppm of O2 and H2O. All sample compositions were observed to be homogenous 

using the EDX (energy-disperse X-ray) method in a JEOL 820 scanning electron microscope. 

Reactions with hydrogen 

The CaAlSi and BaAlSi tablets/ingots were placed in corundum tubes and reacted with 

hydrogen in sealed stainless steel autoclaves. Reactions were carried out at hydrogen 

pressures around 50 bar and the temperature was varied between 150 and 700 o C. 

A stainless steel sealed thermo couple inserted into the powder directly recorded the reaction 

temperature. The result of these reaction series is compiled in table 1 and 2.. Hydride 

formation was observed at temperatures from 550 o C for BaAlSi. Attempts to synthesise 

CaAlSiH from CaAlSi was done in an extensive trial and error series, at pressures from 5 to 

90 bar, starting at temperatures as low as 250 o C. The formation of CaAlSiH was not expected 

as the disintegration into a mixture of CaH2 and CaAl2Si2 is energetically more favourable 

than the reaction to CaAlSiH. During our experiments we got indications that slow kinetics 

for this reaction might initially favour the formation of CaAlSiH. All experiments at higher 

temperature than 270 o C resulted in CaAl2Si2 and amorphous CaH2. 

The first time CaAlSiH lines appeared in the powder diffractograms was in a synthesis done 

at 310 o C for 2 days at 70 bar hydrogen pressure. These lines were broad and a diffuse and a 

large amount of CaAl2Si2 were observed. Under the same conditions the experiment was

epeated but the hydrogenation time was decreased to 3 hours and the sample did now contain 

a significant increased amount of CaAlSiH among remaining CaAlSi and some formed 

CaAl2Si2. The best results were obtained when CaAlSi was hydrogenated at 500 o C for short 

times, 40 minutes at 50 bar of hydrogen pressure. This indicates that the formation of 

CaAlSiH is kinetically controlled. 

Reducing the pressure down to five bars did not produce any hydride. 

A deuterated sample was obtained by reacting BaAlSi with deuterium at a pressure of 50 bar 

and a temperature of 700 o C for 4 days. CaAlSi was reacted with deuterium at 500 o C under 

40 minutes at a pressure of 40 bar. 

Structural characterisation 

All products obtained were characterised by Guinier Hägg x-ray diffraction (Cu Kα1, Si was 

used as internal standard). Lattice parameters (table 3) were obtained from least-squares 

refinement of the corresponding Guinier Hägg x-ray diffractograms with the program Pirum 

[7]. 

All atomic positions of CaAlSiD and BaAlSiD (table 4) were determined from Rietveld 

profile refinements of neutron powder diffraction data from deuteride samples measured at 

room temperature with the PUS powder diffractometer at the JEEP II reactor at Institute for 

Energy Technolgy Kjeller, Institute for Energy Technology (IFE), Kjeller, Norway [8a]. The 

wavelength was 1.5554 Å and the samples were kept in 6 mm vanadium sample holders. The 

program FULLPROF [8b] was used for the refinement of structural parameters of CaAlSiD 

and BaAlSiD. Neutron diffraction spectra of CaAlSiD and BaAlSiD are shown in fig 2 and 3. 

To get a high yield of CaAlSiD, the reaction with deuterium had to be performed within a 

short time of about 40 to 45 minutes at 500 o C. This short time of hydrogenation probably was 

insufficient to get a well defined micro structure of CaAlSiD leading to problem with the 

crystalinity as manifested by anisotropic broadening of the peaks. The (00l) reflections are 

twice as broad as the (hk0) reflections and (hkl) reflections are in between. Intercalation of 

hydrogen along c-axis is probably the reason for this trend in broadening of the reflections. 

Hydrogenation of CaAlSi at higher temperature and/or longer time to obtain a better 

crystallinity lead to decomposition to thermodynamically more stable CaAl2Si2 and CaD2 

reducing the amount of CaAlSiD. 

The crystallinity problem was taken into account by refining the crystal size parameter which 

is implemented in FULLPROF. This procedure was not needed for the refinement of the 

corresponding compound SrAlSiD and BaAlSiD which had been hydrogenated for a longer 4 

days period. The refinement of the crystal size parameters improved the peak fitting, although 

we still detected a deviation in the (110) reflection. 

