Mechanisms of Hydrogen Generation from Tetrameric Clusters of

Dec 23, 2015 - Reversible structural transformations of rubidium and cesium amidoboranes. I.V. Kazakov , A.V. Butlak , P.A. Shelyganov , V.V. Suslonov...
1 downloads 3 Views 3MB Size
Download PDF
Article pubs.acs.org/JPCA

Mechanisms of Hydrogen Generation from Tetrameric Clusters of Lithium Amidoborane Anna V. Pomogaeva,*,† Keiji Morokuma,‡ and Alexey Y. Timoshkin† †

Inorganic Chemistry Group, Institute of Chemistry, St. Petersburg State University, Universitetskaya nab. 7/9, St. Petersburg 199034, Russia ‡ Fukui Institute for Fundamental Chemistry, Kyoto University, Kyoto 606-8103, Japan S Supporting Information *

ABSTRACT: The first-principles study of dehydrogenation mechanism of tetrameric clusters of lithium amidoborane LiNH2BH3, (LiAB)4, is presented. The choice of tetramer is based on the suspicion that dimeric cluster models used in previous theoretical studies are too small to capture the essence of the reaction. Dehydrogenation pathways starting from three isomers of (LiAB)4 tetramers were explored by applying the artificial force induced reaction (AFIR) method at the M06 level of theory. All obtained reaction pathways feature initial dimerization of two LiAB molecules in the tetramer. Formation of intermediates containing the Li3H moiety is a very characteristic feature of all pathways. In the succeeding rate-limiting step of the release of H2 molecule, a hydridic H atom of the Li3H moiety activates a protic H atom of the NH2 group with formation of the Li2H2 moiety in transition state. The most kinetically favorable pathway has the activation enthalpy of 26.6 kcal mol−1, substantially lower than that found for dimeric cluster. The obtained results suggest that only three LiAB molecules directly participate in the elementary reactions.

■

with Li+ changes chemical bonding to ionic type. Li+ acts like a shuttle for the transfer of the negatively charged Hδ− of the BH3 group to the positively charged Hδ+ of the NH2 group, allowing a straightforward release of H2 without phase transformation as in the case of AB.11 It was found that one-step direct intramolecular release of H2 is unfavorable in the case of LiAB dimer. Lee and McKee15 found that dehydrogenation of LiAB goes through formation of Li2H2 moiety in the transition state (TS). Lee et al.15 as well as Swinnen et al.16 conclude that the energetically lowest pathway proceeds with the formation of LiNHBH2··· LiNH2BH3 complex after elimination of the first H2 molecule from the dimer. Kim et al.14 found that a more favorable pathway proceeds via intermolecular Hδ−···Hδ+ interaction and oligomerization via B−N bond formation. This pathway involves BH2NH2···LiH···LiNH2BH3 intermediate where triangular Li2H moiety could be distinguished. Experimental results19 also offer some support to the intermolecular mechanism of hydrogen release. Luedtke et al.19 concluded that the ratedetermining step in the dehydrogenation of MAB involves B−H bond rupture with subsequent formation of metal hydride MH. Experimental observation of absence of dihydrogen bonds in solid LiAB20 indicates that mere H+···H− interaction mechanism is not sufficient. The oligomerization pathway with formation of [NHBHNHBH3]2− chain after the release of 2H2 from the LiAB dimer, proposed in the computational study of Kim et al.,14 is also

INTRODUCTION Intensive search for the environment-friendly renewable energy sources is largely focused on hydrogen as an energy carrier for fuel cells. It is widely accepted that usage of thermodynamically stable materials, where hydrogen is chemically bound to other elements, can solve many problems related to storage, transportation, and utilization of H2. Sorption of molecular H2 in solid state porous materials1−3 has such advantages as fast kinetics and excellent cyclability. On the other hand, materials where H atoms are covalently bound to light elements have a merit of excellent volumetric and gravimetric capacity. Among them are metal hydrides,4 complex hydrides,5,6 borohydrides,7 etc.8,9 Ammonia borane (AB), NH3BH3, with a hydrogen content of 19.6 wt % is among the best candidates for hydrogen storage.10,11 However, dehydrogenation of AB is accompanied by release of toxic borazine. This shortcoming is absent in AB derivatives where one of the hydrogen atoms bonded to nitrogen is replaced by alkali metal. Moreover, thermal dehydrogenation of metal amidoboranes (MAB) starts at lower temperatures, proceeds faster, and is less exothermic than that of AB.12 In this paper our interest is focused on the MAB with the best gravimetric capacity. LiNH2BH3 (LiAB) is found to be an excellent hydrogen source which releases 10.9 wt % of H2 at relatively low temperature (∼90 °C) without gaseous impurities.13 A number of computational and experimental studies have been devoted to the dehydrogenation mechanism of LiAB. Possible dehydration pathways were considered in computational studies for monomeric and dimeric forms of LiAB.14−18 It was shown that the substitution of ammonia proton H+ in AB © XXXX American Chemical Society

