View PDF VersionPrevious ArticleNext Article
 
DOI: 10.1039/C3RA46226A (Paper) RSC Adv., 2014, 4, 4736-4742

Halide substitution in Ca(BH4)2†

Hilde Grovea, Line H. Rudeb, Torben R. Jensenb, Marta Cornoc, Piero Ugliengoc, Marcello Bariccoc, Magnus H. Sørby*a and Bjørn C. Haubacka
aInstitute for Energy Technology, P.O. Box 40, Kjeller, NO-2027, Norway. E-mail: magnuss@ife.no; Fax: +47 63 81 09 20; Tel: +47 63 80 60 00
bCenter for Materials Crystallography (CMC), Interdisciplinary Nanoscience Center (iNANO), and Department of Chemistry, Aarhus University, Langelandsgade 140, DK-8000 Århus C, Denmark
cDipartimento di Chimica and NIS, Università di Torino, Via P. Giuria 7, I-10125 Torino, Italy

Received 29th October 2013 , Accepted 29th November 2013

First published on 2nd December 2013

Abstract

Halide substitution in Ca(BH4)2 has been investigated in ball milled mixtures of Ca(BH4)2 and CaX2 (X = F, Cl, Br) with different molar ratios. In situ synchrotron radiation powder X-ray diffraction measurements of Ca(BH4)2 + CaCl2 with 1[thin space (1/6-em)]:[thin space (1/6-em)]0.5, 1[thin space (1/6-em)]:[thin space (1/6-em)]1 and 1[thin space (1/6-em)]:[thin space (1/6-em)]2 molar ratios reveal that no substitution of Cl− for BH4− occurs from the ball milling process. However, substitution readily occurs after the transitions from α- to β-Ca(BH4)2 and from orthorhombic to tetragonal CaCl2 upon heating above ∼250 °C, which is evident from both contraction of the unit cell and changes in the relative Bragg peak intensities, in agreement with theoretical calculations. Rietveld analyses of the obtained β-Ca((BH4)1−xClx)2 solid solutions indicate compositions from x = 0 to 0.6, depending on the amount of CaCl2 in the parent mixtures. β-Ca((BH4)0.5Cl0.5)2 was investigated by differential scanning calorimetry and has a slightly higher decomposition temperature compared to pure Ca(BH4)2. No substitution with CaF2 or CaBr2 is observed.

Introduction

The efforts to find fuels that are non-polluting and not contributing to the greenhouse effect is on-going. Hydrogen is a clean energy carrier, releasing only water when used in fuel cells. Hydrogen is a gas with a low boiling point and consequently the challenge for hydrogen to become a widely used energy carrier is an efficient storage system. It is possible to store hydrogen gas under pressure, but a safer and much more compact way is to store hydrogen chemically bonded in solid compounds.1 The storage material should have both high gravimetric and volumetric hydrogen content, good cycling ability and fast hydrogen sorption kinetics. NaAlH4 with titanium-based additives was the first complex metal hydrides to be considered for hydrogen storage,2 but the practical gravimetric hydrogen storage capacity is less than 4.5 wt%.3 Metal borohydrides have high gravimetric hydrogen density and are thus of interest as hydrogen storage materials.4,5 Ca(BH4)2 can theoretically store up to 11.6 wt% hydrogen and the volumetric hydrogen content is 108 g H per L. DFT calculation estimates an equilibrium H2 pressure of 1 bar at temperatures below 100 °C,6 which is ideal for hydrogen storage for mobile applications. Experimentally, however, the dehydrogenation process occurs between 360 and 500 °C.6–8 The hydrogen sorption is partially reversible at 90 bar and temperatures from 350 to 420 °C.7 Ca(BH4)2 is generally observed to decompose to CaB6 and CaH2, according to the reaction 3Ca(BH4)2 → CaB6 + 2CaH2 + 10H2, but the process is complicated and several intermediate phases are involved.9–12

The observed dehydrogenation temperature is too high for practical purposes,1 thus the stability has to be reduced for Ca(BH4)2 to become an effective energy carrier. Brinks et al. have shown that the stability of Na3AlH6 can be considerably changed by partly substituting hydrogen with fluorine.13 Several studies have shown that that heavier halides can substitute BH4− in borohydrides, e.g. Cl−, Br−, and I− substitution in LiBH4,14–16 and Cl− substitution in NaBH4,17 Mg(BH4)218 and Ca(BH4)2.19,20 Recently, three new compounds have been reported in the Ca(BH4)2–CaI2 system.21 F− is reported to substitute H− in NaBH4 (ref. 22) and KBH4.23

Several structural modifications of Ca(BH4)2 have been described in the literature.6,24–30 α-Ca(BH4)2 (orthorhombic, space group F2dd) is the stable polymorph at room temperature.25,31 It transforms to tetragonal α′-Ca(BH4)2 through a second-order transformation around 220 °C and further completely to β-Ca(BH4)2 upon heating above 300 °C.25 The β-phase crystallizes with tetragonal symmetry first described in space group P42/m26 and later revised to P[4 with combining macron].25,29,32,33 γ-Ca(BH4)2 is sometimes formed at room temperature. This phase is metastable and crystallizes in the orthorhombic space group Pbca.10,31

The purpose of this work is to study possible substitution of F−, Cl− and Br− in Ca(BH4)2. The most likely substitute for hydrogen is fluorine, because of the similarity in size. On the other hand, the BH4−-ion has similar size (2.05 Å) to both the Cl−-ion (1.81 Å) and the Br−-ion (1.96 Å). These halides might therefore substitute the whole BH4− group.