The SrAlSi starting alloy was studied with the polaris spectrometer at ISIS research centre in 

Rutherford England. The aim was to investigate if the Al/Si network was ordered or not. This 

was not possible to verify but the Rf value for the refinement of the structure decreased from 8 

to 6 % for a slightly puckered Al/Si, in spacegroup P-6m2, compared to a flat as described in 

P6/mmm. This could indicate that the Al/Si network is ordered similar to what Akimtsu et al 

found for CaAlSi as discussed below. 

Diffuse reflectance measurements 

Optical diffuse reflectance measurements were made on finely ground samples of BaAlSiH 

and CaAlSiH at room temperature under an argon gas flow. The spectrum was recorded in the 

region 3200 – 10500 cm -1 with a Bruker Equinox 55 FT-IR spectrometer equipped with a 

diffuse reflectance accessory (Harrick). The measurement of diffuse reflectivity can be used 

to determine band gaps. For this, absorption data were extracted from the reflectance data by 

using the Kubelka-Munk function. 

Thermal investigation 

The equipment setup consisted of a sealed steel autoclave with the sample located in a 

corundum tube, placed in a tube furnace, connected to a pressure meter and a vacuum pump.

A stainless steel sealed thermo couple was inserted into the powder to directly record the 

reaction temperature. The sample chamber was evacuated and closed and the hydrogen 

pressure increase was monitored during the heating of the system. The amount of sample was 

several grams of a finely grinded powder. 

Theoretical calculations 

The electronic structure calculations present in this paper were performed with density 

functional theory (DFT) with the PW91 [9] gradient dependent functional implemented in the 

VASP code [10]. An energy cutoff of 500 eV was used for all the calculations and k-points 

sampling was performed on a Gamma centred Monkhorst-Pack grid [11]. All necessary 

convergence tests were performed and total energies were converged to at least 1 meV. All 

structures were fully relaxed (cell parameters, cell shape and atomic coordinates) with forces 

converging to less than 0.01 eV Å -1 . The equilibrium volume Veq were obtained by fitting E 

vs. V values to a Birch-Murnaghan equation of state. In the initial structural input atoms were 

positioned as suggested from experimental data of SrAlSiH. 

Total energy calculations comparing the stability of the reactions involving CaAlSi, CaAlSiH, 

CaAl2Si2, H2 and CaH2 were also done. 

Inelastic neutron scattering 

The INS measurement of BaAlSiH was carried out using TOSCA [12], a crystal-analyzer 

inverse-geometry spectrometer operating at the ISIS pulsed neutron spallation source 

(Rutherford Appleton Laboratory, UK), accessing an energy transfer from 0 to 4000 cm -1 , 

providing an excellent energy resolution, ΔE, (delta E/E=1.5-3%). Special care was taken to 

prevent possible alkaline earth hydroxide formation by working in an inert-gas glove box. The 

BaAlSiH powder was loaded into a flat aluminum cell and then loaded into the TOSCA 

closed-cycle refrigerator at 20 K. 

The spectrum (fig.4) revealed an impurity which corresponds to BaH2 [13]. 

Phonon calculation requires zero pressure and zero force inside the unit cell and therefore the 

optimized unit cell with parameters listed in S1 (supporting information) was used. 

A 2x2x2 supercell has been afterwards generated and the cell parameters further optimized. 

Previous study on SrAlSiH compound [14] showed that zone-center phonon calculation can 

successfully be applied on this kind of system. That has been slightly improved in the present 

work. 

The vibrational modes at the Г-point of the brillouin zone generated by the supercell are 

calculated using Finite Differences as implemented in Vasp [10]. Each atom is displaced, one 

at a time, by a small distance along each of the Cartesian directions and the Hellmann- 

Feynman force acting on each atom inside the cell is determined. The displacement has to be 

small enough in order to remain in the range of harmonic approximation. Here we used a 

value of 0.01 Å. 

The force-constant matrix is determined by dividing the force by the displacement and as final 

step the dynamical matrix is diagonalized for q=0. These results have been used to calculate 

the INS spectra by the aCLIMAX program [15]. 

