Quantum chemical and experimental insights for the ionic liquid facilitated thermal dehydrogenation of ethylene diamine bisborane

Abstract

The current work reports the judicial selection and subsequent dehydrogenation reaction with ionic liquid (IL) facilitated ethylene diamine bisborane (EDAB). Quantum chemical based COSMO-SAC (COnductor like Screening MOdel Segment Activity Coefficient) model was initially used to screen the ILs as available from Sigma Aldrich. LUMO–HOMO calculation was then performed to analyze the stability of EDAB/IL complexes. The molecular modeling studies converged on the two ILs, namely 1-ethyl-3-methyl imidazolium acetate ([EMIM][OAc]) and 1-butyl-3-methyl imidazolium acetate ([BMIM][OAc]), which were subsequently chosen for the dehydrogenation experiments. The thermal dehydrogenation of EDAB was carried out at 95 °C and 105 °C under vacuum so as to prevent generation of oxygen moieties. A total of 3.96 and 3.52 equivalents of hydrogen were released from the desorption of EDAB/[BMIM][OAc] and EDAB/[EMIM][OAc], respectively, at 105 °C. The purity of released gas was confirmed by gas chromatographic analysis, while the catalytic activity of ILs was confirmed by ¹H NMR characterization of pure EDAB, ILs and EDAB/IL complexes both before and after the reaction. ¹¹B NMR analysis confirms the presence of trigonal boron (sp²) BH₂ group as the only hydrogen containing boron moiety in dehydrogenated EDAB. Further, the two-stage release mechanism of EDAB was also verified by thermogravimetric analysis. High resolution mass spectrometry was able to detect the mass of cyclic repeat units in the polymeric chain containing an sp² BH₂ group.

---

1. Introduction

The quest for new ecofriendly fuels is constantly daunting the scientific community. Currently, renewable as well as non-renewable energy sources are being scrutinized for this purpose. Among them, hydrogen (H₂) is considered globally as an alternative fuel because of its abundant quantity in the universe. Hydrogen fuel is also considered as an environmentally friendly gas as it leaves no pollutants or by-products after burning. It also produces more energy per unit of fuel. Although hydrogen fuel has a promising future, it has some disadvantages for large scale application. The major drawback is its storage, as being light, transportation in ambient conditions is impossible as it has to be compressed at high pressure leading to the use of cryogenics. This makes it an expensive option. On the other hand, a typical hazard for liquid hydrogen is related to its high flammability, especially associated with its physical release from hydrogen storage materials. This led to studies on the chemical storage of hydrogen and the subsequent discovery of chemical hydride. Amine borane complexes¹ are unique class of chemical hydrides containing a high hydrogen content and are currently considered as an attractive option for chemical storage for hydrogen. A substantial amount of hydrogen can be released from AB at temperatures below 200 °C and ambient pressure. The simplest compound in amine borane family is ammonia borane (AB), which contains 19.6 wt% of hydrogen. The dehydrogenation of AB having a yield of 14 wt% is completed in two steps at temperatures below 130 °C. Borazine and ammonia are produced during thermolysis, which causes a major problem in fuel cell applications.^2–4 Furthermore, the dehydrogenation of AB has certain limitations such as: (a) a complete dehydrogenation temperature at 500 °C; and (b) the absence of a method for easily regenerating the end products, such as hydrazine borane. Chemical modifications of AB molecules such as carbon derivatives of AB are alternative options to overcome such problems. Methyl amine borane (MeAB),⁵ sec-butyl amine borane (SBAB)⁶ and ethylene diamine bisborane (EDAB)⁷ are some of the prominent carbon substituted amine boranes. Although these possess a lower hydrogen content compared to AB, formation of volatile compounds such as borazine is not observed in the dehydrogenation.

Kelly and Edwards⁸ first synthesized EDAB by reacting ethylene diamine with diborane under vacuum. They reported the release of two equivalents of hydrogen within 110 °C with prolonged heating and did not find any volatile impurities during the dehydrogenation process. Furthermore, the same group reported a simpler approach to produce EDAB by reacting ethylene diamine dihydrochloride and sodium borohydride in tetrahydrofuran.⁹ Neiner et al.⁷ reported the release of 10 wt% pure hydrogen from EDAB at a temperature less than 200 °C in the two-step process. They proposed an intra and intermolecular dehydrogenation process of EDAB in a temperature range of 100–200 °C. Though the release of hydrogen content is less in comparison to AB, EDAB evolved lesser amount of impurities. The most attractive feature of EDAB is the absence of an induction period during the dehydrogenation process. The rate of dehydrogenation of EDAB is faster than AB at higher temperatures and is more stable at any temperature.⁷ Leardini et al.¹⁰ investigated the dehydrogenation mechanism of EDAB in vacuum as well as inert gas flow by in situ vibrational spectroscopy. The proposed reaction mechanism was slightly different than Neiner et al.⁷ and provided the release of four equivalents of hydrogen during the first and second desorption stages.

The solubilizing property of ILs make them a natural choice as solvents in various applications.^11,12 Recently ILs have been investigated as coordinating media in the dehydrogenation reaction.^13,14 It was reported that the addition of ILs in pure AB suppresses the induction period and helps the faster release of hydrogen. They proposed that faster desorption is achieved because of the formation of the more reactive diammoniate of the diborane intermediate. Thus ILs facilitate milder conditions for the dehydrogenation of AB. The only recent work with IL-facilitated EDAB was carried out by Sahler et al.¹⁵ They observed a higher release of hydrogen equivalents with IL as compared to stand-alone EDAB. However, the selection of IL for dehydrogenation reactions is random. Based on the cation–anion combination, a huge number of ILs are possible. Thus a prior screening method is necessary to select the best possible combination(s) of cation–anion. Furthermore, AB complexes are not soluble in all classes of IL. As ILs promote a stabilizing effect on ionic intermediates¹⁶ and enhance the hydrogen desorption rate, the solubility and stability of IL–amine borane complexes must be the primary criteria for selection of ILs.