Experimental and theoretical methods

Preparation of Ca(BH4)2 + CaX2 samples

Ca(BH4)2·2THF (purchased from Sigma Aldrich) was dried in vacuum (p ≈ 1 × 10−1 mbar) at 160 °C for 24 hours to remove the THF adducts, which yielded a mixture of the α and β polymorphs of Ca(BH4)2 after cooling to ambient temperature. The calcium halides, CaX2, X = F, Br (Aldrich, purity: 99.99%, 99.98%, respectively) and X = Cl (Alfa Aesar, purity: 99.9%) were used as-purchased.

The solvent-free Ca(BH4)2 was mixed with CaX2 in molar ratios of 1[thin space (1/6-em)]:[thin space (1/6-em)]0.5 (X = F, Cl), 1[thin space (1/6-em)]:[thin space (1/6-em)]1 (X = Cl, Br) or 1[thin space (1/6-em)]:[thin space (1/6-em)]2 (X = Cl) and ball milled for 2 hours (2 min milling, 2 min pause, times 60) under argon atmosphere with a ball-to-powder ratio of approximately 30[thin space (1/6-em)]:[thin space (1/6-em)]1 using tungsten carbide bowl (15 mL) and balls (10 mm) in a Fritsch Pulverisette P4. The bowl was sealed with a lid and a polymer O-ring.

All handling of the materials was done in purified Ar atmosphere (<1 ppm O2 and H2O) in an MBraun glove box.

In situ synchrotron radiation powder X-ray diffraction

In situ synchrotron radiation powder X-ray diffraction (SR-PXD) was performed at beamline BM01A at the Swiss-Norwegian Beamlines (SNBL) at European Synchrotron Radiation Facility (ESRF) in Grenoble, France. A MAR345 image plate detector was used and the samples were packed in boron glass capillaries (0.5 mm) and connected to a vacuum pump. The X-ray exposure time for each measurement was 30 s and the readout time was 90 seconds, thus a complete SR-PXD diagram was collected every second minute. The wavelength used was λ = 0.703511 Å. The wavelength and detector geometry were calibrated with LaB6 as an external standard. Different heating rates were used, 3 K min−1 was used for all Ca(BH4)2 + CaCl2 samples, 2 and 3 K min−1 were used for other samples.

One-dimensional diffraction patterns were obtained by integrating the 2D images using the Fit2D program.34 Data analyses were performed using the Rietveld method in the FullProf suite.35 The background was interpolated between manually selected points and the Bragg peaks were modeled by a pseudo-Voigt function with 3 refinable parameters.

Thermal analysis

Differential scanning calorimetry (DSC) was performed with a Netzsch STA449C Jupiter instrument at heating rates of 10 K min−1 in a flow of He (50 mL min−1). The samples were contained in Al2O3 crucibles with tiny holes in the lid to prevent pressure build-up during desorption of gases.

Computational details

The calculations were performed applying the periodic quantum-mechanical software CRYSTAL09 within the density functional theory, PBE functional. The CRYSTAL code utilizes localized Gaussian functions to describe electrons. Ca was described by a 86-511d21G all-electron basis set;36 B by a 6-21G(d) basis set (αsp = 0.124 bohr−2 for the most diffuse shell exponent and αpol = 0.800 bohr−2 for polarization); H by a 31G(p) (αsp = 0.1613 bohr−2 for the most diffuse shell exponent and αpol = 1.1 bohr−2 for polarization),37 F by a 7-311G(d),38 Cl by a 86-311G with an additional d orbital (αsp = 0.125 bohr−2 for the most diffuse shell exponent and αd = 0.750 bohr−2)39 and Br by a 6-311G(d) (αd = 0.154 bohr−2 for the most diffuse shell exponent and αpol = 0.451 bohr−2 for polarization).40 Grimme correction was also considered to take the dispersive contributions into account in all the calculations for geometry optimization and energetic description.41 Phonons at Γ point in the harmonic approximation were computed to derive the thermodynamic functions by diagonalizing the associated mass-weighted Hessian matrix.42

Results and discussion

The Ca(BH4)2–CaF2 system

For Ca(BH4)2–CaF2 with molar ratio 1[thin space (1/6-em)]:[thin space (1/6-em)]0.5, the four phases α-, β-, and γ-Ca(BH4)2 and CaF2 are observed in the diffraction pattern of the as-milled sample. The amount of CaF2, compared to the amount of Ca(BH4)2 calculated from the SR-PXD data, indicates that no F-substitution had occurred. During heating, the phase transformation from the α- to the β-modification is observed between 160 and 220 °C, in addition to the changes in unit cell parameters due to the thermal expansion. The intensities corresponding to the Bragg peaks from CaF2 do not decrease and there were no significant changes in the relative intensities of the Bragg peaks from Ca(BH4)2. This indicates that there is no fluorine substitution neither in the α nor in the β polymorph of Ca(BH4)2.

The Ca(BH4)2–CaCl2 system

After ball milling Ca(BH4)2 and CaCl2 in the molar ratios 1[thin space (1/6-em)]:[thin space (1/6-em)]0.5, 1[thin space (1/6-em)]:[thin space (1/6-em)]1 and 1[thin space (1/6-em)]:[thin space (1/6-em)]2, α-Ca(BH4)2 is the most prominent Ca(BH4)2 polymorph, but a residue of the β-phase is also observed by SR-PXD. CaCl2 is observed in its stable modification (space group Pnnm), but it is also present in two other modifications with space groups Pbcn43 and Pbca44 in the milled Ca(BH4)2–CaCl2 samples. The latter is a high pressure CaCl2 polymorph formed from the Pbcn phase during milling. Some additional peaks from minor impurity phases are also observed, but they have been disregarded in the Rietveld refinements.

