Investigating the Factors Affecting the Ionic Conduction in Nanoconﬁned NaBH 4

: Nanoconﬁnement is an effective strategy to tune the properties of the metal hydrides. It has been extensively employed to modify the ionic conductivity of LiBH 4 as an electrolyte for Li-ion batteries. However, the approach does not seem to be applicable to other borohydrides such as NaBH 4 , which is found to reach a limited improvement in ionic conductivity of 10 − 7 S cm − 1 at 115 ◦ C upon nanoconﬁnement in Mobil Composition of Matter No. 41 (MCM-41) instead of 10 − 8 S cm − 1 . In comparison, introducing large cage anions in the form of Na 2 B 12 H 12 naturally formed upon the nanoconﬁnement of NaBH 4 was found to be more effective in leading to higher ionic conductivities of 10 − 4 S cm − 1 at 110 ◦ C. assem-bled in the glove box under argon atmosphere. EIS was conducted using an alternating current impedance spectroscopy method with a VMP3 potentiostat from BioLogic. The AC impedance measurement was set from 100 mHz to 1 MHz. The measurement was conducted in the temperature range from 25 to 135 ◦ C with an interval of 10 ◦ C. Before each measurement, the sample dwelled for 20 min for temperature equilibration.


Introduction
The development of solid-state electrolytes is critical for the advancement of all solidstate high-energy density batteries. In this respect, several families of inorganic solid-state electrolytes have been intensively investigated because of their non-flammability and capability in leading to a high Li + /Na + transference number and thus safety with extended battery life-time [1]. However, current solid-state electrolytes still suffer from low ionic conductivity [2]. Current materials reported to lead to high ionic conductivity are oxide solid-state electrolytes like La 0.52 Li 0. 35 TiO 2.96 (ionic conductivity of 10 −3 at room temperature) [3]. However, in this case, the poor contact between the electrode and the ceramic electrolyte causes a high interfacial resistance [4]. Sulfide-based solid-state electrolytes have also been found to lead to high ionic conductivity (e.g., Li 9.54 Si 1.74 P 1.44 S 11.7 C l0. 3 has an ionic conductivity of 2.5 × 10 −2 S cm −1 at room temperature) [5], and this is comparable to the ionic conductivity of organic electrolytes [6]; however, sulfide electrolytes have a narrow electrochemical stability window and tend to decompose upon operation and release toxic H 2 S gas [7,8].
Recently, metal borohydrides have drawn considerable interest because of their chemical and electrochemical stability, as well as chemically compatibility with Li/Na metal [9], but modifications are needed to enable high ionic conductivity [10] in complex borohydrides at ambient and not high temperatures. To date, the most common strategies to tailor the ionic conductivity of complex borohydrides are based on anionic substitution [11][12][13] or nanoconfinement approaches [10,14]. For example, partial anionic substitution of BH 4 − by NH 2 in NaBH 4 leads to an increase in ionic conductivity from 1 × 10 −9 to 2 × 10 −6 S cm −1 at 26 • C [11]. Nanoconfining LiBH 4 (LiBH 4 @MCM-41) within the pores of ordered silica scaffolds such as MCM-41 was reported to lead to a high ionic conductivity of 0.2 mS cm −1 at 55 • C instead of 1 × 10 −8 S cm −1 [10] owing to the interface interaction between LiBH 4 and MCM-41 [15,16]. Further testing of such a material showed that nanoconfined LiBH 4 in an Li/S cell could deliver a high capacity of 1220 mAh g −1 at a working voltage of 2 V and Inorganics 2021, 9, 2 2 of 10 at 0.03 C rates for 40 cycles, which is comparable to the sulfide electrolytes [17,18]. To date, such improvements have only been reported for LiBH 4 , with no report on the possibility of such an observation to other borohydrides and in particular NaBH 4 , which is of interest as a solid electrolyte for Na batteries. Except for borohydrides materials, nanoconfinement has been proven as an effective method to tailor the ionic conductivity of electrolyte materials including solid composite polymer electrolytes and ionic liquid electrolytes confined in Metal Organic Frameworks [19][20][21]. Herein, we report on the successful confinement of NaBH 4 into the MCM-41, and the resulting limited improvement in ionic conductivity. As a reference, LiBH 4 was also confined into the scaffold material MCM-41 and the existence of an oxide phase in LiBH 4 @MCM-41 further proved our hypothesis, i.e., the strong oxidation of NaBH 4 upon nanoconfinement in MCM-41 and thus the formation of an extensive oxide phase at the borohydride/MCM-41 interface limits Na and Li conduction. Better improvement in ionic conductivity could be achieved by partially decomposing NaBH 4 into a mixture of Na 2 B 12 H 12 and NaBH 4 .