In this work, we followed a systematic approach for the selection of these ILs for the dehydrogenation of EDAB on the basis of solubility. The solubility of EDAB in ILs was predicted by the coupling of a quantum chemical and statistical mechanical framework, namely the COSMO-SAC model. Furthermore, the stability of the selected EDAB/IL complexes was then analyzed by LUMO–HOMO energies. The IL with the highest solubility of EDAB was then selected for experiments. In the present work, dehydrogenation of EDAB facilitated by imidazolium cation and acetate anion based ILs was investigated in a vacuum. The reaction mixture was characterized both before and after dehydrogenation using ¹H and ¹¹B nuclear magnetic resonance (¹H and ¹¹B NMR). The purity of the released gas was also determined by gas chromatography (GC) analysis. Thermogravimetric analysis (TGA) was performed to analyze the weight loss of the EDAB/IL complex during the thermolysis process. Furthermore, high resolution mass spectrometry (HR-MS) analysis was performed to determine the mass of the final reacted species.

2. Computational details

2.1. Screening of ILs

From the thermodynamic point of view, the infinite dilution activity coefficient (IDAC) predicts the deviation from ideal behavior in a mixture. Heintz et al.¹⁷ first laid down the concept of IDAC of the solute in ILs. Excess Gibbs free energy models such as NRTL, UNIQUAC and UNIFAC can predict IDAC but these models are dependent on parameters generated from experiments and hence are not much use. Interaction parameters involving ILs are scarce in the literature, which makes the UNIFAC model redundant for ILs. Therefore, a predictive model for IDAC on an a priori basis was necessitated. Quantum chemical based COnductor like Screening MOdel-Segment Activity Coefficient (COSMO-SAC),¹⁸ a variant of the COSMO-RS¹⁹ model, is a novel and efficient way to predict IDAC and other thermodynamic properties of a mixture on an a priori basis. The universal COSMO-SAC parameters are reported in ESI.† Screening of solvents by the COSMO-RS method and prediction of IDAC were earlier successfully adopted in extraction studies.^20,21

In COSMO-SAC, molecules are regarded as a collection of surface segments and the chemical potential of each segment is self-consistently determined from a statistical mechanical calculation. The difference in segment activity coefficient between a mixture and a pure liquid gives the segment activity coefficient. The activity coefficient of a molecule is then obtained from the summation over segment activity coefficients. The details of COSMO-SAC can be found in our previous publication.²² The first step in the COSMO-SAC scheme is the generation of COSMO files from quantum chemical calculation. The Gauss View 5.0 (ref. 23) visualization package was used to create initial structures of molecules. Geometry optimization of EDAB and ILs were carried out in the Gaussian 09 (ref. 24) quantum chemical package by density functional theory (DFT) of Becke’s three parameter exact exchange functional together with the gradient corrected correlation functional of Lee, Yang and Parr represented as B3LYP.^25,26 6-311+G(d) basis set was used for geometry optimization. The diffuse function was added to the basis set to accommodate polarization functions. Vibrational analysis (frequency calculation) was followed after geometry optimization. The absence of negative or imaginary frequencies indicated that the structures were at a global minimum. Geometry optimization of ILs were carried out on cations and anions separately assuming complete dissociation of ILs.¹¹ The optimized structures were used for COSMO file generation using the BVP86 (ref. 27) level of density functional theory. The SVP²⁸ basis set was then used in combination with density fitting basis set DGA1.²⁹ The importance of the density fitting basis set lies in the fact that it provides a significant performance gain in DFT calculation and accuracy of the molecular structures and relative energies. The calculation was performed in Gaussian 03.³⁰ The COSMO file contains the surface area of each segment, the charges on each segments and the screening charge density. Distribution of screening charges is typically represented by the σ-profile, which is merely a histogram of the screening charge densities. The COSMO-SAC parameters used in this work are reported in Table S1 of the ESI† and the σ-profile of the EDAB and ILs used in this work are reported in Fig. S1 of the ESI.† The σ-profile was then required to predict thermodynamic properties such as IDAC. As cations and anions were assumed to be completely dissociated, the σ-profile of ILs was simply a linear addition of the σ-profiles of cations and anions. Imidazolium, phosphonium, pyrrolidinium, pyridinium, sulfonium, ammonium and Basionics ILs available through Sigma-Aldrich were considered and the solubility of EDAB was tested in 205 pairs of cation–anion combinations of ILs. Highly soluble ILs were selected as best candidates for IL-facilitated hydrogen generation from EDAB. We quantify the solubility of EDAB in ILs by capacity, which is expressed as

where γ_EDAB^∞ is the IDAC of EDAB in ILs. The higher the value of the capacity, the better will be the solubility and the higher will be the release of hydrogen.

