difluorobenzene, perfluorohexane, chloroform, etc. .... according to p. .... reactions attempted in aromatic solvents (toluene and 1,2-difluorobenzene â the latter.
Supporting Information Facile Formation of Thermodynamically Unstable Novel Borohydride Materials by a Wet Chemistry Route Tomasz Jaron´,*[a] Wojciech Wegner,[b] Karol J. Fijałkowski,[a] Piotr J. Leszczyn´ski,[a] and Wojciech Grochala*[a] chem_201404968_sm_miscellaneous_information.pdf
SUPPORTING INFORMATION Facile formation of thermodynamically unstable novel borohydride materials via a wet chemistry route Tomasz Jaroń,* Wojciech Wegner, Karol J. Fijałkowski, Piotr J. Leszczyński and Wojciech Grochala*
TABLE OF CONTENTS 1. Experimental Section 1.1. General Remarks 1.2. Mechanochemical Synthesis 1.2.1. [(C4H9)4N][Y(BH4)4] ‒ TBAYB: steps I and II in the scheme 1 from the article 1.2.2. α-Y(BH4)3 1.2.3. Attempt for mechanochemical synthesis of Na[Y(BH4)4], sample D_Na 1.2.3. Mechanochemical synthesis of Cs[Y(BH4)4], sample D 1.2.4. M[Al{OC(CF3)3}4], M = Rb, Cs 1.3. Wet synthesis 1.3.1. M[Al{OC(CF3)3}4], M = Li, Na, K 1.3.2. KAlH4 1.3.3. M[Y(BH4)4], M = Li, ..., Cs: step III in the scheme 1 from the article 1.4. Analytical methods and data processing 1.4.1. Fourier Transform Infrared spectroscopy (FTIR) 1.4.2. Powder X–ray diffraction (PXD) 1.4.4. Thermal decomposition (TGA/DSC) and evolved gases analysis (EGA) 2. Supporting Results 3. Additional information 3.1. Early approach to the wet synthesis 3.2. Use of other solvents 3.3. Use of the other precursors [Me4N][Y(BH4)4] and K[Al(OCH(CF3)3]
S1
1. Experimental Section 1.1. General Remarks All manipulations were performed in an inert gas (Ar) atmosphere ‒ in Schlenk-type glassware or in Labmaster DP glovebox from MBraun (95% of pure product (as monitored by FTIR spectroscopy and powder X-ray diffraction). Instead of extraction in Soxhlet apparatus, TBAYB can be purified via its dissolution in a small amount of dichloromethane or chloroform (e.g. 50-100 ml per 10 milimol of TBAYB synthesized) and subsequent filtration and evaporation.[S1] 1.2.2. α-Y(BH4)3 YCl3 and LiBH4 were mixed in ca. 1:3 molar ratio (with a 5 % excess of LiBH4) and milled for 60 min, resulting in the mixture of α-Y(BH4)3 (product) and 3 equivalents of LiCl (byproduct).[S2] 1.2.3. Attempt for mechanochemical synthesis of Na[Y(BH4)4], sample D_Na As the room temperature (RT) mechanochemical reaction of NaBH4 and as-synthesized Y(BH4)3 does not lead to Na[Y(BH4)4],[S3] low temperature milling was tested as the method of synthesis of this substance. NaBH4 and the as-synthesized Y(BH4)3 (containing ca. 3 equivalents of LiCl) in 1:1 molar ratio were sealed under Ar in a stainless steel bowl. The bowl was cooled by liquid nitrogen to the temperature below 0 oC and milled for 5 min. There was no temperature control, however the ice condensed on the outer walls of the bowl remained solid during the whole procedure of synthesis. 1.2.3. Mechanochemical synthesis of Cs[Y(BH4)4], sample D[S4] The as–synthesized Y(BH4)3 (in a mixture with LiCl) was milled with CsBH4 (from Katchem) in 1:1 molar ratio for 60 min in Ar atmosphere, sample D. An alternative route of synthesis via milling of Y(BH4)3 with the product of pre-milled CsCl/LiBH4 mixture (1:1 molar ratio, used without analysis), has also been tested with success. Both routes result in the mixture of Cs[Y(BH4)4], Cs2Li[Y(BH4)6–xClx], LiCl and a small amount of Y(BH4)3 (Fig. 2 in the article). However, the content of Cs[Y(BH4)4] can be maximized by heat treatment up to ca. 130 oC and cooling to the room temperature, sample D_130, cf. Fig. S 3. 