Received: October 10, 2015 Revised: December 6, 2015

A

DOI: 10.1021/acs.jpca.5b09924 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

Figure 1. Structures (in Å and deg), Mulliken charges, and relative energies (kcal mol−1) of LiNH2BH3 tetramers (1−3) and dimers (1d and 2d) optimized at the M06/6-311G(d,p) level of theory. For dimers the energies are obtained as ΔE = 2 × ΔE(dimer) − E(1).

supported by experimental observations by Shimoda.21 However, it is not clear if the dimer approximation is sufficient for the correct description of LiAB dehydrogenation mechanism. The hydrogen release mechanism in the solid state is still a challenge for computational chemistry. Structural, electronic, and energetic properties of crystalline LiAB were explored in several computational studies.22,23 It was shown that the interaction pattern for solid LiAB is rather complex due to interplay of several interactions without a single dominant type.24 Admitting intermolecular mechanism of H2 release and predominantly ionic character of interaction in the LiAB molecule, it is appropriate to assume that the particular condense-phase environment of the LiAB molecule could play very important role in the dehydrogenation process. LiAB crystallizes in the orthorhombic space group Pbca. Two polymorphs with different lattice parameters are known.25 In α-LiAB Li+ ion has tetrahedral coordination. The Li−N bond length is 2.02 Å, and Li−B distances are in range of 2.64−2.68 Å. β-LiAB has lower symmetry and double-sized unit cell compared with the α form. It is formed by alternation of two LiAB layers, where Li−N bond lengths are 1.93 and 2.04 Å for different layers. Tetrahedral coordination of Li+ is distorted in the β-LiAB. Li−B distances are 2.56, 2.57, 2.78 Å and 2.50, 2.51, 2.92 Å in both α and β phases, respectively. However, in spite of different local environments, both α and β phases of LiAB have identical dehydrogenation features. It was noticed that both phases exhibit an endothermic behavior prior the exothermic H2 release and transform to some common phase where local environment of atoms should be different from either α-LiAB or β-LiAB crystalline structures.25 A recently developed theoretical tool, denoted as global reaction route mapping (GRRM) strategy,26 allows automated search of many important possible reaction pathways in complex molecular systems. The artificial force induced reaction (AFIR) method in the GRRM strategy applies an artificial force between fragments of the system for automatic search of many approximate reaction pathways. Each approximate pathway is

further optimized without force and a global map of reaction pathways is obtained. A detailed description of the method could be found in the paper of Maeda et al.26 The strategy has successfully been applied for a number of complex molecular reaction systems. Here we performed a systematic GRRM study of dehydrogenation of the LiAB tetramer as a first step toward investigation of condensed phase hydrogen release mechanisms. Choosing a set of two dimers, we intended to keep succession with dimer investigations to challenge the features of pathways found for the dimer. Different mutual orientations of the dimers were chosen to model different local environments of the atoms, which possibly could occur in solid phase prior the dehydrogenation. We demonstrate that tetramers with different mutual orientations of LiAB molecules exhibit common dehydrogenation features.

■

COMPUTATIONAL METHOD The optimized structures of three tetramers (see Results section) are taken as the starting point for AFIR reaction pathway search. Several different “fragments” were defined, and artificial forces were applied between the fragments, as explained in detail in the Results section. The γ value, the magnitude of the applied artificial force in unit of energy, is taken to be 200 kJ mol−1; this suggests that many reaction pathways with barriers of about 40− 50 kcal mol−1 or less are likely to be found.26 Although this method is not complete, it is definitely better than optimization starting with a few initial guesses. The initial AFIR search of reaction pathways was performed at the M0627/6-31G(d) level, and many (ca. 20−30) approximate pathways have been found. From these AFIR approximate pathways, true equilibrium and transition state structures were fully optimized (of course without artificial force) at the M06/6-311G(d,p) level of theory. The optimized geometries of all considered structures are provided in the Supporting Information. In all the calculations, GRRM14 code28 was used for AFIR search and optimization, B

DOI: 10.1021/acs.jpca.5b09924 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

Figure 2. Fragmentation used in the AFIR method (a) of original tetramer structures, (b) of oligomerization products, and (c) for alternative pathways (see text). Directions of applied artificial forces in the AFIR method are shown by arrows.