2.2. Stability analysis of EDAB/IL complexes

The reactivity and stability of a chemical complex can be explained by its excitation gaps. A detailed picture of interactions occurring among reactants and complexes can be built by a closer inspection of structural interaction and relative excitation energies. There are two different excitation gaps, which are important from a theoretical and practical point of view. The difference between the first ionization potential (IP) and the first electron affinity (EA) is defined as the fundamental gap, E_g, whereas the optical gap, E_opt is the difference between the energy of the lowest dipole allowed excited state and the ground state. Donor–acceptor complexes such as dehydrogenation of amine–borane complexes in the presence of ILs gave better insight to these gaps. The application of many body perturbation theory gives us a solid theoretical and reliable estimate of these excitation gaps at the expense of high computational cost. The computation cost is less intense for DFT and time dependent DFT (TDDFT) calculations for the estimation of excitation gaps. In order to achieve a favorable balance between accuracy and computational cost, DFT calculations are popular in electronic structure calculations. The fundamental condition within the Kohn–Sham DFT formalism states that the negative of the HOMO energy is equal to the IP of the N electron system and the LUMO can be indirectly tuned by considering the (N + 1) electron system of the N electron system.³¹ EA characterizes the susceptibility of a molecule toward attack by nucleophiles whereas IP characterizes the susceptibility of molecules toward attack by electrophiles. Hard electrophiles have a high LUMO energy whereas hard nucleophiles have a low HOMO energy.³² The LUMO–HOMO energy gap is used as a quantum chemical descriptor in establishing correlations for IL containing systems.^33,34 The high stability of molecules is implied by a large LUMO–HOMO gap because of the lower charge transfer in the complexes.

The conventional hybrid functional of DFT predicts quantitatively incorrect results for the fundamental gap. From a theoretical point of view, qualitative as well as quantitative agreement with experimental results justifies the accuracy of any computational method. A range-separated hybrid functional as described in literature^31,35 offers a quantitative agreement with experimental results of fundamental gaps. In this approach, the repulsive Coulomb potential is divided into long range and short range terms. The range separation parameter determines the dominating factor between these two parameters. The range separation parameter, often tuned semiemperically, generally decreases with system size due to delocalization of the electron density with increasing size of the system. Despite its accuracy, the range separation parameter is system-specific and the absence of a universal system-independent value makes it less popular when electronic structure calculation is not the main focus of the work. Kronik et al.³⁵ viewed the optimal tuning as nothing but a mathematical device and a significant departure from the DFT framework where attention is given to searching general functionals closer to the universal functional. In this work, we focus on a qualitative agreement with the stability of EDAB/IL complexes in dehydrogenation experiments. Fundamental gaps obtained from the B3LYP functional have at least some qualitative agreement with experimental data and are better than PBE gaps.³⁵ Here the calculation of the stability of EDAB/IL complexes with the B3LYP functional was performed with the 6-311+G(d) basis set.

3. Experimental studies

3.1. Materials

Ethylene diamine bisborane (96% purity) (EDAB), 1-butyl-3-methylimidazolium acetate (≥95% purity) ([BMIM][OAc]), 1-ethyl-3-methylimidazolium acetate (97% purity) ([EMIM][OAc]) and dimethyl sulfoxide-d₆ (deuteration degree min 99.8%, DMSO) for NMR spectroscopy were purchased from Sigma Aldrich and used as received without further purification. Acetonitrile (HPLC grade) was used for HR-MS analysis.

3.2. Experimental set up

Fig. 1 represents a simplified schematic diagram of the experimental setup. The diagram was drawn by Edraw Max software. The whole setup is divided into two distinct parts. The left side of the setup constitutes a three-necked reactor (A) with a stopper S1, labeled as the ‘reactor side’. The right side of the setup includes a liquid nitrogen condenser (B), condensed gas collector (C) and hydrogen collection measuring burette (D), collectively known as the ‘gas chamber side’. The whole experiment was performed under vacuum where both sides were connected to the vacuum pump. The gauge pressure was maintained at 5 × 10⁻² mbar (±4% instrument error) at both sides. The condenser (B) was filled with liquid nitrogen. An estimated quantity of reaction mixture (15 mg of EDAB and 0.5 ml of IL) was kept in the three neck glass reactor and heated at temperatures of 95 and 105 °C with the help of an oil bath. The reaction was performed at various time intervals. After every time interval, the stopper S1 was opened and gas was allowed to pass from reactor side to gas chamber side. Here liquid nitrogen was used as a condensing agent to absorb the undesirable gas, which evolved during the course of the reaction. Hydrogen being the lightest element however did not get condensed but entered the measuring burette (D) through stopper S3. After a certain time interval, stopper S3 was closed thereby separating the collected gas from rest of the setup. The amount of hydrogen gas released was determined by raising the mercury level through atmospheric pressure. As the ILs were used as-purchased, there is the possibility of the presence of water in the ILs. We address this problem during our experiment in two ways. The entire dehydrogenation reaction was carried out in high vacuum. The reaction temperatures (95 °C and 105 °C) are very close to boiling point of water. So, during the first vacuum period, the water will be evolved from IL. The next precaution is during condensation of gases. We use liquid nitrogen to condense the vapors evolved from the reactor. When vapors come in contact with liquid nitrogen (B and C in Fig. 1), all gases except for hydrogen condense and drain out from the S2 of Fig. 1. It should be noted that hydrogen comes out as the lightest and has a lower boiling point (−252 °C) than liquid nitrogen.

Fig. 1 Simplified schematic diagram of the experimental setup.

Samples of the EDAB/IL mixture before and after reaction were collected in separate vials for further analysis and characterization studies. The confirmatory test of the hydrogen gas was performed by collecting the gas from the marked gas burette using a Hamilton 5.0 ml (22/2′/2) gas syringe. The gas chromatography was conducted in Bruker 450 GC with a Thermal Conductivity Detector or GC-TCD technique for analyzing inorganic gases such as argon, nitrogen, hydrogen, carbon dioxide and hydrocarbon molecules. The peak location and detection of evolved gas was further checked by the GC analysis of pure hydrogen (99.8%) so as to confirm the presence of hydrogen by obtaining a similar peak location.