1.2.4. M[Al{OC(CF3)3}4], M = Rb, Cs Li[Al{OC(CF3)3}4] and excess of MCl (in ca. 1:2 molar ratio, 200% MCl) were mixed and milled for 20-30 min. The product was then extracted with dichloromethane, the solution was filtered, concentrated under vacuum and precipitated with hexane. For M = Cs the crystallization was repeated, which resulted in rentgenographically pure product (according to PXD), while for M = Rb the product has been used as prepared which resulted in a slightly contaminated final product (with unidentified impurities). 1.3. Wet synthesis 1.3.1. M[Al{OC(CF3)3}4], M = Li, Na, K The synthesis of these substances were performed according to the slightly modified synthetic procedures described for Li[Al{OC(CF3)3}4].[S5,S6] The suspension of MAlH4 (LiAlH4 ‒ pure, from S3
Alfa-Aeser, NaAlH4 ‒ purified by dissolution in THF, filtration and evaporation of solvent, KAlH4 ‒ prepared as described below) in toluene (hexane may also be used) was cooled to 0 oC and (CF3)3COH (http://www.fluorochem.co.uk, used without purification) was slowly added with an excess easily allowing for the complete alcoholysis of Al‒H bonds (150-200 %). The reaction mixture was then refluxed for overnight, cooled, and the product was precipitated with excess of hexane, filtered, washed with hexane and dried under vacuum. The lack of Al‒H bonds was evaluated by FTIR spectroscopy; the synthesized Li[Al{OC(CF3)3}4] reveals identical FTIR spectrum to this substance provided by Iolitec provider (http://www.iolitec.de/en/). 1.3.2. KAlH4 A modified procedure reported previously by Dilts and Ashby was used.[S7] To a suspension of KH (used in ca. 2 mol% excess) in diglyme (ca. 0.5‒1 ml per 1 milimol of KH) commercially available LiAlH4 solution in tetrahydrofuran (2 mol dm‒3, Sigma-Aldrich) was added. The mixture was then stirred for 2 h, filtered and the product was precipitated with excess of toluene and dried under vacuum; yield: >75%. 1.3.3. M[Y(BH4)4], M = Li, ..., Cs: step III in the scheme 1 from the article To a suspension or solution of M[Al{OC(CF3)3}4], M = Li, ..., Cs, or Na[B{3,5(CF3)2C6H3}4], in dichloromethane a solution of [(C4H9)4N][Y(BH4)4] in the same solvent was added and the mixture was stirred for 15-120 min. Typically 50-100 ml of DCM per 1 milimol of M[Al{OC(CF3)3}4] and Na[B{3,5-(CF3)2C6H3}4] salts was used while much better solubility of TBAYB allows for dissolution it in only ca. 10 ml of DCM per 1 milimol. The suspension was then allowed to settle and the precipitate was filtered off, washed several times with the fresh portions of DCM and dried under vacuum; yield: >95%. The purity of borohydrides prepared using this method was monitored by FTIR spectroscopy, PXD, and TGA/DSC coupled with EGA. The samples prepared according to this method are collected in Table S1.
S4
Table S1. Summary of the most important samples. TBAYB – [(C4H9)4N][Y(BH4)4], [Al(ORF)4] – [Al{OC(CF3)3}4], [BArF4] – [B{3,5-(CF3)2C6H3}4], DCM - dichloromethane. Symbol
Substrate 1
Substrate 2
Remarks
W_Cs1*
TBAYB in DCM
Cs[Al(ORF)4] in DCM
-----------
W_Rb1
TBAYB in DCM
Rb[Al(ORF)4] in DCM
-----------
W_K1
TBAYB in DCM
K[Al(ORF)4] in DCM
-----------
W_Na1
TBAYB in DCM
Na[Al(ORF)4] in DCM
-----------
W_Na2
TBAYB in DCM
Na[BArF4] in DCM
-----------
W_Li1
TBAYB in DCM
Li[Al(ORF)4] in DCM
reaction at ca. –30 oC to minimize thermal decomposition of the product
D_Na
Y(BH4)3 + 3 LiCl
NaBH4
mechanochemical process according to p. 1.2.2
* - sample denoted as sample W in the article.