Figure 3. Transition states (TS11, TS12, and TS13) and subsequent local minima (S11, S12, and S13) in oligomerization reaction of LiAB tetramers 1, 2, and 3. Numbered atoms distinguish Li3H moieties. For TSs dashed lines show Li3H moiety, and dotted lines show the direction of the imaginary frequency mode. For local minima dashed lines show Li+···H− interactions (where Li+−H− distances are less than 2 Å). Relative energies (kcal mol−1) and Mulliken atomic charges are shown.

utilizing the Gaussian 09 code29 for electronic structure calculations.

in Figure 1. Tetramer denoted as 1 is a cluster formed by head-totail interaction of conventional dimers 1d. Similarly, tetramer 2 is constructed by interaction of the “turned” head-to-head dimers 2d. Tetramer 3 is an intermediate case formed by dimers 1d and 2d. Relative energy values, reported in the text, are enthalpy/free energy at 298 K, relative to the energy of tetramer 1.

■

RESULTS AND DISCUSSION Structure of LiNH2BH3 Tetramers. Three different geometries of LiNH2BH3 tetramers (as well as two different dimers) were at first optimized at the M06/6-311G(d,p) level, as shown C

DOI: 10.1021/acs.jpca.5b09924 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

The Li3H moiety in TS11 has a shape of trigonal pyramid, close to C3v symmetry. In cases of TS12 and TS13 three Li atoms and hydrogen form nearly flat T-shaped structures (Li3 triangles with H atom situated in the middle of the base edge). Interatomic distances inside Li3H moieties are summarized in Table 1. Both

Three chosen tetramers intend to keep a different environment of the Li+ ion. In the tetramer 1 (Figure 1) central Li atoms are coordinated with one N atom and two BH3 groups. In 2, two central Li atoms have three BH3 groups each in the environment. Compared with optimized structures of the related dimers 1d and 2d, a shortening of the B−N bonds in tetramers is observed. The B−N bond lengths in tetramers are in good agreement with the experimental value of 1.561(7) Å in crystal LiAB.18 Increasing the number of the interacting LiAB molecules leads to increase of the Li−N distances from 1.928 Å in 1d to 1.957 Å in 1, making them closer to experimental values 2.02−2.04 Å in solid LiAB.18,25 All three chosen tetramers are close in energy. Setting electronic (enthalpy/free) energy of tetramer 1 to zero, the relative energies of 2 and 3 are −1.54 (−1.57/−1.82) and −0.82 (−0.76/−0.58) kcal mol−1, respectively. Different local environments of central atoms in the tetramers result in slightly different distributions of Mulliken charges. In structure 1 the central Li atoms gain +0.51 charges due to coordination by four hydrogen atoms of BH3 groups. Li atoms in tetramer 2 have charges +0.59 and coordinated by six hydrogen atoms. Even greater differences are observed between 4- and 6-coordinated Li atoms in tetramer 3. Negative values of Mulliken charges at hydrogen atoms in central BH3 groups in tetramers vary from −0.04 to −0.14 e. B. AFIR Search of Initial Oligomerization Pathways of Tetramers. Rational application of the AFIR method involves definition of interacting fragments. Cleavage of the N−B bond is known to be a key process for the dehydrogenation of AB;30 however, it is not a favorable way of LiAB dehydrogenation. Taking into account previous computational studies on LiAB dimers,14−16 for the each studied tetramer we choose one BH2NH2 as the first fragment and all remaining atoms belong to the second fragment (Figure 2a). Because of the symmetry of the starting structure, both 1 and 2 have two nonequivalent BH2NH2 units, and four such units exist in the case of 3. The target for the first step is the B−N interaction between the fragments leading to an oligomer; thus, the N atom in a chosen BH2NH2 fragment was pushed toward each of three B atoms in the remaining fragment. All resulting reaction pathways led to oligomerization with formation B−N bonds resulting in [BH3NH2BH2NH2]−1 chain. AFIR and subsequent optimization gave 23 nonequivalent TSs, with the relative energies ranging from ∼26 to ∼46 kcal mol−1. It should be reminded that some pathways yield intermediates and low-energy TSs connecting these intermediates; however, these involve small configurational changes and do not play the major role in the oligomerization reaction. Therefore, we will not discuss these low-energy bumps on potential energy profiles in the present report (will be discussed in a subsequent detailed accounts) and will focus on the global reaction pathways. Three of the most favorable pathways, starting from different tetramers 1, 2, and 3, respectively, are presented in Figure 3, and the rest of the TSs are shown in Figure S1 of Supporting Information. Energies of the lowest TSs leading to the oligomerization of two LiAB monomers are rather similar for three different starting tetramer configurations 1, 2, and 3: 25.73, 25.53, and 26.44 kcal mol−1 for TS11, TS12, and TS13, respectively. These are substantially lower than the barrier obtained in the past for dimeric species: 31 kcal mol−1 at the same level and 29 kcal mol−1 at the CCSD(T)/aVTZ level.14 These TSs, as well as most other TSs obtained from these starting configurations (Figure S1), are found to have a unique common structural motif. Namely, they all have a Li3H moiety.