It has to be stressed at this stage that eigenvalues for high symmetry K-Points other than 

gamma arise from a zone center calculation on a supercell. This is due to the folding back to 

the gamma point of the points laying on the brillouin zone boundaries. Therefore a further 

increasing of the supercell dimension would increase the brillouin zone sampling; that is not 

required in this work due to the small dispersion of bands in the system. 

In fact the calculated spectrum shows a LO-TO splitting for the bending mode at 884 cm -1 in 

good agreement with the experiment. Previous work on SrAlSiH did not show this feature due 

to the use of symmetry consideration in order to reduce the number of displacements giving as 

results eigenvalues just at the Г-point of the single cell brillouin zone. 

As in the case of Ae=Sr, the spectrum of BaAlSiH presents two narrow high intensity bands 

which can unambiguously be assigned to [Al-H] bending (884 cm -1 ) and stretching (1189 cm - 

1 ) modes, respectively.

The overtones arising from the internal modes are identified with peaks at 1770 cm -1 

(bending-mode overtone), at 2089 cm -1 (bending-stretching combination mode) and at 2285 

cm -1 (stretching-mode overtone). 

The stretching-mode overtone shows a deviation from the harmonic approximation when 

comparing the calculated and experimental INS spectra with a shift downward by about 100 

cm -1 . 

The anharmonicity of the modes can be studied with fitting the potential energy curves for the 

displacement of the H atom along x or y (in plane) or z (along the Al-H bond). The 

k 2 3 4 2 

x -αx +βx with k=mω . Table 5 shows the calculated 

polynomials used for the fit is U= 2 

energies for the Ground state E0 and the first and second excited states. The corresponding 

transition frequencies were calculated and the shift delta=2ω1-ω2 due to the anharmonicity 

was estimated with a satisfactory agreement with the experimental observed value. 

For future reference the INS spectra for CaAlSiH has been calculated (fig 4) with the [Al-H] 

bending mode at 765 cm -1 and the stretching mode at 1200 cm -1 . 

Results and Discussion 

BaAlSiH and CaAlSiH are both isostructural with the earlier reported SrAlSiH. BaAlSiH and 

SrAlSiH were found to be final products whereas SrAl2H2 could be further hydrogenated to 

different alanates [12]. For (Sr,Ba)Ga2 the corresponding final compounds are (Sr,Ba)Ga4 and 

(Sr,Ba)H2 respectively [17]. This is similar to CaAlSi with its unstable hydride, CaAlSiH that 

decomposes into CaAl2Si2 and calcium hydride, as was observed to occur during the 

experiments according to: 

2CaAlSiH � CaAl2Si2 + CaH2 ∆ G = -0.7 eV (Total energy values for the computationally 

relaxed structures was used to calculate ∆ G at 0 K.) 

When CaAlSi rearrange to form CaAl2Si2 and CaH2 higher activation energy is required than 

forming CaAlSiH as the metal atoms have to be arranged in to new structures. Thus the 

formation of CaAlSiH is more kinetically favoured although the total energy would favour a 

separation into CaAl2Si2 and CaH2. 

The new hydrides require higher synthesis temperatures compared to the synthesis of 

SrAl2H2. Probably this is due to a different bonding situation related to the π electron 

interactions in the AeAlSi ring structure making the compound more resistant towards 

hydrogen. This is supported by shorter bond distances between the p-block elements in - 

[AlSi] 2- - compared to the longer bonds in the -[Al2] 2- - ion. In SrAl2 each negatively charged 

aluminium atom bonds to four neighbour atoms, hence aluminium is four coordinated in 

contrast to -[AlSi] 2- - where each network atom connects to three other. The Al-Si distance 

among these hydrides is around 2.5 Å. The [-Al-Al-] distance in SrAl2 varies between 2.76 to 

3.00 Å. 

It has been discussed if silicon and aluminium distributes randomly or ordered in the network. 

A study made by Akimitsu et al [18] on CaAlSi indicate that the atoms in the poly anionic 

network are ordered and therefore the symmetry description of CaAlSi is better described by 

space group P-6m2 instead of P6/mmm. As mentioned above our neutron diffraction 

measurement on SrAlSi favoured a slightly puckered anionic network layer in spacegroup P- 

6m2 suggesting an ordered distribution on the Al-Si sites also in SrAlSi. 