1.4. Analytical methods and data processing 1.4.1. Fourier Transform Infrared spectroscopy (FTIR) FTIR spectra of all solid products were measured using Vertex 80v FT–IR spectrometer (Bruker). Anhydrous KBr was used as a pellet material (200 mg per pellet). 1.4.2. Powder X–ray diffraction (PXD) PXD patterns of solids (sealed under argon inside 0.3, 0.5 or 1 mm quartz capillaries) were measured using two diffractometers: a) Bruker D8 Discover diffractometer (parallel beam; the CuKα1 and CuKα2 radiation intensity ratio of ca. 2:1), denoted here as CuKα, b) Panalytical X'Pert Pro diffractometer (parallel beam; the CoKα1 and CoKα2 radiation intensity ratio of ca. 2:1), denoted here as CoKα. 1.4.3. X–ray diffraction analysis and structure refinement. The following software was used: Crystal Sleuth[S8] – preliminary data analysis; Mercury, VESTA[S9] – structure visualization, simulation of powder patterns; Jana2006[S10] – Rietveld S5
refinement. During the Rietveld refinement of all structures described here the restraints have been applied to the H–B–H angles and B–H distances in BH4 groups: 109.47o (with a standard uncertainty, s.u. = 0.01o) and 1.15 Å (s.u. = 0.001 Å), respectively. Similarly, the geometry of the [Y(BH4)4]– ion has been restrained: Y–Hbridge distances were kept equal (with s.u. = 0.01 Å), while for Li[Y(BH4)4], due to the low quality experimental data, the B–Y–B angles were additionally restrained to 109.47o (s.u. = 0.1o). Due to such restraints, Hterminal atoms were positioned collinear with Y–B contacts. Atomic displacement parameters (ADP) of H were fixed as 1.5 ADP of adjacent B atom (riding model), while for Li[Y(BH4)4] one isotropic atomistic displacement factor was refined for all atoms. The background for structure refinement was described by 30–36 Legendre polynomials; pseudo-Voigt peak shape function was used. In Li[Y(BH4)4] lithium atom has been initially placed at 4 k Wyckoff position (0 0 z), as for scandium analogue, with the occupancy of 0.5. However, refinement led to higher symmetry localized position, 2a (0 0 1/4), and finally this position was fixed. Unfortunately, due to slow decomposition of the sample during the measurement and the presence of unknown impurities, our PXD data are of low quality which hamper reliable refinement of the position of lightweight lithium atom, cf. Fig. S11. 1.4.4. Thermal decomposition (TGA/DSC) and evolved gases analysis (EGA) Thermal decomposition of the samples placed inside Al2O3 crucibles was investigated with a combined thermogravimeter (TGA) and differential scanning calorimeter (DSC) from Netzsch – STA 409 PG, at a constant Ar (99.9999 %) flow of 70 ml/min, and at 5 K min-1 heating rate. The evolved gases were analyzed with a quadruple mass spectrometer (MS) QMS 403 C (Pfeiffer Vacuum), connected to the TGA/DSC device by quartz capillary preheated to 200 oC to avoid condensation of low-boiling volatiles.
S6
2. Supporting Results
Figure S2. Powder X-ray Diffraction (PXD) measured and refined for the sample W_Cs1 (CuKα).