Table 1. Interatomic Distances (Å) of Li3H Moieties in Transition States (M06/6-311G(d,p) Level of Theory) interatomic distance

TS11

TS12

TS13

H(1)−Li(1) H(1)−Li(2) H(1)−Li(3) Li(1)−Li(2) Li(2)−Li(3) Li(1)−Li(3)

1.766 1.840 1.829 2.781 2.889 2.786

1.776 1.946 1.742 2.773 2.667 3.508

1.840 1.786 1.797 2.633 3.553 2.781

pyramidal and T-shaped Li3H moieties in TSs have structures similar the two isomers of isolated Li3H cluster,31 found in a previous theoretical study. For example, in free pyramidal Li3H cluster, calculated at the same level of theory, Li−H distances are 1.89, 1.89, and 1.90 Å, a little longer than in TS11 and all Li−Li distances are 2.61 Å, a little shorter than in TS11. The Li3H clusters have a substantial mixing of ion pair (H−−Li3+) and covalent bonding character;32 H atom is activated by the three Li atoms. The central H atom in the Li3H moiety gains more charges in T-shaped structures (TS12, TS13, S12, S13) than in pyramidal structure (TS11, S11). The oligomerization products S11, S12, and S13 (Figure 3) have energies of 8.60, 14.60, and 9.83 kcal mol−1, with reverse barriers of 25.73, 25.53, and 26.44 kcal mol−1 to get back to respective TSs. In the structures of S11, S12, and S13, the new N−B bond has been completely formed to give dimeric BH3NH2BH2NH2 unit in the peripheral, with the Li atoms of the central Li3H cluster interacting more strongly with the neighboring H−B bonds and N atoms, than in the TSs. In compounds shown in Figure 3, hydrogen atoms situated between three Li atoms have Mulliken charges −0.41, −0.61, and −0.59 for S11, S12, and S13, respectively. Analysis of the Mulliken charge distribution allows to choose active fragments for the next GRRM step. Nearly equal to H(1) atoms, positive Mulliken charges +0.43, +0.62, and +0.59 were found at Li(1) atoms in the complexes S11, S12, and S13, respectively. Mulliken charges at Li(2) are +0.57, +0.59, and +0.54 and at Li(3) are +0.48, +0.53, and +0.54 for S11, S12, and S13, respectively. Thus, the Li(1)H(1) moiety (Figure 3), as shown in the Figure 2b, was chosen as one fragment and all other atoms as the other fragment to apply artificial forces in the next step. The LiH fragment was pushed toward each of the N atoms in the S11, S12, and S13 molecular systems. The entire oligomerization reaction step from a tetramer through a “Li3H” TS to the dimeric intermediate involves changes of many bond distances and angles. In order to clarify how these geometrical changes take place, a series of snapshots along the reaction pathway, TS01 and TS11 toward intermediate S11, have been calculated and are shown in Figure 4. As discussed above, the early stage of the reaction is not a single barrier reaction but rather comprises of traversing through different configurations of the tetramer with low barriers. The major changes in geometry in the early stage from 1 involve the rotation of the left bottom BaH3NaH2 monomer round the interacting Lia atom and break up the favorable intermonomer interaction, requiring a large amount of energy, and the energy of D

DOI: 10.1021/acs.jpca.5b09924 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

H is held by three Li atoms, Lia, Lib, and Lid, with the characteristic Li3H motif. TS01 of 25.39 kcal mol−1 is the barrier for cleavage of the BH2NH2 unit. Its energy is essentially the same as TS11. All intermediate complexes like S01 and S02 are van der Waals isomeric of similar energy that is only slightly less than energies of TS01 or TS11. It should be noted that alternative ways of fragmentation were considered as well. For example, one of central Li atoms and the closest hydridic H atom of BH3 group were treated as one fragment. This LiH fragment was pushed toward each of N atoms in the tetramer by means of AFIR method (Figure 2c, left). However, this partition did not result in a reaction pathway; if such a path exists, its energy is likely to be higher than 200 kJ mol−1, the adopted γ value. A dehydrogenation of the tetramer by substitutions of H atom in NH2 group of any LiAB molecule with lithium is unlikely. Another way of fragmentation, cleavage of N− B bonds was tested considering a BH3 group as the first fragment (Figure 2c, right). This partition also did not provide a reaction pathway within the predefined γ value, indicating that such a reaction requires much larger barrier. Thus, it appears that the reaction pathways going through the dimeric intermediates are kinetically the most preferable. C. AFIR Search of Hydrogen Molecule Release Pathways of Tetramers. Application of AFIR to oligomeric