Infiltration of NaBH 4 in MCM-41
To confirm the degree of infiltration of NaBH 4 into MCM-41, Brunauer-Emmett-Teller (BET) measurements were carried out. As expected, melt infiltration of NaBH 4 into MCM-41 resulted in a pore volume reduction from the original MCM-41 value of 1.02 to 0.02 cm 3 g −1 , and this corresponded to a 78% pore filling. As reflected by BET, the melt infiltration of NaBH 4 into MCM-41 also led to a significant decrease of the surface area of NaBH 4 @MCM-41 from 1110.91 m 2 g −1 for pristine MCM-41 to 3.5 m 2 g −1 (Table S1 and Figure S1).
Further analysis by transmission electron microscopy (TEM) showed that the ordered porous structure of MCM-41 was filled upon nanoconfinement of NaBH 4 with the disappearance of a clear porous structure ( Figure S2a). Additionally, elemental mapping showed a signal of Na and Si overlapping, and this was taken as additional evidence that NaBH 4 was melt infiltrated within the porosity of MCM-41 ( Figure S3). X-ray diffraction (XRD) patterns of the pristine and nanoconfined NaBH 4 are shown in Figure 1a. For NaBH 4 @MCM-41, all the peaks are assigned to NaBH 4 , which indicates that no detectable additional phase was formed during the melt infiltration process at 560 • C. However, all the diffraction peaks showed a significant boarding and shift to lower diffraction angles as compared with pristine NaBH 4 (Figure 1b), which suggested some confinement of the borohydride [22][23][24]. Indeed, a shift has been reported to occur for confined borohdyrides owing to the lattice strain imposed by the MCM-41 scaffold [22,25], as further revealed by the small angle X-ray diffraction analysis of MCM-41 and NaBH 4 @MCM-41. As shown in Figure 1c, the main (100) diffraction peak of MCM-41 broadens. This peak also shifted to higher diffraction angles, further indicating that NaBH 4 was located within the internal pore of MCM-41 [26]. From these results, it can thus be concluded that NaBH 4 is infiltrated within the porosity of MCM-41. Another indirect evidence of the nanoconfinement is the shift in the dehydrogenation peak of NaBH 4 upon infiltration in MCM-41, from 550 to 520 • C ( Figure S4).