Figure S3. Analysis of the most Cs[Y(BH4)4]-rich sample prepared mechanochemically (products of mechanochemical reaction of Y(BH4)3 + 3 LiCl and CsBH4 heated to 130 oC and rapidly cooled to RT), D_130 (red)[S4] and W_Cs1 (blue). (a) PXD patterns (CuKα radiation), (b) FTIR spectra. The strongest extraneous signals (from various impurities) have been marked with asterisks. Note the significant contamination of mechanochemically prepared sample by Cs2Li[Y(BH4)6-xClx] (ca. 7.3 wt%) and LiCl (ca. 32.3 wt%). S7
Figure S4. Powder X-ray Diffraction (PXD) measured and refined for the sample W_Rb1 (CuKα). The most intense signals from unidentified impurities (probably from the contaminated Rb[Al(OR F)4] precursor) are seen on the difference curve.
Figure S5. FTIR spectra of (a) sample W_Rb1[S4] and, (b) Rb[Y(BH4)4] prepared mechanochemically in the reaction between Y(BH4)3 + 3 LiCl and RbBH4. The strongest extraneous signals (from various impurities) have been marked with asterisks.
S8
Figure S6. Powder X-ray Diffraction (PXD) measured and refined for the mixture of Y(BH4)3 + 3 LiCl and NaBH4 milled for 5 min, sample D_Na (CuKα). Na[Y(BH4)4] is undetectable.
Figure S7. FTIR spectra of (a) sample D_Na, (b) freshly prepared sample W_Na2 (containing mostly non-decomposed Na[Y(BH4)4], (c) Y(BH4)3 for comparison. For the sample D_Na note the lack of absorption band at ca. 2480 cm-1, characteristic for [Y(BH4)4]– anion (marked with a dashed line).
S9
Figure S8. PXD patterns of the sample W_Na1 and the results of Rietveld phase analysis. The signals from unknown impurity phase are marked with asterisks; this phase was not taken into account for phase analysis. Sample (a) was stored at RT for ca. 13 h more than sample (b); relative amount of the products remains unchanged (the estimated standard deviations are given in brackets). The signals from Na[Y(BH4)4] has not been marked; NaBH4 shows more intensive signals at higher angles.
S 10
Figure S9. PXD patterns of the sample W_Na2 measured at RT. The estimated time after mixing the reagents has been given. The top pattern is a fast initial scan (5 min), the other have been measured for 2 h. Initially Na[Y(BH4)4] decomposes at high rate, while after several hours practically no change is observed. Note the lack of diffraction peaks from unidentified impurities; main signals from NaBH4 are observed at higher angles (outside the scope of this plot).
S 11
Figure S10. PXD measured and refinement for the sample W_Na2 (CuKα). Sample was stored for ca. 44 h at the room temperature; at this point the diffraction pattern was stable in time, therefore a reliable refinement was possible.
S 12
Figure S11. Powder X-ray Diffraction (PXD) measured and refined for the product of reaction of [(C4H9)4N][Y(BH4)4] with Li[Al{OC(CF3)3}4], sample W_Li1 (CoKα). The signals from impurities (predominantly Li[Al{OC(CF3)3}4] substrate) not taken into account during the refinement and phase analysis, were marked with asterisks (*). According to the Quantitative Phase Analysis program from Materials Studio package (Accelrys) the residual diffraction intensity (the area below diffraction peaks with subtracted background) from unidentified impurities comprises ca. 3.4 % of the total diffraction intensity from Li[Y(BH4)4] and α-Y(BH4)3. The signals from LiBH4 (the second product of Li[Y(BH4)4] decomposition are at the noise level, therefore this phase was not added to the refinement.
S 13
Figure S12. FTIR spectra of the samples: (a) W_Li1, (b) W_Na1, (c) W_K1, (d) W_Rb1, (e) W_Cs1. The strongest signals from organic impurities have been marked with asterisks.