3.3. Analysis

3.3.1. ¹H solution nuclear magnetic resonance (NMR). ¹H solution NMR was recorded in DMSO-d₆ at room temperature on a 600 MHz Nuclear Magnetic Resonance (NMR) spectrometer made by Bruker. Seven samples namely pure EDAB, pure [BMIM][OAc], pure [EMIM][OAc], EDAB/[BMIM][OAc] (before reaction), EDAB/[BMIM][OAc] (after reaction), EDAB/[EMIM][OAc] (before reaction) and EDAB/[EMIM][OAc] (after reaction) were recorded.

3.3.2. ¹¹B solution nuclear magnetic resonance (NMR). ¹¹B solution NMR was recorded in DMSO-d₆ at room temperature on a 500 MHz Nuclear Magnetic Resonance (NMR) spectrometer made by Bruker. Four samples namely EDAB/[BMIM][OAc] (before reaction), EDAB/[BMIM][OAc] (after reaction), EDAB/[EMIM][OAc] (before reaction) and EDAB/[EMIM][OAc] (after reaction) were recorded.

3.3.3. Thermogravimetric analyzer (TGA). The thermogravimetric analysis (TGA) of EDAB/[BMIM][OAc] and EDAB/[EMIM][OAc] was performed on a Mettler Toledo thermo gravimetric analyzer (TGA/SDTA 851® model). Samples were heated from 30 to 700 °C in a 60 ml min⁻¹ flow of N₂ at heating rate of 5 °C min⁻¹.

3.3.4. High resolution mass spectrometry (HR-MS). High resolution mass spectra of EDAB/[BMIM][OAc] (after reaction) was recorded in +APCI mode on an Agilent Accurate-Mass Q-TOF LC/MS 6520, and peaks are given in m/z (% of basis peak).

4. Results and discussion

4.1. Selection of ILs

It is a well-known fact that ILs are used to generate hydrogen from EDAB at milder conditions i.e. lowering of the induction period when compared to pure EDAB. We presumed that ILs would not have gone through any structural changes. They act merely as a polar medium for the stabilization of the intermediate by lowering the activation energies of dehydrogenation. In such a context, the COSMO-SAC theory was used to predict the infinite dilution activity coefficient (IDAC) and subsequent capacity (eqn (1)). The primary criterion for selecting an IL for EDAB was higher solubility. Furthermore, selection was also considered keeping in mind the formation of acids from the ILs themselves. In this circumstance, the halogen based ILs produce acids on heating. Thus the halogen based ILs were excluded from our studies. Further we limited our studies to the 205 pairs of cation–anion combination, which are commercially available from Sigma-Aldrich. We divided these ILs into two sections based on the calculated IDAC (or capacity) namely: very high solubility (i.e. capacity ≥ ∼10⁵) and medium solubility (∼10² < capacity < 10⁵). The frequency diagram for very high solubility ILs is plotted in Fig. 2. Imidazolium, phosphonium, pyrrolidinium, pyridinium, sulfonium, ammonium and Basionics types of cations are considered for the prediction of the solubility of EDAB in ILs. In Fig. 2, ILs were grouped based on anions rather than cations. From Fig. 2, based on anion performance, we obtained three acetate, five methyl carbonate, two hydrogen sulfate and six dibutyl phosphate based ILs in which EDAB is highly soluble. Among them, solubility of EDAB in acetate based ILs is the highest. Thus we converged to the IL having acetate anion for the dehydrogenation studies. We rejected choline acetate as it has high melting point (85 °C) and is also closer to the reaction temperature. The IDAC and capacity of [EMIM][OAc] and [BMIM][OAc] are given in Table 1.

Fig. 2 Frequency of ILs. Frequency represents the number of times an anion has shown a capacity of ≥10⁵.

Table 1 IDAC and capacity of EDAB

Name of the solvent	IDAC of EDAB	Capacity of EDAB
[EMIM][OAc]	2.96 × 10^−14	3.37 × 10^13
[BMIM][OAc]	3.16 × 10^−13	3.16 × 10^12

---

Subsequently, we conducted a LUMO–HOMO study to analyze the stability of EDAB/IL complexes. With the LUMO–HOMO gap, we determined the most suitable acetate IL for the dehydrogenation. LUMO is a probable location for bond formation as incoming electrons will fill in LUMO, while HOMO provides the capacity in terms of energy to make external interactions. The stability of complexes is determined by higher negative values of LUMO/HOMO. The higher the negative values of LUMO or HOMO, the less stable the complex becomes and therefore the more reactive the complex. LUMO energies are more important for ILs involving complexes, as LUMO sites dominate the overlapping of incoming HOMO orbitals. Thus it becomes a prominent factor for external interaction connectivity. The LUMO–HOMO energy of pure EDAB, ILs and EDAB/IL complexes are reported in Table 2. Pure ILs have an almost equal LUMO energy (−0.94802 eV for [EMIM][OAc] and −0.93578 eV for [BMIM][OAc]). Thus it is not possible to determine the preference between two pure ILs. However, based on the LUMO energy of the complexes, the EDAB/[BMIM][OAc] complex has a LUMO energy of −1.44625 eV, which is lower than that of the EDAB/[EMIM][OAc] complex. This implies that EDAB/[BMIM][OAc] is less stable than EDAB/[EMIM][OAc] and thus more reactive. Plots of the LUMO–HOMO gaps of EDAB/IL complexes are reported in Fig. 3. For both complexes, LUMOs are located on the cationic imidazolium ring thus making it a favorable location to create transition or intermediate states with EDAB. In both cases, the HOMO lies on the anionic acetates. The lowering of the LUMO energies occurs due to formation of a donor–acceptor during dehydrogenation. Intermediate carbene formation might play a role during imidazolium IL supported dehydrogenation mechanism of EDAB.¹⁶ In addition to possible carbene formation from imidazolium cations, which can exert a strong effect on the hydrogen yield, the basicity of anions also influences the yield. Thus we presume that the combined effect of intermediate carbene formation from the imidazolium cation coupled with the presence of a basic acetate anion lowers the LUMO energy significantly. Based on the LUMO energy, [BMIM][OAc] is more favorable than [EMIM][OAc] for hydrogen release from EDAB. The LUMO–HOMO energy gap of pure EDAB and pure ILs are given in Fig. S2–S4 of the ESI.†