The sample Ca(BH4)2–CaCl2 1[thin space (1/6-em)]:[thin space (1/6-em)]1 was studied with SR-PXD during heating up to 360 °C, and the results are shown in Fig. 1. Rietveld refinement of the SR-PXD data collected at about 40 °C gives α-Ca(BH4)2 (36 wt%), CaCl2 Pnnm (20 wt%) and CaCl2 Pbca (33 wt%) as the main components, with minor amounts of β-Ca(BH4)2 (2 wt%) and CaCl2 Pbcn (7 wt%). Ca(BH4)2 and CaCl2 in 1[thin space (1/6-em)]:[thin space (1/6-em)]1 molar ratio corresponds to a mass ratio of 39[thin space (1/6-em)]:[thin space (1/6-em)]61. This is close to the refined values when summing all the Ca(BH4)2 phases and CaCl2 phases: 38 wt% and 62 wt%, respectively. This indicates that most of the sample is crystalline and also that no reaction between the compounds has occurred during ball milling. The refined unit cell parameters for α-Ca(BH4)2 at 40 °C are: a = 8.755(6) Å, b = 13.143(6) Å and c = 7.532(5) Å. These are in good agreement with literature values (see Table 1) and the relative intensities of the Bragg peaks from α-Ca(BH4)2 fits well with published crystal structure data,29 thus supporting that no substitution has taken place at this temperature. Substitution would be evident as a decrease in the unit cell size and marked changes in the relative intensities due to the smaller size and higher X-ray scattering cross section for Cl− compared to BH4−.


image file: c3ra46226a-f1.tif
Fig. 1 In situ SR-PXD measured for Ca(BH4)2 + CaCl2 in molar ratio 1[thin space (1/6-em)]:[thin space (1/6-em)]1, heating rate 3 K min−1. The temperature increases from 40 to 360 °C. (a) gives a plot of the evolution of Bragg peak intensities. (b) 3D plot of selected 2θ area. λ = 0.703511 Å.
Table 1 Crystallographic parameters for pure and chloride containing Ca(BH4)2 compounds at different temperatures
Compound Crystal system Space group a/Å b/Å c/Å V/Å3 T/°C
a This work.
α-Ca(BH4)2 (ref. 29) Orthorhombic F2dd 8.745 13.105 7.495 858.95 RT
α-Ca(BH4)2a Orthorhombic F2dd 8.755(6) 13.143(6) 7.532(5) 867(1) 40
β-Ca(BH4)2 (ref. 25) Tetragonal P[4 with combining macron] 6.9189(1)   4.3471(1) 208.1 RT
β-Ca(BH4)2 (ref. 32) Tetragonal P[4 with combining macron] 6.9509(5)   4.3688(3) 211.1(2) 160
β-Ca(BH4)2 (ref. 45) Tetragonal P[4 with combining macron] 7.005(5)   4.414(7) 216.6(4) 350
β-Ca((BH4)0.7Cl0.3)2a Tetragonal P[4 with combining macron] 6.7725(6)   4.3074(6) 197.6(1) 350
β-Ca((BH4)0.5Cl0.5)2a Tetragonal P[4 with combining macron] 6.673(4)   4.268(2) 190.0(2) 350
β-Ca((BH4)0.4Cl0.6)2a Tetragonal P[4 with combining macron] 6.542(3)   4.214(2) 180.3(1) 350

The phase transformation from α-Ca(BH4)2 to β-Ca(BH4)2 takes place in the temperature range 150–200 °C, see Fig. 1. At about 260 °C a shoulder starts to appear on the high angle side of the β-Ca(BH4)2 peaks. Upon further heating, the diffraction peaks from the β-polymorph is gradually shifted to higher scattering angle, indicating contraction of the unit cell. This is in agreement with substitution of BH4− with the smaller Cl−. The following reaction seems to take place:

β-Ca(BH4)2 + xCaCl2 → (1 + x)β-Ca((BH4)1−xClx)2

Furthermore, from the Rietveld refinement the relative intensities of the peaks for the proposed β-Ca((BH4)1−xClx)2 do not match the diffraction pattern for pure β-Ca(BH4)2. Exchanging some of the BH4− with Cl− in the Rietveld refinement of β-Ca(BH4)2, results in a better fit to the observed data. Fig. 2 shows the refinement for substituted Ca((BH4)1−xClx)2 giving x = 0.53(3) at 350 °C.


image file: c3ra46226a-f2.tif
Fig. 2 Rietveld refinement for SR-PXD data of ball milled Ca(BH4)2 and CaCl2 in molar ratio 1[thin space (1/6-em)]:[thin space (1/6-em)]1 at 350 °C. Three phases are refined: CaCl2 P42/mnm (top tick marks), substituted β-Ca((BH4)1−xClx)2 with x = 0.5 (middle) and CaCl2 Pbcn (bottom tick marks). Experimental data are shown as dots, and the calculated profile is a solid line. The bottom line shows the difference plot. λ = 0.703511 Å.