S 14
3. Additional information 3.1. Early approach to the wet synthesis A number of potential precursors towards the metathesis reaction (Step III in Scheme 1 in the publication) have been preliminarily tested. This involves: LiBH4, NaBH4, [Li(CH3OCH2CH2OCH3)2][B(C6H5)4], Na[B(C6H5)4], Mg(SO3CF3)2, NaO-n-C4H9, NaOC(CH3)3, Mg(OC2H5)2, LiOCH(CH3)2, Na[N(Si(CH3)3)2], NaBHEt3 and others. These substances either do not react with [Bu4N][Y(BH4)4] in dichloromethane or reactions do not lead to the inorganic borohydrides (also toluene and THF in various proportions to dichloromethane were tested). 3.2. Use of other solvents Because dichloromethane is a weakly solvating reaction medium, it does not dissolve easily the precursors towards ionic exchange (metathesis) reactions. Even the salts of various weakly coordinating anions (especially with the smallest cations) are not very well soluble in this solvent. Therefore we have tested several other solvents, including toluene (known to be a relatively good solvent of these salts, especially at elevated temperature[S5]), 1,2-difluorobenzene (of a significantly higher dielectric constant than that of DCM), perfluorohexane (a mixture of isomers) and acetonitrile (ACN), alone or in mixtures with DCM. Perfluorohexane appeared not to be a good solvent for both [Bu4N][Y(BH4)4] and K[Al{OC(CF3)3}4] used for testing, also this solvent does not mix with dichloromethane. Interestingly, reactions attempted in aromatic solvents (toluene and 1,2-difluorobenzene ‒ the latter is a very good solvent for both reagents) resulted in no precipitate after mixing the solutions, only a weak opalescence from a colloid appeared, but no precipitate formed after a few days. These solution were not analyzed. Such behavior is most probably a consequence of complexing of potassium ions by electron density-donating aromatic rings. Three samples with acetonitrile added to dichloromethane were tested: (a) pure ACN; (b) 0.5 milimol K[Al{OC(CF3)3}4] dissolved in the mixture of 10 ml DCM + 2.5 ml ACN added to [Bu4N][Y(BH4)4] dissolved in 10 ml of pure DCM; (c) as in b), but 0.5 ml ACN was added. While addition of ACN highly increases the solubility of K[Al{OC(CF3)3}4] in DCM, reaction lead to KBH4 instead of expected K[Y(BH4)4], the yttrium-containing compounds remained dissolved, Fig. S13.
S 15
Figure S13. FTIR spectra of the products of reaction between [Bu4N][Y(BH4)4] and K[Al{OC(CF3)3}4] in acetonitrile and dichloromethane mixtures (a), (b), (c) ‒ description is contain in the text above, compared with the spectrum of KBH4, (d). 3.3. Use of the other precursors [Me4N][Y(BH4)4] and K[Al(OCH(CF3)3] These precursors are weakly soluble in dichloromethane and their use results in very large amount of organic impurities in the samples, Figure S14.
S 16
Figure S14. FTIR spectra of the products of reaction between [Bu4N][Y(BH4)4] and K[Al(OCH(CF3)3] in dichloromethane. Note the very strong absorption from organic impurities (seen especially around 2900 cm-1 and below 1000 cm-1, where the borohydride-related bands are not present).
Supporting References [S1] T. Jaroń, W. Wegner, M. K. Cyrański, Ł. Dobrzycki, W. Grochala, J. Solid State Chem. 2012, 191, 279-282. [S2] T. Jaroń, W. Grochala, Dalton Trans. 2010, 39, 160-166. [S3] T. Jaroń, W. Grochala, Dalton Trans. 2011, 40, 12808-12817. [S4] T. Jaroń, W. Wegner, W. Grochala, Dalton Trans. 2013, 42, 6886-6893. [S5] I. Krossing, Chem. Eur. J. 2001, 7, 490-502. [S6] I. Krossing, A. Reisinger, Coord. Chem. Rev. 2006, 250, 2721-2744. [S7] J. A. Dilts, E. C. Ashby, Inorg. Chem. 1972, 11, 1230-1236. [S8] T. Laetsch, R. Downs, Abstracts from the 19th General Meeting of the International Mineralogical Association, Kobe, Japan, 23-28 July 2006. [S9] K. Momma, F. Izumi, J. Appl. Cryst. 2011, 44, 1272-1276. [S10] V. Petricek, M. Dusek, L. Palatinus, Jana2006. Structure Determination Software Programs. Institute of Physics, Praha, Czech Republic, 2006.
S 17