Comparing the observed Ae-D atomic distances in the hydrides indicate an ionic contribution 

to the bonding. The Ae-D distance in BaAlSiD is 2.58Å, in SrAlSiD 2.48 Å and for CaAlSiD 

2.42 Å, these distances are comparable to Ae-H bond distance in the correspondingly binary 

hydride. The Al-H in the hydrides was 1.71 Å for SrAl2H2 vs. 1.78 Å for SrAlSiD and for 

BaAlSiD it is 1.73 Å and CaAlSiD 1.74 Å. Respectively this trend could be connected to the 

alkaline earth metal ion, in AeAlSiH, that in addition to an ionic description also has a 

covalent contribution that weekly attracts electron density from the poly anion, -[SiAl-H]- 2- , 

through bridging hydrogen. Therefore AeAlSiH has a certain covalent character although 

these hydrides are described within the Zintl phase concept. The polarisation of the network 

depends mainly on the charge and radius of the positive ion. For the elements Ca, Sr and Ba 

ion radius are 0.99, 1.13 and 1.35 Å [19]. Mg 2+ has got a radius which is 65 % of that of Ca 2+ . 

The smaller Mg 2+ radius would result in strong covalent interactions in a hypothetic hydride.

Attempts to produce MgAlSi and its hydride from arc melting of the elements and by heating 

a mixture of MgH2, Si and Al, however all ended in Mg2Si and Al. 

Theoretic calculations showed a more puckered poly anionic network with a smaller positive 

ion as for the calcium ion. This trend was not clearly confirmed from the experimental data. 

Upon hydrogenation CaAlSi undergoes a unit cell volume expansion of 3.67 [Å/H-atom] 

compared to 2.45 [Å/H-atom] for SrAlSiH and 1.65 [Å/H-atom] for BaAlSiH. For the Casystem 

the a-axis shrinks from 4.1902(4) Å to 4.1278(7) Å, the c-axis expands from 4.3994(6) 

to 4.7618(9). Upon hydrogenation such a large expansion of cell volume [Å/H-atom] is 

common for metal hydrides and usually causes the hydride alloy to disintegrate into a fine 

powder. But despite this larger volume expansion and the fact that the a-axis shrinks and the 

c-axis expands, we interestingly did not observe any disintegration of our particles. 

Decomposing the hydrides by heating, in an evacuated steal autoclave, indicated that 

CaAlSiH is more unstable than the others. In these experiments the kinetics can not be 

determined. But several qualitative experiments done under vacuum showed that hydrogen 

release was observed to start at 520 o C for SrAlSiH and at 500 o C for BaAlSiH. For the 

CaAlSiH-system all decomposition experiments gave back the starting material, CaAlSi, and 

the observed hydrogen release started at 420 o C. 

The DOS curves in fig 5 for BaAlSiH and CaAlSiH show features similar to SrAlSiH [20] 

with a narrow gap at the Fermi level, 0.71 eV and 0.42 eV respectively, compared to 0.65 eV 

for SrAlSiH. Band structure calculation indicated as previously reported for SrAlSiH an gap 

of indirect nature between the high symmetry point A and M. The indirect gap were measured 

and confirmed to be indirect with a value of n=2 in the Tauc plot. The measured gaps were 

0.47 eV for BaAlSiH and 0.74 eV for CaAlSiH. These measured values are of the right order 

of magnitude but numerically reversed with respect to the calculated values. It should be kept 

in mind, however, that small bandgaps are both difficult to calculate as well as to measure 

correctly. Errors in the fit procedure as well as, in this case, uncertainties in hydrogen content 

and possible topological disorder add to these difficulties [21]. Interestingly it might however 

be possible to modify the band gap of the hydrides by continuously substituting (Al-H) units 

with silicon, with electron lone pairs, in the Al-Si network. Both AeAl2H2 and AeSi2 are 

expected, or have been observed, to be metals whereas the intermedial AeAlSiH are 

semiconductors as discussed above. 

Acknowledgements: 

Dr. Juan Rodriguez-Carvajal for his help with refining the anisotropic peak broadening in the 

CaAlSiH spectra. 

Dr Kjell Jansson for his help regarding transmittance/absorbance measurements. 

We are grateful to Professor Ulf Hålenius at Swedish Natural Museum of Natural History for 

assistance with IR measurements. 