The onset of chloride substitution in β-Ca(BH4)2 follows shortly after the reported phase transformation temperature of CaCl2 from the orthorhombic Pnnm phase to a tetragonal phase described in space group P42/mnm at 235 °C.46 The transition is of the second-order and proceeds by contraction of the a-axis and expansion of the b-axis in the orthorhombic phase until they become identical at the phase transition temperature. It is interesting to note that this high-temperature phase of CaCl2 (HT-CaCl2) can be regarded as isostructural to β-Ca(BH4)2 if the non-spherical symmetry of BH4− is disregarded. The difference in space group, P42/mnm vs. P[4 with combining macron], is due to the tetrahedral symmetry of the BH4− anion compared to spherical symmetry of Cl−. The phase transformation of CaCl2 induce only subtle changes in the diffraction pattern which are not directly observable in the present data, due to extensive peak overlap with Ca(BH4)2. However, HT-CaCl2 was used in Rietveld refinements of data collected above the transition temperature, resulting in good fits to the data.

For the sample with molar ratio 1[thin space (1/6-em)]:[thin space (1/6-em)]1, the refinements show that there is still 15 wt% of CaCl2 left in the sample at 350 °C, suggesting that the substituted fraction x in Ca(BH4)2 must be somewhat smaller than 0.5. Specifically, calculation of the degree of substitution from the weight fractions of the different phases gives x = 0.45(5). For the 1[thin space (1/6-em)]:[thin space (1/6-em)]2 and 1[thin space (1/6-em)]:[thin space (1/6-em)]0.5 samples at 350 °C, the Rietveld refinements of anion site occupancies give compositions x = 0.81(1) and x = 0.46(2), respectively. From the refined phase fractions x should be 0.60(6) for the 1[thin space (1/6-em)]:[thin space (1/6-em)]2 sample and 0.30(3) for the 1[thin space (1/6-em)]:[thin space (1/6-em)]0.5 sample. Thus, the refined ratios of the BH4− and Cl− in the three substituted phases all overestimate the amounts of Cl− compared to the nominal compositions. This could be due to the high anisotropy of BH4−, making it difficult to determine exact positions of the H atoms and, consequently, making refinement less stable. The compositions estimated from the phase fractions, i.e. from the ratio of unreacted CaCl2 and substituted Ca(BH4)2, are deemed to be more reliable. Hence, the compositions of the three substituted phases are given with one decimal precision as Ca((BH4)0.5Cl0.5)2 (x = 0.5), Ca((BH4)0.4Cl0.6)2 (x = 0.6) and Ca((BH4)0.7Cl0.3)2 (x = 0.3) for the samples 1[thin space (1/6-em)]:[thin space (1/6-em)]1, 1[thin space (1/6-em)]:[thin space (1/6-em)]2 and 1[thin space (1/6-em)]:[thin space (1/6-em)]0.5, respectively.

Table 1 shows significant contraction in unit cell dimensions for all Cl-substituted β-Ca(BH4)2 sample. The changes are anisotropic with larger changes in the a-axis than the c-axis. This is can be explained by a greater difference in the a-axis than the c-axis between pure β-Ca(BH4)2 (a = 7.0 Å, c = 4.4 Å at 350 °C) and HT-CaCl2 (a = 6.4 Å, c = 4.2 Å at 350 °C).46 The changes in unit cell parameters are approximately proportional to the amount of Cl substitution into the lattice, in agreement with Vegard's law.47

DSC measurements were performed for pure Ca(BH4)2 and Ca(BH4)2–CaCl2 ball milled in molar ratio 1[thin space (1/6-em)]:[thin space (1/6-em)]1, respectively, see Fig. 3. For the pure Ca(BH4)2, there is a sharp endothermic peak in the DSC data at about 370 °C. This is consistent with release of hydrogen from Ca(BH4)2. The same feature is seen in the 1[thin space (1/6-em)]:[thin space (1/6-em)]1 Ca(BH4)2–CaCl2 sample, but the temperature of the event is increased to 390 °C, indicating a slight stabilization of the Cl-containing compound. SR-PXD data show that CaHCl is a decomposition product. This is in agreement with our recent report that the Ca(BH4)2–CaI2 decompose to CaHI.21 No other crystalline decomposition products are observed in the present study.


image file: c3ra46226a-f3.tif
Fig. 3 DSC data for Ca(BH4)2 (dashed) and Ca(BH4)2–CaCl2 (1[thin space (1/6-em)]:[thin space (1/6-em)]1) (solid) with heating rate 10 K min−1.

The Ca(BH4)2–CaBr2 system

Ca(BH4)2 and CaBr2 were ball milled in the molar ratio 1[thin space (1/6-em)]:[thin space (1/6-em)]1. CaBr2 (V = 196 Å3) is isostructural with the Pnnm modification of CaCl2 (168 Å3). When the Ca(BH4)2–CaBr2 sample is heated, the SR-PXD data shows the transformation from α- to β-Ca(BH4)2 in the temperature range 150–200 °C. There are no significant changes in unit cell parameters beside thermal expansion, no change in relative intensities of the β-Ca(BH4)2 diffraction peaks and the intensities of the CaBr2 peaks do not decrease. This indicates that no appreciable amount of Br− is substituted into Ca(BH4)2, which is confirmed with Rietveld refinement showing a good fit without any substitution. The lack of substitution can be explained by the high transition temperature of the transition from Pnnm to P42/mnm for CaBr2, which occurs at 553 °C.46 Thus, Ca(BH4)2 and CaBr2 do not have any isostructural relationship in the investigated temperature range, which makes substitution less favorable.

The in situ SR-PXD measurement shows that CaHBr is a decomposition product of the 1[thin space (1/6-em)]:[thin space (1/6-em)]1 Ca(BH4)2 + CaBr2 mixture above 360 °C. No other crystalline decomposition products are observed.