Table 2 LUMO, HOMO energies and LUMO–HOMO energy gap

Name of complex	LUMO energy/eV	HOMO energy/eV	LUMO–HOMO energy gap/eV
EDAB	−1.08081	−7.95423	6.87342
[EMIM][OAc]	−0.94802	−5.32486	4.37684
[BMIM][OAc]	−0.93578	−5.32432	4.38854
EDAB/[EMIM][OAc]	−1.10884	−6.18826	5.07942
EDAB/[BMIM][OAc]	−1.44625	−5.87098	4.42473

---

Fig. 3 LUMO–HOMO energy gap of EDAB/IL complex. (a) EDAB/[EMIM][OAc], (b) EDAB/[BMIM][OAc].

4.2. Dehydrogenation experiments with acetate based ILs

Based on our screening results, the dehydrogenation experiment of EDAB with imidazolium cation and acetate anion was performed at 95 °C and 105 °C. The time resolved hydrogen production of the two different set of experiments are shown in Fig. 4. The trend clearly reflects the absence of any induction period. Pure EDAB also does not show any induction period⁷ and releases 2.14 equivalent of H₂ at 120 °C.¹⁵ The rate of hydrogen desorption is nearly the same for both the systems until 50 min. After 50 min, the desorption rate of EDAB/[EMIM][OAc] (black circle) at 95 °C is lower and 2.59 equivalent of H₂ is released after 225 min of reaction. This yield is still significantly higher than 2.14 equivalent of hydrogen released by pure EDAB when heated at 120 °C. This proves the advantage of using ILs in dehydrogenation of EDAB. When the same mixture is heated at 105 °C, 3.52 equivalents of hydrogen (empty circle) are released. The rate of desorption was found to increase during 50–80 min. The higher yield of hydrogen shows the temperature dependency in dehydrogenation. EDAB/[BMIM][OAc] releases a higher amount of equivalent hydrogen when the reaction was performed at 95 and 105 °C. A total of 3.15 and 3.96 equivalents of H₂ is generated at 95 and 105 °C, respectively. The higher amount of equivalent H₂ release from EDAB/[BMIM][OAc] is also confirmed by the LUMO–HOMO calculation as the complex is found to be less stable than EDAB/[EMIM][OAc] complex. The dehydrogenation rates of EDAB/[BMIM][OAc] for both temperatures are almost similar until 80 min of operation. After 80 min, an increase of desorption rate at 105 °C suggests a delay in the effect of temperature for longer alkyl chains of imidazolium cation. Thus at 105 °C, we obtained an equivalent of hydrogen almost equal to that of theoretical prediction.¹⁰ This fact emphasizes the catalytic role of ILs in dehydrogenation. Using a vacuum, we achieved 3.96 equivalent of hydrogen release from EDAB/[BMIM][OAc] at 105 °C, which is more than the 3.81 equivalents of hydrogen at 120 °C as reported in the literature.¹⁵ This establishes that the milder conditions of dehydrogenation can be achieved by acetate ILs. The cumulative hydrogen production at the end of the reaction is shown in Fig. 5.

Fig. 4 Time resolved equivalent hydrogen release from EDAB/[EMIM][OAc] and EDAB/[BMIM][OAc] complexes at 95 °C and 105 °C. (●) EDAB/[EMIM][OAc] at 95 °C, (○) EDAB/[BMIM][OAc] at 95 °C, (■) EDAB/[EMIM][OAc] at 105 °C, (□) EDAB/[BMIM][OAc] at 105 °C.

Fig. 5 Cumulative hydrogen generation from EDAB facilitated by ILs at 95 °C and 105 °C. Hydrogen generation of pure EDAB at 120 °C is taken from the literature.¹⁵

4.3. Thermogravimetric analysis

Fig. 6 represents the TGA profiles of EDAB/[EMIM][OAc] and EDAB/[BMIM][OAc]. Two hydrogen desorption phenomena, as reported in the literature¹⁰ are clearly visible in Fig. 6 for both the systems. For the EDAB/[BMIM][OAc] system, the first weight loss of 16.42% occurs at 140 °C and the second major weight loss of 72.77% takes place at 250 °C. After the second major loss, there is a gradual loss of 6.31% when the sample is heated from 250 °C to 700 °C, thereby leaving a residual mass of 4.48%. For EDAB/[EMIM][OAc], the first desorption takes place at 164 °C with 6.37% mass loss and the second one at 254 °C with a corresponding weight loss of 89.09%. The concluding mass loss of 4.16% is observed from 254 °C to 700 °C. The residual mass was calculated to be 0.36%. In TGA analysis, the EDAB/[EMIM][OAc] complex requires a ∼20 min longer time for its first mass loss as compared to the EDAB/[BMIM][OAc] complex. This suggests that EDAB forms a more stable complex in [EMIM][OAc] than [BMIM][OAc]. This also confirms the relative stability analysis as predicted from HOMO–LUMO studies.