To further determine any amorphous phases that may have formed during the nanoconfinement of NaBH 4 in MCM-41 at 560 • C, Fourier Transform Infrared Spectrometery (FTIR)and nuclear magnetic resonance (NMR) analyses were carried out. By FTIR (Figure 2), the typical BH stretching and bending vibrations corresponding to the BH 4 − anion in NaBH 4 were observed in NaBH 4 @MCM-41 in the range from 2400 to 2200 cm −1 and at 1091 cm −1 , in agreement with previous reports [27]. The broad peaks between 3800 and 3200 cm −1 were assigned to OH stretching modes corresponding to a partial oxidation of NaBH 4 in contact with the walls of the MCM-41. As a scaffold material, MCM-41 possesses silanol (Si-OH) and hydrogen-bonded terminal hydroxyl (Si-OH-O-Si) groups located within its internal structure [28]. Therefore, during the melt infiltration of NaBH 4 , it is not surprising that silanol and/or hydroxyl groups readily react with NaBH 4 to lead to the formation of boron oxide phases in NaBH 4 @MCM-41, as evidenced by the peaks at 1626, 883, and 794 cm −1 owing to the B-O vibrational modes ( Figure 2) [29][30][31]. To eliminate the concern regarding the formation of other oxide compounds (e.g., NaO 2 ), we carefully checked the FTIR spectrum of NaBH 4 @MCM-41 in the range of 800-400 cm −1 ( Figure S8), and the only peak located at 473 cm −1 was assigned to NaBO 4 instead of NaO 2 [32]. A similar oxidation has previously been observed upon the infiltration of LiBH 4 in SBA-15, and in this case, this led to the formation of LiBO 2 [33]. Besides these oxidized phases, the peak observed at 2496 cm −1 was attributed to the formation of Na 2 B 12 H 12 , which is commonly reported to occur upon a partial decomposition of NaBH 4 [34,35]. These results are also in agreement with the 11 B NMR spectrum ( Figure 3) showing the typical resonance of the BH 4 − anion centered at −41.95 ppm [36]; a single sharp peak at −2 ppm, assigned to boron in tetrahedral BO 4 environments [37]; and at −15.58 ppm, a peak corresponding to the dodecahedral [B 12 H 12 ] 2− anion in Na 2 B 12 H 12 [38].  case, this led to the formation of LiBO2 [33]. Besides these oxidized phases, the peak observed at 2496 cm −1 was attributed to the formation of Na2B12H12, which is commonly reported to occur upon a partial decomposition of NaBH4 [34,35]. These results are also in agreement with the 11 B NMR spectrum ( Figure 3) showing the typical resonance of the BH4 − anion centered at −41.95 ppm [36]; a single sharp peak at −2 ppm, assigned to boron in tetrahedral BO4 environments [37]; and at −15.58 ppm, a peak corresponding to the dodecahedral [B12H12] 2− anion in Na2B12H12 [38].  . Arrhenius plot of NaBH4@MCM-41 and pristine NaBH4. The cause for the small "jump" in ionic conductivity for NaBH4 is unknown as it does not correspond to any known phase transition.

Ionic Conductivity of Nanoconfined NaBH4
The ionic conductivity of the pristine NaBH4 and NaBH4@MCM-41 was determined by electrochemical impedance spectroscopy (EIS) (Figure 3). The ionic conductivity of NaBH4@MCM-41 is 10 times higher (i.e., 7.4 × 10 −10 S cm −1 at 20 °C) than that of pristine NaBH4 in the temperature range of 20-70 °C. However, above 70 °C, the ionic conductivity of NaBH4 and NaBH4@MCM-41 was found to be of the same magnitude. To our surprise, encapsulating NaBH4 into MCM-41 did not significantly enhance the ionic conductivity. This may be because of the extensive reaction of nanoconfined NaBH4 with the walls of MCM-41 during melt infiltration, which resulted in the formation of an insulating oxide phase [39]. Indeed, boron oxide phases (e.g., NaBO4) have been reported to be poor ionic conductors [40]. In previous reports, it has been suggested that the ionic conductivity of MCM-41 nanoconfined LiBH4 was the result of the LiBH4/MCM-41 interface promoting the reorientation of BH4 − [10,25]. To verify this hypothesis, we reproduced the nanoconfinement of LiBH4 by melt infiltration in MCM-41 [10]. Successful nanoconfinement of LiBH4 was verified by BET and thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC) analysis ( Figures S5 and S6). In this case, the ionic conductivity of LiBH4@MCM-41 was found to be 1.34 × 10 −7 S cm −1 at 110 °C ( Figure S9), which is four times lower in magnitude as compared with the reported value of 10 −3 S cm −1 [10]. Through careful analysis by FTIR ( Figure S7 . Arrhenius plot of NaBH 4 @MCM-41 and pristine NaBH 4 . The cause for the small "jump" in ionic conductivity for NaBH 4 is unknown as it does not correspond to any known phase transition.