Figure 4. Snapshots of the lowest oligomerization pathway from 1 via TS11 toward S11 with some of intermediate local minima. Relative energies (kcal mol−1) of TSs and local minima are shown. Letters a−d denote different monomeric units in 1. Red circles mark H atoms that become involved in Li3H moiety.

1a, for instance, becomes as high as 24 kcal mol−1. The rotation, however, eventually brings in a favorable interaction between one of the hydrogen atoms, H on Ba, and Lib and Lid of the other dimer units. As the reaction proceeds, the bond distance H−Ba becomes longer and H is pulled closer Lia, Lib, and Lid. At TS01

Figure 5. Transition states and subsequent local minima for the hydrogen molecule release step starting from oligomerization products, S11, S12, and S13. Dashed lines and numbered atoms in TSs distinguish Li2H2 moieties. Dotted lines N(1)−H(1)−H(2) in TSs show direction of the imaginary frequency mode. Relative energies (kcal mol−1) of TSs and reaction products are shown. E

DOI: 10.1021/acs.jpca.5b09924 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A intermediates S11−S13 leads to release of one molecule of H2 from the (LiAB)4 tetramer. Among multiple TSs leading to the hydrogen release from the oligomeric intermediates, three most energetically favorable are presented in Figure 5. Other 28 obtained TSs are listed in Figure S2. TS21, TS22, and TS23 originate from different structures, S11, S12, and S13, respectively, but have similar energies, 31.53, 31.21, and 32.54 kcal mol−1, respectively. These TSs are the highest points on the entire reaction pathways and are the rate-limiting TSs of the reaction. All TSs have similar structural motifs: a Li2H2 structure where H atoms are situated between two Li cations. Li−H interatomic distances (see Table 2) differ insignificantly for all three TSs, but they are notably longer than that in the Li3H triangular motif in the oligomerization TSs.

Figure 6. Dehydrogenation reaction pathways (relative energies with respect to 1) starting from LiAB tetramers 1 (line), 2 (dashed line), and 3 (dotted line). The dash-dotted line marks intramolecular dehydrogenation pathway.

Table 2. Interatomic Distances (Å) and Bond Angles (deg) of Li2H2 Moieties in Transition States (M06/6-311G(d,p) Level of Theory) interatomic distance/bond angle

TS21

TS22

TS23

H(1)−Li(1) H(1)−Li(2) H(2)−Li(1) H(2)−Li(2) Li(1)−H(1)−Li(2) Li(1)−H(2)−Li(2)

1.959 2.190 2.013 2.040 122.3 127.6

2.292 1.951 2.053 1.954 82.03 88.48

2.146 1.879 1.996 1.917 86.5 89.8

along the stepwise pathways leading to the intermolecular H2 release through the oligomerization of two LiAB monomers. The dehydrogenation reaction is found to be a two-step process in the global sense. As discussed above, each step actually involves several low-energy intermediates and low barrier TSs, corresponding to isomerization of multicentered complexes with many possible configurations. Therefore, in the present report, we intentionally neglected these small bumps on the potential energy profiles. Overall, the studied reaction consists of two major steps: oligomerization and intermolecular dehydrogenation. Thus, only the lowest TSs are presented in Figure 6 for pathways 1, 2, and 3. As already mentioned, with the use of AFIR, it is possible to identify all the intermediates along small bumps in the reaction pathway, but in the present study we are focused on finding the global characteristic features of the reaction pathways. The first (oligomerization) step is found to be endothermic for all three considered pathways. Formation of S12 where three monomeric units are forming [BH3NH2BH2NH2LiNH2BH3]− structure (Figure 3) is the most endothermic. The second (hydrogen release) step is slightly endothermic in the case of 1 and 3 (by 3.6 and 0.3 kcal mol−1, respectively) and slightly exothermic (−1.4 kcal mol−1) for the pathway 2. Gibbs energy change is disfavorable for the oligomerization step but favorable for the subsequent hydrogen release process. ΔG°298 for the second step is −4.2, −7.4, and −5.5 kcal mol−1 for 1, 2, and 3, respectively. The overall dehydrogenation process is predicted to be endothermic in the gas phase. This agrees well with previous computational studies for LiAB dimers,14−16 but is in contrast with slightly exothermic decomposition in the solid state, reported experimentally.13 The complexation of reaction products in the solid state is expected to lower the energy of the dehydrogenated product with respect to the crystal structure of LiAB.