DSC measured on the 1[thin space (1/6-em)]:[thin space (1/6-em)]1 ratio of Ca(BH4)2–CaBr2 (not shown) reveal an increase in decomposition temperature which is comparable to that observed for the Cl-substituted sample, even though Br substitution is not taking place. This observation indicates that changed reaction paths, evident from the decomposition products CaHCl and CaHBr, could have a larger impact on the decomposition temperature than the actual halide substitution.

Theoretical calculations

Theoretical modelling has been used to estimate the solubility of F−, Cl− and Br− into β-Ca(BH4)2. A single unit cell of β-Ca(BH4)2 with the P[4 with combining macron] symmetry has been considered, and the four BH4− units have been progressively substituted by a halide in the case of Cl− and Br−, while F substituted H-atoms rather than whole BH4− groups. Since the unit cell of the pure β-Ca(BH4)2 (Ca2B4H16) contains 16 H-atoms, a very large number of symmetry-non-equivalent configurations should be considered for single H-to-F substitutions in the BH4− units. So, F-substitution has been conducted following the same procedure described for orthorhombic LiBH4,48 where a “locality principle” was established. It simply states that the lowest energy is obtained when all four H-atoms are substituted by F-atoms in the BH4− units. As a consequence, only BF4− to BH4− unit substitutions have been considered. A single substitution corresponds to a fraction x = 0.25, two substitutions to x = 0.50 and, finally, three substitution to x = 0.75. It has to be noticed that for x = 0.5, two configurations are possible. In this case, the solution with the lowest energy was considered. The results, expressed in terms of enthalpy of mixing as a function of composition, are reported in Fig. 4.
image file: c3ra46226a-f4.tif
Fig. 4 Computed enthalpy of mixing for solid solution of beta calcium borohydride, where BH4− units have been progressively substituted by BF4−, Cl− and Br−. All values are per formula unit.

It is clear that BF4− substitution leads to a significant positive enthalpy of mixing, suggesting strong immiscibility. On the other hand, both Cl− and Br− substitutions appear more favorable, giving an enthalpy of mixing close to zero. As shown in Fig. 4, the enthalpy of mixing in the case of x = 0.75 gives a negative value around −3 kJ per mol per formula unit for both Cl− and Br− substitutions. The two cases are however very different with respect to variation of the unit cell volume, as shown in Fig. 5, where the calculated volume variation of the unit cell is shown as a function of halide substitution. A decrease in volume is indeed computed for Cl− substitution, in line with the experimental trend. Calculations suggest that a maximum volume change occurs at x = 0.75. Calculated values are underestimated, with respect to experimental results, by a few percent points, due to systematic errors of the adopted functionals and basis set or to neglecting temperature effects in the calculations. Calculated volume variations for Br− substitution are much smaller in comparison to those obtained for Cl− substitution, giving an almost constant volume as a function of composition. This behavior can be easily rationalized on the basis of the ionic radius of Br−, which is much closer to BH4− than Cl−. In turn, the significant volume decrease for the Cl− substituted structures forces the BH4 group to rotate in such a way that the H–H intermolecular repulsive contacts are minimized. This effect does not occur for the Br substituted structures, as the volume change is minor compared to pure Ca(BH4)2 and, consequently, the BH4 groups maintain their pristine orientation. The computed structural data (not reported here for brevity) shows that the unit cell contraction for x = 0.5 Cl− is anisotropic (see Table S1 in ESI†) with larger change in a- than c-axis, in agreement with the experiments (see Table 1).


image file: c3ra46226a-f5.tif
Fig. 5 Variations of the unit cell volume for solid solution of β-Ca(BH4)2, where BH4− units have been progressively substituted by Cl− (circles) and Br− (squares). Continuous lines (computed), dashed line (experimental).

In order to estimate the solubility of Cl− and Br− in β-Ca(BH4)2 from mixtures with the corresponding halides, thermodynamics of CaCl2 and CaBr2 compounds have to be considered. For both CaCl2 and CaBr2, the most stable structure at T = 25 °C has the Pnnm symmetry. To estimate the effect of halide substitution, the free energy of CaCl2 and CaBr2 structures have been computed considering, as a reference, the high-temperature structures, crystallizing with the P42/mnm symmetry.46 In order to be coherent, the thermodynamic data have been recalculated considering the two structures within the P[4 with combining macron] symmetry (subgroup of P42/mnm) of the reference β-Ca(BH4)2 structure, obtaining very similar results (see Tables S1 and S2 in ESI†). The predicted structures of the two high temperature phases are in good agreement with the experimental data, as shown in Table S2 in ESI.† Calculations at T = 25 °C on CaCl2 give ΔH = −6.8 kJ per mol per formula unit and ΔS = 0.5 J per mol per K per formula unit, respectively, for the phase transition from the P42/mnm to the Pnnm structure. Calculations on CaBr2 led to a lower enthalpy change, corresponding to a value of ΔH = −3.0 kJ per mol per formula unit, but a much higher entropy change, equal to ΔS = −5.3 J per mol per K per formula unit. This large variation of entropy for CaBr2 compared to CaCl2 may result from the fact that very low vibrational frequencies, which dominate the entropy value, are much smaller for CaBr2 than for CaCl2, due to the higher mass of Br. The calculated values of enthalpy and entropy should imply a phase transition from orthorhombic (Pnnm) to tetragonal (P42/mnm) symmetry at about 297 °C for CaBr2, underestimated with respect to the experimental transition temperature of 553 °C. On the contrary, a higher stability of Pnnm with respect to the P42/mnm structure is predicted at all temperatures for the CaCl2 crystal.