Fig. 6 TGA of EDAB/IL complexes. The complexes decompose in presence of nitrogen atmosphere at a heating rate of 5 °C min⁻¹.

4.4. ¹H NMR analysis

We conducted an ¹H NMR study to characterize pure EDAB and ILs as well as EDAB/IL complexes before and after reactions. A comparative discussion of ¹H NMR plots of pure EDAB, ILs and EDAB/IL complexes will determine the possible formation of any product after dehydrogenation and, furthermore, the role of IL in dehydrogenation. We adopted an approach which relates the number of hydrogens with their areas to the functional groups or attachments of EDAB and IL. We have successfully applied this approach for the determination of experimental timelines in liquid–liquid equilibria studies.³⁶ Here, the area of a certain peak was counted as unity and taken as the reference. The area of other peaks was integrated with reference to unity. This method is particularly useful to locate peaks of compounds in mixtures and end products after reaction. An ¹H NMR plot of EDAB is shown in Fig. 7(a). The chemical shift for –BH₃ appears between 1.13–1.67 ppm, which agrees with previous work.¹⁰ The chemical shift at 5.24 ppm is assigned to –NH₂ due to the presence of an electronegative atom in the vicinity of nitrogen. EDAB, having four hydrogen atoms connected to a nitrogen atom, contributes to an area of unity implying a value of 0.25 for each hydrogen. The chemical shift at 2.6 ppm assigned to –CH₂ (four hydrogen atoms) also has an area of unity, implying a contribution of 0.25 to each hydrogen atom. –BH₃ having six hydrogen atoms is located at 1.5 ppm with a peak area of 1.5. This also confirms to a contribution of 0.25 i.e., 1.5/6. Furthermore, the chemical shifts also verify the range as reported in the literature.¹⁰ Chemical shifts of pure [EMIM][OAc] are given in Fig. 7(b). The attached hydrogen is the acidic hydrogen atom of imidazolium core (shown as ‘f’) and corresponds to 9.89 ppm.³⁶ The other hydrogen atoms of the imidazolium ring displays chemical shift of 7.88 and 7.78 ppm. Chemical shift at 1.39–1.61 ppm are identified as methyl groups of ethyl chain of cation and acetate anion (identified as ‘d’ in Fig. 7(b)). Chemical shift for hydrogen attached with –N–CH₃ and –N–CH₂ are shown to appear between 3.86–4.21 ppm.

Fig. 7 Plot for ¹H NMR. (a) Pure EDAB, (b) pure [EMIM][OAc], (c) EDAB/[EMIM][OAc] before reaction, (d) EDAB/[EMIM][OAc] after reaction, (e) pure [BMIM][OAc], (f) EDAB/[BMIM][OAc] before reaction, (g) EDAB/[BMIM][OAc] after reaction.

The accurate identification of peaks in the pure compound will help us to locate the peaks for functional groups in the mixture. In the mixture, a shift in the peaks may be observed but the overall trend remains the same. Fig. 7(c) represents the ¹H NMR of EDAB/[EMIM][OAc] mixture before reaction. The chemical shifts of hydrogen in the imidazolium ring are identified at 9.70, 7.83 and 7.74 ppm, respectively. Hydrogens of –N–CH₂ and –N–CH₃ groups are located between 3.86–4.19 ppm. The hydrogen atoms in the alkyl group of cation are visible between 1.39–1.62 ppm. The chemical shift at 5.67 ppm is attributed to –NH₂ (four hydrogen atoms) with an area of 1.60. Similarly the chemical shift at 2.62 ppm is due to –CH₂ (four hydrogen atoms) also contributing an area of ∼1.60. A flat peak near 1.2 ppm can be assigned to the –BH₃ peak because a similar flattened peak was observed for pure EDAB (see Fig. 7(a)).

The ¹H NMR plot of EDAB/[EMIM][OAc] mixture after reaction is also reported in Fig. 7(d). We observe that hydrogens of the imidazolium ring are present between 7.85–10.16 ppm, which is the same trend as of Fig. 7(c). Further we observe chemical shifts for hydrogens of –N–CH₂ and –N–CH₃ in between 3.67–4.22 ppm. Similarly, hydrogens of the –CH₃ group of IL are identified between 1.39–1.60 ppm. It is observed that all the peaks of [EMIM][OAc] are in the same area ratio as with its initial structure. An important observation is the absence of a peak corresponding to –NH₂, thus suggesting that it has reacted and all hydrogens have evolved. There is a small peak (with area of 0.02) observed at 2.62 ppm attributed to the –CH₂ group of EDAB. Considering that the –CH₂ group (four hydrogen atoms) remains unreacted in the dehydrogenation process, each hydrogen has an area of 0.005. From this, we can conclude that hydrogens are liberated from the –NH₂ group.

A similar trend is also observed in the ¹H NMR of pure [BMIM][OAc], EDAB/[BMIM][OAc] before and after reaction as shown in Fig. 7(e–g). With the results of ¹H NMR analysis of the two systems, it can be concluded that the release of hydrogen is mainly from the –NH₂ attachment of EDAB as both ILs retain their original structure. Thus ILs merely help in facilitating the dehydrogenation process at a significantly lower temperature than pure EDAB.