Formation of the trapezoidal Li2H2 motif, where both two Li and two H atoms are arranged in one plane, was noted by Lee and McKee15 in a TS leading to hydrogen molecule release from LiAB dimer. At the M06/6-311G(d,p) level of theory, the energy of this TS was 41.02 kcal mol−1 higher with respect to the dimer 1d (Figure 1). Li−H interatomic distances in the Li2H2 moiety of 1.747 and 2.008 Å and H−H distance of 0.999 Å. The Gibbs free energy ΔG°298 of the TS is 33.95 kcal mol−1 is slightly lower than 36.2 kcal mol−1 reported in the original paper15 at the CCSD(T)/6-311++G(3d,2p) level of theory. We note that TS21, TS22, and TS23 are lower than the TS leading to H2 release from LiAB dimer. It should be noted that reaction intermediates S11, S12, and S13 can be considered as complexes between LiBH 3 NH 2 BH 2 NH 2 , LiH, and two LiNH2BH3 where LiH is a part of the Li3H moiety. LiH provides delivery of hydridic hydrogen atom to the protic hydrogen atom of NH2 group with subsequent release of H2 molecule. Energetically most favorable pathways through TS21, TS22, or TS23 reveal that detachment of Hδ+ proceeds not from the LiAB monomer but from the NH2 group attached to two B atoms in LiBH3NH2BH2NH2 moiety (Figure 5). In the final product after hydrogen release, a complex is formed where the N atom is bound to two Li cations. In the S22 product the N(1) atom is equally bonded to Li(1) and Li(2) atoms with bond length of 2.00 Å. The bonds are longer than N−Li bonds in LiAB monomers (Figure 1), but they are within a sum of Li and N covalent radii that indicate mixed ionic−covalent nature of these bonds. In the case of S21 and S23 the N(1)−Li(1) and N(1)− Li(2) distances are slightly different. In S21 N(1)−Li(1) bond length is 1.99 Å and N(1)−Li(2) is 2.06 Å. The respective values for S23 are 1.96 and 2.05 Å. D. Overall Favorable Reaction Pathways. Figure 6 shows the three lowest dehydrogenation pathways found in the present work. The single step concerted dehydrogenation pathway for 1 is also shown for comparison; the corresponding TS1 (Figure S3) is nearly twice higher in energy than the rate-limiting TSs

■

CONCLUSIONS The present study of (LiAB)4 tetramers allowed to identify main features of the hydrogen release mechanism. The AFIR methodology makes it possible to cover a wide spectrum of reactions that could happen in the complex system where a huge number of conformations are possible due to dense network of mixed-nature interactions. Application of artificial forces directs the reaction of interest, provides main steps of the reaction, and helps to select pathways within a practically acceptable energy F

DOI: 10.1021/acs.jpca.5b09924 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A limit. We find that the choice of γ value of 200 kJ mol−1 is appropriate to find many low-energy pathways of this complicated reaction. Within this limit, intramolecular Hδ−··· Hδ+ interaction, direct release of LiH, and direct cleavage of B−N bond are unlikely processes for the (LiAB)4 tetramer. Instead of this, energetically preferable pathway for dehydrogenation goes through oligomerization followed by the hydrogen release. It was noticed that formation of triangular Li3H moieties with shapes corresponding to isomers of free Li3H cluster is a characteristic motif of structures on the oligomerization pathway. Thus, three Li+ cations are directly involved in formation of transition states and intermediates on this reaction pathway. Hydrogen atom trapped in the Li3H moiety is found to be significantly hydric in nature to activate essentially protic hydrogen atom of NH2 group leading to H2 elimination. Kinetically favorable pathways imply H2 evolution via delivery of the Hδ− of the Li3H moiety to the first N atom of LiBH3NH2BH2NH2 oligomer. In lowest TSs for the H2 release from the oligomeric intermediates, a Li 2 H 2 moiety is distinguished where both Hδ− and Hδ+ atoms are suited between two Li atoms. In the final product of dehydrogenated N atom is coordinated with both of the Li atoms. Cleavage of the N−H bond was found the most energy consuming on the pathway, although the activation energy is comparable with energy of TS leading to the oligomerization. It should be noted that structurally and energetically similar reaction pathways are found starting from three different (LiAB)4 isomers. Finally, we point out that the most significant and favorable intermolecular interaction in the tetramers is realized via formation of the Li3H moiety prior the oligomerization. Thus, it is reasonable to assume that it may be sufficient to consider reaction pathway of LiAB trimers to reproduce the dehydrogenation process properly. Results of such studies will be presented in a forthcoming publication.