Partial funding by the European Commission DG Research (contract SES6-2006-518271/ 

NESSHY) is gratefully acknowledged.

References: 

[1] S. K. Konovalov, B. Bulychev, M. Zhurnal, Neorganicheskoi Khimii 37(12) (1992) 2640- 

2642. 

[2] B. Bogdanović, M. Schwickardi, J. Alloys Compd. 253-254 (1997) 1-9. 

[3] D. Blanchard, H.W. Brinks, B.C. Hauback, P. Norby, J. Muller, J. Alloys Compd. 404-406 

(2005) 743-747. 

[4] F. Gingl, T. Vogt, E. Akiba, J. Alloys Compd. 306 (2000) 127-132. 

[5 a] J. Nagamatsu, N. Nakagawa, T. Muranaka, Y. Zenitani, J. Akimitsu, Nature 410 (2001) 

63-64. 

[5 b] B. Lorenz, J. Lenzi, J. Cmaidalka, R.L. Meng, Y.Y. Sun, Y.Y. Xue, C.W. Chu, Physica 

C 383 (2002) 191-196. 

[6] M. Affronte, S. Sanfilippo, M. Nuñez-Regueiro, O. Laborde, S. LeFloch, P. Bordet, M. 

Hanfland, D. Levi, A. Palenzona, G. L. Olcese, Physica B. 1117 (2000) 284-288. 

[7] P. E. Werner, Ark. Kemi 31 (1969) 513-515. 

[8 a] B. C. Hauback, H. Fjellvåg, O. Steinsvoll, K. Johansson, O.T. Buset, J. Jorgensen, J. 

NeutronRes. 8 (2000) 215-232. 

[8 b] J. Rodriguez-Carvajal, "FULLPROF: A Program for Rietveld Refinement and Pattern 

Matching Analysis", Abstracts of the Satellite Meeting on Powder Diffraction of the XV 

Congress of the IUCr, Toulouse, France (1990) 127. 

[9] P. E. Blöchl, Phys. Rev. B 50 (1994) 17953-17979; G. Kresse, J. Joubert, Phys. Rev. B 59 

(1999) 1758-1775. 

[10] G. Kresse, J. Hafner, Phys. Rev. B 47 (1993) 558-561; G. Kresse, J. Furthmüller, Phys. 

Rev. B 54 (1996) 11169-11186. 

[11] H. J. Monkhorst, J. D. Pack, Phys. Rev. B 13 (1972) 5188-5192. 

[12] D. Colognesi, M. Celli, F. Cilloco, R.J. Newport, S.F. Parker, V. Rossi-Albertini, F. 

Sacchetti, J. Tomkinson, M. Zoppi, Appl. Phys. A 74 Suppl. 1 (2002) 64-66. 

[13] D. Colognesi, G. Barrera, A.J. Ramirez-Cuesta, M. Zoppi, J. Alloys Compd. 427 (2007) 

18-24. 

[14] M. H. Lee, O. F. Sankey, T. Björling, D. Moser, D. Noréus, S. F. Parker,U. 

Häussermann, Inorg. Chem. 46 (2007) 6987-6991. 

[15] A.J. Ramirez-Cuesta Comput, Phys. Commun. 157 (2004) 226-238. 

[16] F. Gingl, T. Vogt, E. Akiba, J. Alloys Compd. 306 (2000) 127-132. 

[17] T. Björling, D. Noréus, U. Häussermann, J Am Chem Soc. 128(3) (2006) 817-824. 

[18] S. Kuroiwa, T. Kakiuchi, H. Sagayama, H. Sawa, J. Akimitsu, Physica C: 

Superconductivity In Press, Corrected Proof (2007). 

[19] Table of periodic properties of the elements. Sargent-Welch Scientific company. 7300 

Linder Avenue, Skokie, Illinois 600076. 

[20] T. Björling, D. Noréus, K. Jansson, M. Andersson, E. Leonova, M. Eden, U. Hålenius, U. 

Häussermann, Angewandte Chemie 44 (2005) 7269-7273. 

[21] S. Knief, W. von Niessen, Phys. Rev. B 59[20] (1999) 12940-6.