4.5. ¹¹B NMR analysis

Since [BMIM][OAc] facilitated dehydrogenation of EDAB gives a higher equivalent release of hydrogen, we conducted ¹¹B solution NMR for before and after dehydrogenation samples of EDAB/[BMIM][OAc]. ¹¹B NMR analysis will give us a qualitative agreement of the boron species by comparing the respective peaks as reported in literature. As the imidazolium-acetate ILs do not possess boron species, the qualitative comparison will also reflect the possible reaction mechanism proposed in the literature.^7,10 Fig. 8(a) represents a ¹¹B NMR plot of EDAB/[BMIM][OAc] before reaction. A chemical shift at −19.61 ppm indicates a near agreement of sp³ BH₃ groups in EDAB as reported in the literature.⁷ After the dehydrogenation (Fig. 8(b)), the only noticeable chemical shift is at 20.52 ppm, which closely resembles the reported chemical shift of trigonal boron (sp²) i.e. 24 ppm.⁷ We consider this as a conclusive proof for the presence of an sp² BH₂ group in imidazolium-acetate supported dehydrogenation of EDAB. This is due to the fact that the chemical shifts for sp³ BH and BH₃ emerge on the negative side of the scale.⁷ An isolated BH₄⁻ group gives a sharp resonance at −39 ppm and asymmetric BH₂ resonance at −10 ppm.⁷ A single peak of 20.52 ppm nullifies the possibility of having BH₄⁻ and asymmetric BH₂ in the end product. Thus, only trigonal boron (sp²) is a possible presence as the end-product of dehydrogenation.

Fig. 8 Plot for ¹¹B NMR. (a) EDAB/[BMIM][OAc] before reaction, (b) EDAB/[BMIM][OAc] after reaction.

With this possibility, the possible reaction mechanism in imidazolium-acetate facilitated dehydrogenation of EDAB was deduced. As per the reported work of Neiner et al.,⁷ the presence of intramolecular dehydrogenation mechanism is ruled out due to the absence of BH₄⁻, BH₃ and BH peaks. Neiner et al.⁷ and Leardini et al.¹⁰ proposed intermolecular reaction mechanisms which differ in their intermediate products. Both the studies identified the end species having both BH and BH₂ functional groups. In our dehydrogenation experiments, the final product has merely a moiety having the sp² BH₂ group. This step is depicted as the penultimate species in the reaction mechanism presented by Leardini et al. (part b of their Scheme 1).¹⁰

4.6. HR-MS

In order to confirm the ¹¹B observation on a trigonal boron (sp²) BH₂ group, we conducted HR-MS analysis to find out the mass of the polymeric end product. It should be noted that the amount of released hydrogen cannot be explained by the degree of polymerization.¹⁶ We conducted HR-MS analysis in the positive APCI mode and the spectra is plotted in Fig. 9. Three distinctive peaks are observed in the +APCI scan. We assigned the peak at 139.1238 to the imidazolium cation by the mass of imidazolium cation calculated for C₈H₁₅N₂ (139.1230). The next two prominent peaks at 242.0734 and 339.1637 are assigned to the cyclic moiety with sp² BH₂. These species are given as the penultimate structures of the Scheme 1, mechanism ‘b’, of Leardini et al.¹⁰ For n = 3, the exact mass of C₈H₄₄B₈N₈ is 340.4433 which is close to the HR-MS spectra (339.1637). For n = 2, the exact mass of C₆H₃₁B₅N₆ is 242.3075 which again compares well to our mass at 242.0734. Thus from the analysis of mass spectra, we obtained the penultimate species of n = 2, 3 of the Scheme 1 as proposed and presented by Leardini et al.¹⁰

Fig. 9 APCI-HR-MS plot of EDAB/[BMIM][OAc] after reaction.

5. Conclusion

IL-facilitated thermal decomposition of EDAB was carried out at two temperatures (95 °C and 105 °C) under vacuum. The desorption of pure EDAB shows no induction period, while the acetate based ILs promote the dehydrogenation at lower temperatures by giving higher equivalents of hydrogen. Thus, from an economic point of view, this work has developed milder conditions for chemical hydrogen storage applications. The COSMO-SAC method was initially used to screen the ILs from a large IL database on the basis of the solubility of EDAB in ILs. This reduced the number of ILs to only imidazolium-acetate based ILs over a vast database of imidazolium, phosphonium, pyrrolidinium, pyridinium, sulfonium, ammonium and Basionics cations. The COSMO-SAC calculation predicted the highest solubility of EDAB in [EMIM][OAc] and [BMIM][OAc]. LUMO–HOMO calculations were performed to compare the relative stability of the EDAB/IL complex. EDAB in [BMIM][OAc] was proved to be less stable owing to its higher LUMO energy than [EMIM][OAc]. Furthermore, the less stable EDAB/[BMIM][OAc] complex was found to release hydrogen earlier than EDAB/[EMIM][OAc]. This phenomenon was confirmed from the TGA characterization of both complexes. We observed a higher dehydrogenation rate with both temperature and alkyl chains of imidazolium ring. The generation of hydrogen from EDAB/[BMIM][OAc] at 105 °C was nearly equal to theoretical hydrogen generation reported in literature¹⁰ and more than that reported by Sahler et al.¹⁵ Thus, we claim that the dehydrogenation process in a vacuum facilitates the creation of milder condition. ¹H NMR characterization of reactants and products confirms the structural integrity of ILs and highlights the catalytic role of ILs in dehydrogenation process. ¹¹B NMR confirms the presence of trigonal boron (sp²) BH₂ as the only boron containing moiety in dehydrogenation experiments. Further analysis of high resolution mass spectra of dehydrogenated products detects the presence of EDAB having n = 2 and 3 repeat units in the end product.