■

(3) Xia, Y.; Yang, Z.; Zhu, Y. Porous Carbon-Based Materials for Hydrogen Storage: Advancement and Challenges. J. Mater. Chem. A 2013, 1, 9365−9381. (4) Dornheim, M.; Doppiu, S.; Barkhordarian, G.; Boesenberg, U.; Klassen, T.; Gutfleisch, O.; Bormann, R. Hydrogen Storage in Magnesium-Based Hydrides and Hydride Composies. Scr. Mater. 2007, 56, 841−846. (5) Orimo, S. I.; Nakamori, Y.; Eliseo, J. R.; Zuttel, A.; Jensen, C. M. Complex Hydrides for Hydrogen Storage. Chem. Rev. 2007, 107, 4111− 4132. (6) Ley, M. B.; Jepsen, L. H.; Lee, Y.-S.; Cho, Y. W.; von Colbe, J. M. B.; Dornheim, M.; Rokni, M.; Jensen, J. O.; Sloth, M.; Filinchuk, Y.; et al. Complex Hydrides for Hydrogen Storage − New Perspectives. Mater. Today 2014, 17, 122−128. (7) Jaroń, T.; Wegner, W.; Grochala, W. M[Y(BH4)4] and M2Li[Y(BH4)6−xClx] (M = Rb, Cs): New Borohydride Derivatives of Yttrium and their Hydrogen Storage Properties. Dalton Trans. 2013, 42, 6886− 6893. (8) Grochala, W.; Edwards, P. P. Thermal Decomposition of the NonInterstitial Hydrides for the Storage and Production of Hydrogen. Chem. Rev. 2004, 104, 1283−1316. (9) Jepsen, L. H.; Ley, M. B.; Lee, Y.-S.; Cho, Y. W.; Dornheim, M.; Jensen, J. O.; Filinchuk, Y.; Jørgensen, J. E.; Besenbacher, F.; Jensen, T. R. Boron−Nitrogen Based Hydrides and Reactive Composites for Hydrogen Storage. Mater. Today 2014, 17, 129−135. (10) Stephens, F. H.; Pons, V.; Baker, R. T. Ammonia−Borane: the Hydrogen Source par Excellence? Dalton Trans. 2007, 2613−2626. (11) Zhang, J.; Lee, J. W. Progress and Prospects in Thermolytic Dehydrogenation of Ammonia Borane for Mobile Applications. Korean J. Chem. Eng. 2012, 29, 421−431. (12) Chua, Y. S.; Chen, P.; Wu, G.; Xiong, Z. Development of Amidoboranes for Hydrogen Storage. Chem. Commun. 2011, 47, 5116− 5129. (13) Xiong, Z.; Yong, C. K.; Wu, G.; Chen, P.; Shaw, W.; Karkamkar, A.; Autrey, T.; Jones, M. O.; Johnson, S. R.; Edwards, P. P.; et al. HighCapacity Hydrogen Storage in Lithium and Sodium Amidoboranes. Nat. Mater. 2008, 7, 138−141. (14) Kim, D. Y.; Singh, N. J.; Lee, H. M.; Kim, K. S. Hydrogen-Release Mechanisms in Lithium Amidoboranes. Chem. - Eur. J. 2009, 15, 5598− 5604. (15) Lee, T. B.; McKee, M. L. Mechanistic Study of LiNH2BH3 Formation from (LiH)4 + NH3BH3 and Subsequent Dehydrogenation. Inorg. Chem. 2009, 48, 7564−7575. (16) Swinnen, S.; Nguyen, V. S.; Nguyen, M. T. Potential Hydrogen Storage of Lithium Amidoboranes and Derivatives. Chem. Phys. Lett. 2010, 489, 148−153. (17) Shevlin, S. A.; Kerkeni, B.; Guo, Z. X. Dehydrogenation Mechanisms and Thermodynamics of MNH2BH3 (M = Li, Na) Metal Amidiboranes as Predicted from First Principles. Phys. Chem. Chem. Phys. 2011, 13, 7649−7659. (18) Wu, H.; Zhou, W.; Yildirim, T. Alkali and Alkaline-Earth Metal Amidoboranes: Structure, Crystal Chemistry, and Hydrogen Storage Properties. J. Am. Chem. Soc. 2008, 130, 14834−9. (19) Luedtke, A. T.; Autrey, T. Hydrogen Release Studies of Alkali Metal Amidoboranes. Inorg. Chem. 2010, 49, 3905−3910. (20) Najiba, S.; Chen, J. High-Pressure Study of Lithium Amidoborane Using Raman Spectroscopy and Insight into Dihydrogen Bonding Absence. Proc. Natl. Acad. Sci. U. S. A. 2012, 109, 19140−19144. (21) Shimoda, K.; Doi, K.; Nakagawa, T.; Zhang, Y.; Miyaoka, H.; Ichikawa, T.; Tansho, M.; Shimizu, T.; Burrell, A. K.; Kojima, Y. Comparative Study of Structural Changes in NH3BH3, LiNH2BH3, and KNH2BH3 During Dehydrogenation Process. J. Phys. Chem. C 2012, 116, 5957−5964. (22) Ramzan, M.; Silvearv, F.; Blomqvist, A.; Scheicher, R. H.; Lebègue, S.; Ahuja, R. Structural and Energetic Analysis of the Hydrogen Storage Materials LiNH2BH3 and NaNH2BH3 from ab initio calculations. Phys. Rev. B: Condens. Matter Mater. Phys. 2009, 79, 132102.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpca.5b09924. Figures S1−S3; optimized structures of all reactants, transition states, and intermediates (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*E-mail [email protected]; Ph +7 (812) 428 4071. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was financially supported by the Russian Science Foundation grant 14-13-00151. Research was carried out using computational resources provided by Resource Center “Computer Center of SPbSU” and a computer cluster at FIFC, Kyoto University.