According to thermodynamic calculations, a similar behavior is expected for Br− and Cl− substituted solid solutions. In particular, considering ideal entropy of mixing, a negative free energy of mixing is expected at about 350 °C. In this case, because of the low value of the enthalpy of mixing calculated for x = 0.5, a two-phase equilibrium between Cl-poor, i.e. Ca[(BH4)0.75Cl0.25]2, and Cl-rich, i.e. Ca[(BH4)0.25Cl0.75]2, solid solutions should occur for 0.25 < x < 0.75. This trend suggests a specific stability of terminal solid solutions, likely because of a combination of geometric and electronic effects. Similar trend is estimated for Br substitutions. On the other hand, the presence of CaCl2 and CaBr2 (i.e. x = 1) in their stable structures could also limit the formation of halide-rich solid solutions. In addition, the calculated closer stability of the P42/mnm structure with respect to the Pnnm for CaBr2 compared to CaCl2 would suggest an easier solubility. The reasons for the disagreement between the calculations and experimental findings are not clear. It could be that kinetic effects, related to ionic size, play a crucial role during experiments. The discrepancies could also be due to uncertainties in the calculations, as the absence of the predicted Pnnm to P42/mnm phase transformation for CaBr2 in the experimentally explored temperature range, can hinder the predicted solubility. It is worth noting that a stabilization of β-Ca((BH4)1−xClx)2 solid solutions with respect to pure β-Ca(BH4)2 reduces the driving force for the transformation into dehydrogenated products, and thus possibly increasing the dehydrogenation temperature, as observed experimentally (see Fig. 3).

The concept of anion substitution in borohydrides remains not fully explored, however, some trends in the structural chemistry are revealed from this work and the literature. For substitution of borohydrides with the heavier halides, (i) the solid containing the smaller anion, e.g. CaCl2, tends to dissolve into the compound containing the larger anion, β-Ca(BH4)2, and the structure of the latter tends to be preserved in the obtained solid solution. This trend can be interpreted as an increase in the lattice energy due to the clearly observed decrease in the unit cell volume, which may create an internal ‘chemical pressure’.49 (ii) Some polymorphs of metal borohydrides are more prone to perform anion substitution, than others, e.g. β-Ca(BH4)2 and not α-Ca(BH4)2 as shown here. (iii) This work also highlights that isomorphism may be more important to facilitate anion substitution as compared to similarities in anion radii. This explains that CaBr2 did not dissolve in Ca(BH4)2, despite the fact that r(BH4−) ∼ r(Br−). In fact, in some cases anion substitution may occur in both compounds, which is previously observed for the LiBH4–LiBr, LiBH4–LiI and NaBH4–NaCl systems, possibly due to the fact that β-LiBr, β-LiI and h-LiBH4 as well as NaBH4 and NaCl are isostructural.50–53

Conclusions

Possible halide substitutions in Ca(BH4)2 have been investigated by experimental and theoretical (ab initio) methods. For mixtures with CaX2, X = F, Cl and Br, substitution is only observed for CaCl2, and no substitution with CaF2 and CaBr2 was found.

The absence of solid solubility in the Ca(BH4)2–CaF2 system is rationalized by a positive calculated enthalpy of mixing. Substitution with Cl− is only observed into the β-modification of Ca(BH4)2 after heating the ball milled sample above 250 °C which is above the orthorhombic-to-tetragonal phase transformation temperature for CaCl2, while no substitution is observed in α-Ca(BH4)2. This is explained by the isostructural relationship between β-Ca(BH4)2 and the tetragonal high-temperature modification of CaCl2. The resulting phase Ca(BH4)1−xClx has been observed with x from 0 to 0.6, depending on the Ca(BH4)2[thin space (1/6-em)]:[thin space (1/6-em)]CaCl2 ratio in the initial mixture. However, it is possible that there is full solubility between the two phases due to their isostructural relationship. The decomposition temperature of β-Ca((BH4)0.5Cl0.5)2 was found to be slightly increased compared to pure Ca(BH4)2, likely due to a reduction of driving force for the dehydrogenation reaction. Br− substitution in β-Ca(BH4)2 is predicted by theoretical calculations, but it is not observed experimentally. This is explained by the lack of orthorhombic-to-tetragonal phase transition in the experimental temperature range even though it is predicted by the calculations.

Acknowledgements

This work was financially supported by the European Commission FP7 project FLYHY (grant no. 226943), the Fuel Cell and Hydrogen Joint Undertaking (FCH JU) project BOR4STORE (grant no. 303428) and the RENERGI and SYNKNØYT programs of the Research Council of Norway. The skilful assistance from the project team at the Swiss-Norwegian Beam Line at the ESRF is gratefully acknowledged.