Acknowledgements

The work reported in this article was financially supported by a research grant (SB/S3/CE/063/2013) under the Science and Engineering Research Board (SERB), Department of Science and Technology (DST), Government of India. The authors further acknowledge the Central Instrument Facility of Indian Institute of Technology Guwahati (IIT Guwahati) and Indian Institute of Science Education and Research Bhopal (IISER-Bhopal) for providing the ¹H and ¹¹B NMR facility. Due acknowledgements are also due to the Analytical Laboratory of Department of Chemistry (IIT Guwahati) for letting us record the HR-MS spectra.

References

1. U. Eberle, M. Felderhoff and F. Schüth, Angew. Chem., Int. Ed., 2009, 48, 6608.
2. F. H. Stephens, V. Pons and R. Tom Baker, Dalton Trans., 2007, 2613.
3. A. Al-Kukhun, H. T. Hwang and A. Varma, Ind. Eng. Chem. Res., 2011, 50, 8824.
4. R. Halseid, P. J. S. Vie and R. Tunold, J. Power Sources, 2006, 154, 343.
5. C. A. Jaska, K. Temple, A. J. Lough and I. Manners, J. Am. Chem. Soc., 2003, 125, 9424.
6. S. S. Mal, F. H. Stephens and R. T. Baker, Chem. Commun., 2011, 47, 2922.
7. D. Neiner, A. Karkamkar, M. Bowden, Y. Joon Choi, A. Luedtke, J. Holladay, A. Fisher, N. Szymczak and T. Autrey, Energy Environ. Sci., 2011, 4, 4187.
8. H. C. Kelly and J. O. Edwards, J. Am. Chem. Soc., 1960, 82, 4842.
9. H. C. Kelly and J. O. Edwards, Inorg. Chem., 1963, 2, 226.
10. F. Leardini, M. J. Valero-Pedraza, E. Perez-Mayoral, R. Cantelli and M. A. Bañares, J. Phys. Chem. C, 2014, 118, 17221.
11. T. Banerjee, K. K. Verma and A. Khanna, AIChE J., 2008, 54, 1874.
12. A. A. P. Kumar and T. Banerjee, Fluid Phase Equilib., 2009, 278, 1.
13. M. E. Bluhm, M. G. Bradley, R. Butterick, U. Kusari and L. G. Sneddon, J. Am. Chem. Soc., 2006, 128, 7748.
14. D. W. Himmelberger, L. R. Alden, M. E. Bluhm and L. G. Sneddon, Inorg. Chem., 2009, 48, 9883.
15. S. Sahler, H. Konnerth, N. Knoblauch and M. H. G. Prechtl, Int. J. Hydrogen Energy, 2013, 38, 3283.
16. S. Sahler, S. Sturm, M. T. Kessler and M. H. G. Prechtl, Chem.–Eur. J., 2014, 20, 8934.
17. A. Heintz, D. V. Kulikov and S. P. Verevkin, J. Chem. Eng. Data, 2001, 46, 1526.
18. S. T. Lin and S. I. Sandler, Ind. Eng. Chem. Res., 2002, 41, 899.
19. A. Klamt, J. Phys. Chem., 1995, 99, 2224.
20. F. Eckert and A. Klamt, AIChE J., 2002, 48, 369.
21. T. Banerjee and A. Khanna, J. Chem. Eng. Data, 2006, 51, 2170.
22. D. Kundu and T. Banerjee, Ind. Eng. Chem. Res., 2011, 50, 14090.
23. R. Dennington, T. Keith and J. Millam, GaussView (Version 5), Semichem Inc., Shawnee Mission, KS, 2009.
24. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J. A. Montgomery Jr, J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, T. Keith, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, O. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski and D. J. Fox, Gaussian 09 (Revision D.01), Gaussian, Inc., Wallingford, CT, 2013.
25. A. D. Becke, J. Chem. Phys., 1993, 98, 5648.
26. C. Lee, W. Yang and R. G. Parr, Phys. Rev. B: Condens. Matter Mater. Phys., 1988, 37, 785.
27. J. P. Perdew, Phys. Rev. B: Condens. Matter Mater. Phys., 1986, 33, 8822.
28. A. Schäfer, H. Horn and R. Ahlrichs, J. Chem. Phys., 1992, 97, 2571.
29. C. Sosa, J. Andzelm, B. C. Elkin, E. Wimmer, K. D. Dobbs and D. A. Dixon, J. Phys. Chem., 1992, 96, 6630.
30. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J. A. Montgomery Jr, T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth, P. Salvador, J. J. Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D. Daniels, M. C. Strain, O. Farkas, D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul, S. Clifford, J. Cioslowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong, C. Gonzalez and J. A. Pople, Gaussian 03 (Revision C.02), Gaussian, Inc., Wallingford CT, 2004.
31. M. E. Foster and B. M. Wong, J. Chem. Theory Comput., 2012, 8, 2682.
32. M. Karelson, V. S. Lobanov and A. R. Katritzky, Chem. Rev., 1996, 96, 1027.
33. R. Anantharaj and T. Banerjee, Fluid Phase Equilib., 2010, 293, 22.
34. S. R. Pilli, T. Banerjee and K. Mohanty, J. Mol. Liq., 2015, 207, 112.
35. L. Kronik, T. Stein, S. Refaely-Abramson and R. Baer, J. Chem. Theory Comput., 2012, 8, 1515.
36. R. Anantharaj and T. Banerjee, Fluid Phase Equilib., 2011, 312, 20.

---

Footnote

† Electronic supplementary information (ESI) available: The COSMO-SAC parameters and the sigma profiles of EDAB, imidazolium cations and acetate anions. Further it also depicts the HOMO–LUMO energy gap of EDAB, [EMIM][OAc] and [BMIM][OAc]. See DOI: 10.1039/c5ra10625g

---