■

REFERENCES

(1) Bénard, P.; Chahine, R. Storage of Hydrogen by Physisorption on Carbon and Nanostructured Materials. Scr. Mater. 2007, 56, 803−812. (2) Hirscher, M.; Panella, B.; Schmitz, B. Metal-Organic Frameworks for Hydrogen Storage. Microporous Mesoporous Mater. 2010, 129, 335− 339. G

DOI: 10.1021/acs.jpca.5b09924 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A (23) Chittari, B. L.; Tewari, S. P. First Principles Calculations of LiNH2 BH3, LiNH3 BH4, and NaNH2 BH3. Phys. Status Solidi B 2014, 251, 898−906. (24) Wolstenholme, D. J.; Titah, J. T.; Che, F. N.; Traboulsee, K. T.; Flogeras, J.; McGrady, G. S. Homopolar Dihydrogen Bonding in AlkaliMetal Amidoboranes and Its Implications for Hydrogen Storage. J. Am. Chem. Soc. 2011, 133, 16598−16604. (25) Wu, C.; Wu, G.; Xiong, Z.; David, W. I. F.; Ryan, K. R.; Jones, M. O.; Edwards, P. P.; Chu, H.; Chen, P. Stepwise Phase Transition in the Formation of Lithium Amidoborane. Inorg. Chem. 2010, 49, 4319− 4323. (26) Maeda, S.; Ohno, K.; Morokuma, K. Systematic Exploration of the Mechanism of Chemical Reactions: the Global Reaction Route Mapping (GRMM) Strategy Using the ADDF and AFIR Methods. Phys. Chem. Chem. Phys. 2013, 15, 3683−3701. (27) Zhao, Y.; Truhlar, D. G. The M06 Suite of Density Functionals for Main Group Thermochemistry, Thermochemical Kinetics, Noncovalent Interactions, Excited States, and Transition Elements: Two New Functionals and Systematic Testing of Four M06-Class Functionals and 12 other Functionals. Theor. Chem. Acc. 2008, 120, 215−241. (28) Maeda, S.; Harabuchi, Y.; Osada, Y.; Taketsugu, T.; Morokuma, K.; Ohno, K. GRRM14, 2014. (29) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; et al. Gaussian 09 Revision C.01; Gaussian: Wallingford, CT, 2009. (30) Nguyen, V. S.; Matus, M. H.; Grant, D. J.; Nguyen, M. T.; Dixon, D. A. Computational Study of the Release of H2 from Ammonia Borane Dimer (BH3NH3)2 and Its Ion Pair Isomers. J. Phys. Chem. A 2007, 111, 8844−8856. (31) Wu, C. H.; Jones, R. O. Stability and Structure of LinH Molecules (n = 3−6): Experimental and Density Functional Study. J. Chem. Phys. 2004, 120, 5128−5132. (32) Montgomery, J. A., Jr.; Michels, H. H.; Güner, O. F.; Lammertsma, K. The Structure and Bonding of Li3H Ion-Pair States. Chem. Phys. Lett. 1989, 161, 291−296.

H

DOI: 10.1021/acs.jpca.5b09924 J. Phys. Chem. A XXXX, XXX, XXX−XXX