Ionic Conductivity of Nanoconfined NaBH 4
The ionic conductivity of the pristine NaBH 4 and NaBH 4 @MCM-41 was determined by electrochemical impedance spectroscopy (EIS) (Figure 3). The ionic conductivity of NaBH 4 @MCM-41 is 10 times higher (i.e., 7.4 × 10 −10 S cm −1 at 20 • C) than that of pristine NaBH 4 in the temperature range of 20-70 • C. However, above 70 • C, the ionic conductivity of NaBH 4 and NaBH 4 @MCM-41 was found to be of the same magnitude. To our surprise, encapsulating NaBH 4 into MCM-41 did not significantly enhance the ionic conductivity. This may be because of the extensive reaction of nanoconfined NaBH 4 with the walls of MCM-41 during melt infiltration, which resulted in the formation of an insulating oxide phase [39]. Indeed, boron oxide phases (e.g., NaBO 4 ) have been reported to be poor ionic conductors [40]. In previous reports, it has been suggested that the ionic conductivity of MCM-41 nanoconfined LiBH 4 was the result of the LiBH 4 /MCM-41 interface promoting the reorientation of BH 4 − [10,25]. To verify this hypothesis, we reproduced the nanoconfinement of LiBH 4 by melt infiltration in MCM-41 [10]. Successful nanoconfinement of LiBH 4 was verified by BET and thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC) analysis (Figures S5 and S6). In this case, the ionic conductivity of LiBH 4 @MCM-41 was found to be 1.34 × 10 −7 S cm −1 at 110 • C ( Figure S9), which is four times lower in magnitude as compared with the reported value of 10 −3 S cm −1 [10]. Through careful analysis by FTIR ( Figure S7), strong B-O vibrations were observed in LiBH 4 @MCM-41 as compared with pristine LiBH 4 . These were assigned to lithium borates (Li 2 B 4 O 7 ) [41], and indicated a strong oxidation of the borohydride phase during nanoconfinement. The discrepancy between the ionic conductivity of the current LiBH 4 @MCM-41 and reported data may thus be attributed to the extensive oxidation of the complex borohydride within the scaffold material. Here, we emphasize the inevitable reaction between the pore wall in MCM-41 and complex borohydrides and the resulting negative effect on ionic conduction.
One factor that may explain the increase in ionic conductivity observed for NaBH 4 @MCM-41 below 70 • C is the presence of Na 2 B 12 H 12 , which was reported to lead to improved ionic conductivity [42]. Through integrating the chemical shifts of B 12 H 12 located at −15.6 ppm in NaBH 4 @MCM-41 (Figure 4), 18% of amorphous Na 2 B 12 H 12 appeared in NaBH 4 @MCM-41 owing to decomposition of NaBH 4 (Figure 4). We believed that the appearance of the Na 2 B 12 H 12 in the nanoconfined material is due to the interaction between the NaBH 4 and scaffold materials, triggering a partial decomposition of NaBH 4 upon its oxidation and the release of B 2 H 6 , further reacting with NaBH 4 to lead to Na 2 B 12 H 12 [43,44].
Inorganics 2021, 9, x FOR PEER REVIEW 6 of 11 One factor that may explain the increase in ionic conductivity observed for NaBH4@MCM-41 below 70 °C is the presence of Na2B12H12, which was reported to lead to improved ionic conductivity [42]. Through integrating the chemical shifts of B12H12 located at −15.6 ppm in NaBH4@MCM-41 (Figure 4), 18% of amorphous Na2B12H12 appeared in NaBH4@MCM-41 owing to decomposition of NaBH4 (Figure 4). We believed that the appearance of the Na2B12H12 in the nanoconfined material is due to the interaction between the NaBH4 and scaffold materials, triggering a partial decomposition of NaBH4 upon its oxidation and the release of B2H6, further reacting with NaBH4 to lead to Na2B12H12 [43,44]. To investigate if the presence of Na2B12H12 within NaBH4 can effectively lead to improved