References

  1. L. Schlapbach and A. Zuttel, Nature, 2001, 414, 353–358 CrossRef CAS PubMed.
  2. B. Bogdanovic and M. Schwickardi, J. Alloys Compd., 1997, 253–254, 1–9 CrossRef CAS.
  3. J. M. Bellosta von Colbe, O. Metz, G. A. Lozano, P. K. Pranza, H. W. Schmitz, F. Beckmann, A. Schreyer, T. Klassen and M. Dornheim, Int. J. Hydrogen Energy, 2012, 37, 2807–2811 CrossRef CAS PubMed.
  4. H.-W. Li, Y. Yan, A. Orimo, S. Züttel and C. M. Jensen, Energies, 2011, 4, 185–214 CrossRef CAS.
  5. L. H. Rude, T. K. Nielsen, D. B. Ravnsbæk, U. Bösenberg, M. B. Ley, B. Richter, L. M. Arnbjerg, M. Dornheim, Y. Filinchuk, F. Besenbacher and T. R. Jensen, Phys. Status Solidi A, 2011, 208, 1754–1773 CrossRef CAS.
  6. K. Miwa, M. Aoki, T. Noritake, N. Ohba, Y. Nakamori, S. Towata, A. Zuttel and S. Orimo, Phys. Rev. B: Condens. Matter Mater. Phys., 2006, 74, 155122 CrossRef.
  7. J.-H. Kim, S.-A. Jin, J.-H. Shim and Y. W. Cho, Scr. Mater., 2008, 58, 481–483 CrossRef CAS PubMed.
  8. E. Rønnebro and E. H. Majzoub, J. Phys. Chem. B, 2007, 111, 12045–12047 CrossRef PubMed.
  9. Y. Kim, S.-J. Hwang, J.-H. Shim, Y.-S. Lee, H. N. Han and Y. W. Cho, J. Phys. Chem. C, 2012, 116, 4330–4334 CAS.
  10. M. D. Riktor, M. H. Sørby, K. Chlopek, M. Fichtner, F. Buchter, A. Zuettel and B. C. Hauback, J. Mater. Chem., 2007, 17, 4939–4942 RSC.
  11. M. D. Riktor, Y. Filinchuk, P. Vajeeston, E. G. Bardaji, M. Fichtner, H. Fjellvag, M. H. Sørby and B. C. Hauback, J. Mater. Chem., 2011, 21, 7188–7193 RSC.
  12. I. Llamas-Jansa, O. Friedrichs, M. Fichtner, E. G. Bardaji, A. Züttel and B. C. Hauback, J. Phys. Chem. C, 2012, 116, 13472–13479 CAS.
  13. H. W. Brinks, A. Fossdal and B. C. Hauback, J. Phys. Chem. C, 2008, 112, 5658–5661 CAS.
  14. H. Maekawa, M. Matsuo, H. Takamura, M. Ando, Y. Noda, T. Karahashi and S. I. Orimo, J. Am. Chem. Soc., 2009, 131, 894–895 CrossRef CAS PubMed.
  15. L. Mosegaard, B. Moller, J. E. Jorgensen, Y. Filinchuk, Y. Cerenius, J. C. Hanson, E. Dimasi, F. Besenbacher and T. R. Jensen, J. Phys. Chem. C, 2008, 112, 1299–1303 CAS.
  16. L. M. Arnbjerg, D. B. Ravnsbæk, Y. Filinchuk, R. T. Vang, Y. Cerenius, F. Besenbacher, J. E. Jorgensen, H. J. Jakobsen and T. R. Jensen, Chem. Mater., 2009, 21, 5772–5782 CrossRef CAS.
  17. J. E. Olsen, M. H. Sørby and B. C. Hauback, J. Alloys Compd., 2011, 509, L228–L231 CrossRef CAS PubMed.
  18. S. Hino, J. E. Fonnelop, M. Corno, O. Zavorotynska, A. Damin, B. Richter, M. Baricco, T. R. Jensen, M. H. Sørby and B. C. Hauback, J. Phys. Chem. C, 2012, 116, 12482–12488 CAS.
  19. C. Rongeat, I. Lindemann, A. Borgschulte, L. Schultz and O. Gutfleisch, Int. J. Hydrogen Energy, 2011, 36, 247–253 CrossRef CAS PubMed.
  20. J. Y. Lee, Y. S. Lee, J. Y. Suh, J. H. Shim and Y. W. Cho, J. Alloys Compd., 2010, 506, 721–727 CrossRef CAS PubMed.
  21. L. H. Rude, Y. Filinchuk, M. H. Sørby, B. C. Hauback, F. Besenbacher and T. R. Jensen, J. Phys. Chem. C, 2011, 115, 7768–7777 CAS.
  22. L. H. Rude, U. Filsø, V. D'Anna, A. Spyratou, B. Richter, S. Hino, O. Zavorotynska, M. Baricco, M. H. Sørby, B. C. Hauback, H. Hagemann, F. Besenbacher, J. Skibsted and T. R. Jensen, Phys. Chem. Chem. Phys., 2013, 15, 18185–18194 RSC.
  23. R. H. Heyn, I. Saldan, M. H. Sørby, C. Frommen, B. Arstad, A. M. Bougza, H. Fjellvag and B. C. Hauback, Phys. Chem. Chem. Phys., 2013, 15, 11226–11230 RSC.
  24. Y. Filinchuk, D. Chernyshov and V. Dmitriev, Z. Kristallogr., 2008, 223, 649–659 CrossRef CAS.
  25. Y. Filinchuk, E. Rønnebro and D. Chandra, Acta Mater., 2009, 57, 732–738 CrossRef CAS PubMed.
  26. F. Buchter, Z. Lodziana, A. Remhof, O. Friedrichs, A. Borgschulte, P. Mauron, A. Zuttel, D. Sheptyakov, G. Barkhordarian, R. Bormann, K. Chlopek, M. Fichtner, M. Sørby, M. Riktor, B. Hauback and S. Orimo, J. Phys. Chem. B, 2008, 112, 8042–8048 CrossRef CAS PubMed.