Figure captions: 

Fig 1: Crystal structure of AeAlSi (a) and AeAlSiH (b) viewed along [110]. Red, blue, and 

green circles denote Ae, Al, Si, and H atoms, respectively. 

Fig 2: Rietveld graph of neutron data for CaAlSiD. Bragg R-factor: 4.99 % Rf-factor: 2.51 % 

Fig 3: Rietveld graph of neutron data for BaAlSiD. Bragg R-factor: 5.20 % RF-factor: 4.49 % 

Fig 4: Measured (red plots) and calculated (blue plots) INS spectra of AeAlSiH. The 

calculations correspond to a supercell gamma-point treatment. Intensities from impurities are 

marked with asterisks. The two plots from the bottom correspond to previous SrAlSiH, the 

middle plot is CaAlSiH and the two in the top represents the INS spectra for BaAlSIH. 

Fig 5: Electronic density of state of the AeAlSiH systems. 

Table 1 

CaAlSi + H2 (H2 pressure 90 bar. The the last two experiments done at 60 bar.) 


250 48 No hydride was present(Only CaAlSi was observed). 

260 48 No hydride was present(Only CaAlSi was observed). 

270 48 No hydride was present. Formation of CaAl2Si2. 

450 48 CaAlSiH was formed. Formation of CaAl2Si2. 

500 0.833 Almost single phased CaAlSiH was formed. 

Table 2 

BaAlSi + H2 (H2 pressure 70 bar) 


200 48 No hydride was present(Only BaAlSi was observed). 

500 48 No hydride was present(Only BaAlSi was observed). 

600 48 BaAlSiH was formed. Small peaks of an unknown phase 


700 48 BaAlSiH was formed. Small peaks of an unknown phase 


Table 3 

Compound BaAlSiH BaAlSiD BaAlSi CaAlSi CaAlSiH CaAlSiD 

space group, Z P3m1, 1 P3m1, 1 P6/mmm,1 P6/mmm,1 P3m1, 1 P3m1,1 

a, Å 4.3186(4) 4.3087(6) 4.3026(6) 4.1902(4) 4.1278(7) 4.1316(7) 

b, Å 4.3186(4) 4.3087(6) 4.3026(6) 4.1902(4) 4.1278(7) 4.1316(7) 

c, Å 5.2080(9) 5.203(1) 5.1441(9) 4.3994(6) 4.7618(9) 4.766(1) 

V, Å 3 84.12 83.66 82.47 66.89 70.27 70.45 

a/c 0.829 0.828 0.836 0.952 0.867 0.867 

Refinement results for AeAlSi( H/D) (cell parameters obtained from X-ray powder data) (Cu 

Kα1; Si standard)

Table 4 

Atom X y z B occ. 

Ba 0 0 0 0.7(2) 1 

AL 2/3 1/3 0.538(2) 1.1(2) 1 

Si 1/3 2/3 0.448(2) 0.7(2) 1 

D 2/3 1/3 0.870(2) 1.5(1) 1 

Atom X y z B Occ. 

Ca 0 0 0 1.1(2) 1 

AL 2/3 1/3 0.548(5) 0.4(1) 1 

Si 1/3 2/3 0.462(2) 0.4 (1) 1 

D 2/3 1/3 0.915(5) 3.6(2) 1 

Atomic positions of BaAlSiD and CaAlSiD determined from Rietveld profile refinements of 

neutron powder diffraction data from BaAlSiD measured at Kjeller, Institute For Energy 

technology (IFE), Norway. 

Table 5 

Al-H stretching Al-H bending 

E0 [eV] 0.0752 0.0542 

E1 [eV] 0.2132 0.1635 

E2 [eV] 0.3341 0.2742 

ω1 = (E1-E0)/ħ [cm -1 ] 1112 880 

ω2 = (E1-E0)/ħ [cm -1 ] 2087 1772 

Δ=2ω1-ω2 138 -12 

Numerically calculated energies for the ground state E0, the first, E1, and the second, E2, 

excited states, and the normal-mode frequencies ω1 and ω2 using the potential energy curvefit 

polynomials from figure 5. The frequency shift Δ is related to the deviation from the 

harmonic approximation. 