ionic conductivities, we synthesised NaBH4 with higher amounts of Na2B12H12. After the reaction between B2H6 and NaBH4, 43% of the starting NaBH4 was converted to Na2B12H12 ( Figure S10). Such a mixture of NaBH4 and Na2B12H12 is stable up to 300 °C (Figure S11) and exhibited a drastic enhancement in ionic conductivity, with an ionic conductivity of 10 −6 S cm −1 at 110 °C ( Figure 5). This is two magnitudes higher than pristine NaBH4 [45]. Thus, the enhancement in ionic conductivity observed in nanoconfined NaBH4@MCM-41 can be assigned to the existence of the B12H12 2− anion. Anion rotation is believed to significantly enhance cation hopping [46,47], and B12H12 − exhibits an intrinsically high dynamic motion [48]. Accordingly, the enhanced ionic conductivity in NaBH4@MCM-41 was attributed to the co-existence of B12H12 2− and BH4 − . We thus believe that fine tuning the ratio between Na2B12H12 and NaBH4 will further increase the ionic conductivity of NaBH4 and other complex borohydrides. Further investigations along this path are underway. To investigate if the presence of Na 2 B 12 H 12 within NaBH 4 can effectively lead to improved ionic conductivities, we synthesised NaBH 4 with higher amounts of Na 2 B 12 H 12 . After the reaction between B 2 H 6 and NaBH 4 , 43% of the starting NaBH 4 was converted to Na 2 B 12 H 12 ( Figure S10). Such a mixture of NaBH 4 and Na 2 B 12 H 12 is stable up to 300 • C ( Figure S11) and exhibited a drastic enhancement in ionic conductivity, with an ionic conductivity of 10 −6 S cm −1 at 110 • C ( Figure 5). This is two magnitudes higher than pristine NaBH 4 [45]. Thus, the enhancement in ionic conductivity observed in nanoconfined NaBH 4 @MCM-41 can be assigned to the existence of the B 12 H 12 2− anion. Anion rotation is believed to significantly enhance cation hopping [46,47], and B 12 H 12 − exhibits an intrinsically high dynamic motion [48]. Accordingly, the enhanced ionic conductivity in NaBH 4 @MCM-41 was attributed to the co-existence of B 12 H 12 2− and BH 4 − . We thus believe that fine tuning the ratio between Na 2 B 12 H 12 and NaBH 4 will further increase the ionic conductivity of NaBH 4 and other complex borohydrides. Further investigations along this path are underway.

Synthesis
All the operations were carried out under an inert atmosphere in an argon-filled LC-Technology glove box (<1 ppm O2 and H2O, Salisbury, MA, USA). Sodium borohydride (NaBH4, 99%) was purchased from Sigma-Aldrich (Sydney, NSW, Australia) and further purified. ZnCl2 (≥98%) was purchased from Ajax Finechemn (Sydney, NSW, Australia), and dried at 120 °C overnight on a Schlenk line under vacuum (0.01 MPa) before use. Lithium borohydride (LiBH4, 95%) was purchased from Acros (Sydney, NSW, Australia). Prior to use, LiBH4 was purified following the reported procedures [49]. MCM-41 was purchased from ACS materials and dried under vacuum at 200 °C for 2 h before use to remove any water traces.

Synthesis of the Nanoconfined Complex Borohydrides
For the nanoconfinement of NaBH4 into MCM-41 (noted NaBH4@MCM-41), a mixture of 1.5 g NaBH4 grinded with 1.27 g MCM-41 was heated at 5 °C min −1 and kept at 560 °C for 3.5 h under 8 MPa H2 pressure. This temperature was chosen because NaBH4 melts at 500 °C ( Figure S4) LiBH4@MCM-41 was synthesized as per previous report [10]. LiBH4 infiltration was carried out to fill 100% of the pore volume of MCM-41. This was achieved by mixing 0.85 g of LiBH4 with 0.127 g of MCM-41 in a mortar and pestle. The mixture was then placed in a stainless-steel sample holder and heated to 295 °C at 5 °C min −1 under an H2 pressure of 10 MPa. Infiltration was done at this temperature for 30 min.