  27. M. Aoki, K. Miwa, T. Noritake, N. Ohba, M. Matsumoto, H. W. Li, Y. Nakamori, S. Towata and S. Orimo, Appl. Phys. A: Mater. Sci. Process., 2008, 92, 601–605 CrossRef CAS PubMed.
  28. G. Barkhordarian, T. R. Jensen, S. Doppiu, U. Bosenberg, A. Borgschulte, R. Gremaud, Y. Cerenius, M. Dornheim, T. Klassen and R. Bormann, J. Phys. Chem. C, 2008, 112, 2743–2749 CAS.
  29. Y. S. Lee, Y. Kim, Y. W. Cho, D. Shapiro, C. Wolverton and V. Ozolins, Phys. Rev. B: Condens. Matter Mater. Phys., 2009, 79, 104107 CrossRef.
  30. F. Buchter, Z. Lodziana, A. Remhof, O. Friedrichs, A. Borgschulte, P. Mauron, A. Zuttel, D. Sheptyakov, L. Palatinus, K. Chlopek, M. Fichtner, G. Barkhordarian, R. Bormann and B. C. Hauback, J. Phys. Chem. C, 2009, 113, 17223–17230 CAS.
  31. T. Noritake, M. Aoki, M. Matsumoto, K. Miwa, S. Towata, H. W. Li and S. Orimo, J. Alloys Compd., 2010, 491, 57–62 CrossRef CAS PubMed.
  32. T. J. Frankcombe, J. Phys. Chem. C, 2010, 114, 9503–9509 CAS.
  33. E. H. Majzoub and E. Rønnebro, J. Phys. Chem. C, 2009, 113, 3352–3358 CAS.
  34. A. P. Hammersley, Internal Report ESRF-98-HA01T, 1998.
  35. J. Rodríguez-Carvajal, Physica B, 1993, 192, 55 CrossRef.
  36. L. Valenzano, F. J. Torres, D. Klaus, F. Pascale, C. M. Zicovich-Wilson and R. Dovesi, Z. Phys. Chem., 2006, 220, 893–912 CrossRef CAS.
  37. R. Dovesi, C. Ermondi, E. Ferrero, C. Pisani and C. Roetti, Phys. Rev. B: Condens. Matter Mater. Phys., 1984, 29, 3591–3600 CrossRef CAS.
  38. R. Nada, C. R. A. Catlow, C. Pisani and R. Orlando, Model. Numer. Simul. Mater. Sci., 1993, 1, 165–187 CrossRef CAS.
  39. E. Apra, M. Causa, M. Prencipe, R. Dovesi and V. R. Saunders, J. Phys.: Condens. Matter, 1993, 5, 2969–2976 CrossRef CAS.
  40. L. A. Curtiss, M. P. McGrath, J. P. Blaudeau, N. E. Davis, R. C. Binning and L. Radom, J. Chem. Phys., 1995, 103, 6104–6113 CrossRef CAS PubMed.
  41. S. Grimme, J. Comput. Chem., 2006, 27, 1787–1799 CrossRef CAS PubMed.
  42. F. Pascale, C. M. Zicovich-Wilson, F. L. Gejo, B. Civalleri, R. Orlando and R. Dovesi, J. Comput. Chem., 2004, 25, 888–897 CrossRef CAS PubMed.
  43. P. E. Tomaszewski, Phase Transitions, 1992, 38, 127–220 CrossRef CAS.
  44. M. Olejak-Chodan and H. A. Eick, J. Solid State Chem., 1987, 69, 274–279 CrossRef CAS.
  45. M. Riktor, personal communication, 2012.
  46. C. J. Howard, B. J. Kennedy and C. Curfs, Phys. Rev. B: Condens. Matter Mater. Phys., 2005, 72, 214114 CrossRef.
  47. L. Vegard, Z. Phys., 1921, 5, 17–26 CrossRef CAS.
  48. M. Corno, E. Pinatel, P. Ugliengo and M. Baricco, J. Alloys Compd., 2011, 509, S679–S683 CrossRef CAS PubMed.
  49. V. D'Anna, L. M. L. Daku, H. Hagemann and F. Kubel, Phys. Rev. B: Condens. Matter Mater. Phys., 2010, 82, 024108 CrossRef.
  50. J. E. Olsen, M. H. Sørby and B. C. Hauback, J. Alloys Compd., 2011, 509, L228–L231 CrossRef CAS PubMed.
  51. L. H. Rude, O. Zavorotynska, L. M. Arnbjerg, D. B. Ravnsbæk, R. A. Malmkjaer, H. Grove, B. C. Hauback, M. Baricco, Y. Filinchuk, F. Besenbacher and T. R. Jensen, Int. J. Hydrogen Energy, 2011, 36, 15664–15672 CrossRef CAS PubMed.
  52. L. H. Rude, E. Groppo, L. M. Arnbjerg, D. B. Ravnsbæk, R. A. Malmkjaer, Y. Filinchuk, M. Baricco, F. Besenbacher and T. R. Jensen, J. Alloys Compd., 2011, 509, 8299–8305 CrossRef CAS PubMed.
  53. D. B. Ravnsbæk, L. H. Rude and T. R. Jensen, J. Solid State Chem., 2011, 184, 1858–1866 CrossRef PubMed.

Footnote

† Electronic supplementary information (ESI) available: Unit cell parameters and volumes of Ca(BH4)2 with different levels of Cl or Br substitution calculated from DFT; unit cell parameters and volumes of CaCl2 and CaBr2 in space groups Pnnm and P42/mnm calculated from DFT. See DOI: 10.1039/c3ra46226a
This journal is © The Royal Society of Chemistry 2014
Click here to see how this site uses Cookies. View our privacy policy here.