Fig 1

Fig 2 

Fig 3

* * * 

* * * 

Fig 4 



SrAlSiH 

500 1000 1500 2000 2500 

Energy transfer [cm -1 ]

DOS 

10 

8 

6 

4 

2 

DOS 

8 

6 

4 

2 

Fig 5 

0 

-1.0 -0.5 0 0.5 1.0 

E-E F [eV] 


Sr Al Si H 


0 

-10 -8 -6 -4 -2 0 2 

E-E F [eV]

Supporting information 

Characterisation of two new Zintl phase hydrides BaAlSiH and CaAlSiH 

T. Björling*, B. Hauback, T. Utsumi, D. Moser, D. Noréus*, U. Häussermann*** 

S 1 

BaAlSiH CaAlSiH SrAlSiH 

space group, Z P3m1, 1 P3m1, 1 P3m1, 1 

a, Å 4.3381 4.1471 4.2261 

b, Å 4.3381 4.1471 4.2261 

c, Å 5.2289 4.7573 4.9646 

V, Å 3 85.22 70.86 76.79 

density, g/cm 3 3.77 2.25 3.11 

Data obtained from the VASP calculation program. 

S 2 

Atom X Y Z 

Ba 0 0 0.0045 

Al 2/3 1/3 0.5412 

Si 1/3 2/3 0.4629 

D 2/3 1/3 0.8739 

Atomic positions obtained from the VASP calculation program 

S 3 

Atom X Y Z 

Ca 0 0 0.004 

Al 2/3 1/3 0.546 

Si 1/3 2/3 0.436 

D 2/3 1/3 0.914 

Atomic positions obtained from the VASP calculation program 

S 4 

Atom X Y Z 

Sr 0 0 0.001 

Al 2/3 1/3 0.541 

Si 1/3 2/3 0.447 

D 2/3 1/3 0.894 

Atomic positions obtained from the VASP calculation program.

S 5 


Ba—H 2.596 Å 

Al—H 1.740 Å 

Al—Si 2.538 Å 

Al—Si—Al 117.45° 

Si—Al—Si—Al 31.2° 

Al—Si—Al—Si -31.2° 

Shortest distances and angles calculated from theoretical data in S 1 and S 2. 

S 6 

SrAlSiH 

Sr—H 2.497 Å 

Al—H 1.754 Å 

Al—Si 2.485 Å 

Al—Si—Al 116.50° 




S 7 


Ca—H 2.433 Å 

Al—H 1.750 Å 

Al—Si 2.451 Å 

Al—Si—Al 115.59° 




S 8 

BaAlSiD CaAlSiD 

Ae—D 2.578(3) Å 2.420(4) 

Al—D 1.73(2) Å 1.75(3) 

Si—D 3.32(1) Å 3.22 (2) 

Al—Si 2.531(3) Å 2.420(4) 

Shortest distances between the atoms in one unit cell obtained from experimental data.

S 9 

S 10 

A 

k z 

M 

L 

H 

Γ k y 

k x 

Brillouin Zone of HCP-lattice 

Tauc Plot of BaAlSiH. The square (n=2) of the optical-absorption coefficient vs the photon 

energy. The dotted line shows the Tauc extrapolatio.[1]. 

K

S 11 

Tauc Plot of CaAlSiD. The square (n=2) of the optical-absorption coefficient vs the photon 

energy. The line shows the Tauc extrapolation. 

[1] Tauc, R. Grigorovici and A. Vancu, Phys. Stat. 1966, Sol. 15, 627.

S 12 

Potential energy curve for BaSiAlH. The potential curve show the anharmonic behavior when 

the H atom is displaced along the Al-H bonding axis (the z axis). 

S 13 

E-E F [eV] 

Energy [eV] 

5 

0 

-5 

-10 

Γ 

0.8 

0.6 

0.4 

0.2 

0.0 

K 

M 


Γ 

x-direction 

y-direction 

z-direction 

-0.3 0.0 0.3 

Displacement of H atom [Å] 

A 

L 

H 

A 

E-E F [eV] 

5 

0 

-5 

-10 

Γ 

K 

M 


Electronic band structure of the AeAlSiH systems showing the indirect nature of the band 

gap. 

Γ 

A 

L 

H 

A

Ae = Ca, Sr and Ba - Department of Materials and Environmental ...

You also want an ePaper? Increase the reach of your titles

Delete template?

Save as template?