Synthesis of the Mixture of NaBH4 and Na2B12H12 Was via Solid-Gas Reaction
NaZn2(BH4)5 was synthesised by ball milling pristine NaBH4 and ZnCl2 with a molar ratio of 2:1 with a Retsch MM301 mill operated at a frequency of 20 Hz. The mixture of NaBH4 and ZnCl2 was milled in a stainless-steel vial (25 mL) containing a single stainlesssteel ball (15 mm diameter) for 10 min. Sodium borohydride (120 mg) was ball-milled in a similar manner under high purity argon in a 15 mL stainless steel vial filled with a single stainless-steel ball (1.5 g and 15 mm diameter).
The synthesis of NaBH4/Na2B12H12 mixed compounds was undertaken in an in-house built Sievert apparatus with a customized sample holder, which had two compartments

Synthesis
All the operations were carried out under an inert atmosphere in an argon-filled LC-Technology glove box (<1 ppm O 2 and H 2 O, Salisbury, MA, USA). Sodium borohydride (NaBH 4 , 99%) was purchased from Sigma-Aldrich (Sydney, NSW, Australia) and further purified. ZnCl 2 (≥98%) was purchased from Ajax Finechemn (Sydney, NSW, Australia), and dried at 120 • C overnight on a Schlenk line under vacuum (0.01 MPa) before use. Lithium borohydride (LiBH 4 , 95%) was purchased from Acros (Sydney, NSW, Australia). Prior to use, LiBH 4 was purified following the reported procedures [49]. MCM-41 was purchased from ACS materials and dried under vacuum at 200 • C for 2 h before use to remove any water traces.

Synthesis of the Nanoconfined Complex Borohydrides
For the nanoconfinement of NaBH 4 into MCM-41 (noted NaBH 4 @MCM-41), a mixture of 1.5 g NaBH 4 grinded with 1.27 g MCM-41 was heated at 5 • C min −1 and kept at 560 • C for 3.5 h under 8 MPa H 2 pressure. This temperature was chosen because NaBH 4 melts at 500 • C ( Figure S4) LiBH 4 @MCM-41 was synthesized as per previous report [10]. LiBH 4 infiltration was carried out to fill 100% of the pore volume of MCM-41. This was achieved by mixing 0.85 g of LiBH 4 with 0.127 g of MCM-41 in a mortar and pestle. The mixture was then placed in a stainless-steel sample holder and heated to 295 • C at 5 • C min −1 under an H 2 pressure of 10 MPa. Infiltration was done at this temperature for 30 min.
3.1.2. Synthesis of the Mixture of NaBH 4 and Na 2 B 12 H 12 Was via Solid-Gas Reaction NaZn 2 (BH 4 ) 5 was synthesised by ball milling pristine NaBH 4 and ZnCl 2 with a molar ratio of 2:1 with a Retsch MM301 mill operated at a frequency of 20 Hz. The mixture of NaBH 4 and ZnCl 2 was milled in a stainless-steel vial (25 mL) containing a single stainlesssteel ball (15 mm diameter) for 10 min. Sodium borohydride (120 mg) was ball-milled in a similar manner under high purity argon in a 15 mL stainless steel vial filled with a single stainless-steel ball (1.5 g and 15 mm diameter).
The synthesis of NaBH 4 /Na 2 B 12 H 12 mixed compounds was undertaken in an in-house built Sievert apparatus with a customized sample holder, which had two compartments separated by a stainless-steel mesh (20 µm porosity). Then, 100 mg of ball milled NaBH 4 was placed on the top of the mesh and 400 mg NaZn 2 (BH 4 ) 5 was placed at the bottom of the sample holder. The reaction was carried out at 150 • C for 4 h under 10 MPa H 2 . The materials synthesized by this solid-gas reaction were then ball milled in a 15 mL stainless steel vial filled with a single stainless steel ball (1.5 g and 15 mm diameter) at a frequency of 20 Hz for 10 min, and the previous synthesis route was repeated to maximise the Na 2 B 12 H 12 yield.

Characterization
The crystalline nature of the materials was determined by X-ray diffraction (XRD) using a Philips X'pert Multipurpose XRD system (Bruker, Preston, Victoria, Australia)operated at 40 mA and 45 kV with a monochromatic Cu Kα radiation (λ = 1.541 Å)-step size = 0.01 • , time per step = 10 s·step −1 . The materials were protected against oxidation from the air by a Kapton foil.
Infrared analysis was carried out on a Bruker Vertex 70 V (Bruker, Preston, Victoria, Australia) equipped with a Harrick diffuse reflectance Praying Mantis accessory. The materials were loaded in an air-tight chamber in the glovebox and the chamber was fitted on the Praying Mantis. Spectra were acquired with a 1 cm −1 resolution with an MCT-detector.
Solid-state 11 B magic angle spinning (MAS) nuclear magnetic resonance (NMR) experiments were carried out on a narrow-bore Bruker Biospin Avance III solids-700 MHz spectrometer (Bruker, Preston, Victoria, Australia) with a 16.4 Tesla superconducting magnet operating at a frequency of 224.7 MHz 11 B nucleus. Approximately 3-10 mg of material was packed into 4 mm zirconia rotors fitted with Kel-f ® caps or vespel caps, respectively. The 4 mm rotors were spun in a double resonance H-X probe head at 14 kHz at the magic angle. The 11 B spectra were acquired with hard 1 to 3 µs radio frequency pulses corresponding to a 30 • tip angle. The recycle delays of up to 10 s were used to ensure full relaxation of the signals of all nuclei, and up to 512 transients were co-added to ensure sufficient signal to noise. The spectra were obtained at room temperature and chemical shifts were referenced using a 1 M NaCl (aq) for 23 Na, NaBH 4(s) for 11 B, and 1 M LiCl (aq) for 7 Li. The spectral deconvolution was carried out using the Dmfit software [50].
Thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC) in conjunction with mass spectrometry (MS) were conducted at 10 • C min −1 under an argon flow of 20 mL min −1 using a Mettler Toledo TGA/DSC 3 (Mettler Toledo, Melbourne, Victoria, Australia) coupled with an Omnistar (Pfeiffer) MS. Masses between m/z = 2 and 100 were followed and 40 µL alumina crucibles were used.
The MCM-41 and nanoconfined NaBH 4 morphology and elemental mapping were determined by transmission electron microscopy (TEM) and scanning transmission electron microscopy (STEM) using a Philips CM200 (Philips, Eindhoven, Netherlands) operated at 200 kV. For TEM analysis, the materials were ground and dispersed in cyclohexane, sonicated for a few seconds, and then dropped onto a carbon coated copper grid. Brunauer-Emmett-Teller (BET) was performed using a Micromeritics TriStar 3000 Analyzer from Micrometrics Instrument Corporation (Norcross, GA, USA).
The ionic conductivity of the materials was determined by electrochemical impedance spectroscopy (EIS) (Biologic, Sydney, NSW, Australia). Here, 30 mg of materials was placed in a 10 mm diameter die and uniaxially cold pressed at 9 MPa with a hydraulic press to make a pellet of 0.03 cm thickness and 0.712 cm 2 . Two polished sheets of stainless steel were used as electrodes. The pellet was then placed into a controlled environment sample holder (CESH) from BioLogic (Biologic, Sydney, NSW, Australia). The cell was assembled in the glove box under argon atmosphere. EIS was conducted using an alternating current impedance spectroscopy method with a VMP3 potentiostat from BioLogic. The AC impedance measurement was set from 100 mHz to 1 MHz. The measurement was conducted in the temperature range from 25 to 135 • C with an interval of 10 • C. Before each measurement, the sample dwelled for 20 min for temperature equilibration.

Conclusions
In conclusion, we encapsulated NaBH 4 into MCM-41 and the resulting NaBH 4 @MCM-41 material was expected to lead to higher Na + ionic conductivities. However, any enhancement in ionic conductivity was found to be hindered by the formation of oxide phases in NaBH 4 @MCM-41, and similar observations made with LiBH 4 @MCM-41 tend to prove this hypothesis. We thus propose that the inevitable oxidation between complex borohydrides and MCM-41 may have a negative effect on the ionic conductivity. However, the presence of the large cage-like anions (B 12 H 12 ) 2− formed upon the partial decomposition of the borohydrides during their nanoconfinement may lead to a more effective path to improve their ionic conductivity.