- Email: [email protected]
The mechanism of the chain-growth of ammoniaborane: A classic Lewis pairs catalysed by a Frustrated Lewis Pairs
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e9
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/he
The mechanism of the chain-growth of ammoniaborane: A classic Lewis pairs catalysed by a Frustrated Lewis Pairs Kun Wang a,b,c,**, Long-jiu Cheng b, Jian-guo Zhang c, Xue-bin Yu a,* a
Department of Materials Science, Fudan University, Shanghai 200433, PR China Department of Chemistry, Anhui University, Hefei, Anhui 230601, PR China c State Key Laboratory of Explosion Science and Technology, Beijing Institute of Technology, Beijing 100081, PR China b
article info
abstract
Article history:
Ammonia borane (AB) is an outstanding hydrogen-storage material, which has been gained
Received 20 October 2017
many attentions. It can be regenerated from polyborazylene based on the experimental
Received in revised form
study, but the uncontrollable BeN dehydro-coupling in the polymerization from AB to
30 December 2017
borazine is still a problem for the continuous application of AB. Frustrated Lewis Pairs
Accepted 1 January 2018
(FLPs) is a landmark discovery to dissociate small molecules by activating strong nonpolar
Available online xxx
bonds reversibly, which is recently applied to directly polymerize three molecular AB to a cyclotriborazylene (CTB, B3N3H12, which transforms into borazine after releasing 3 equiv.
Keywords:
H2). Here we analyse how a dimethylxanthene-derived FLPs catalyse AB to CTB. The
Frustrated Lewis Pairs (FLPs)
cyclization experiences three steps by our calculation: the connection between FLPs and AB
Metal-free hydrogenation
(1), the growth of BeN chain (2) and an intermolecular dehydrogenation between terminal
Hydrogen-storage material
N-Hdþ and
Density functional theory
nation of AB and broadens the applicability of this concept in the relevant catalyst design.
Reaction kinetic study
d
HeB (3), which gives us a clue of a possible reversibility of the dehydroge-
© 2018 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
Introduction Hydrogen energy resource is a kind of clean energy where hydrogen is stored in condensed materials by physisorption or chemisorption as atomic or ionic states [1,2]. Boron-nitrogen compounds always appear potential application in the field of the solid state hydrogen-storage materials. Ammonia borane (AB), with a molecular formula of H3NBH3, was a star compound in the area of hydrogen storage materials owing to its favourable stability and high hydrogen content (up to 19.5 wt%) [3e5]. Lots of people studied the decomposition
pathway of AB and its derivate [6e8]. Dixon's results show e BH3 group catalyse the dehydrogenation of AB very efficiently and decrease its energy gap of 6.1 kcal/mol after the formation of BH3NH3eBH3 [9]. Based on the research of Shimoda and Fijalkowski, the major mechanism of decomposition is evolution of hydrogen by N-Hdþ … dHeB directly [10]. The strong dihydrogen bond also gives AB much higher volumetric density due to the classic intramolecular interaction between Lewis acid and Lewis base [11]. AB is proved to be regenerated from polyborazylene experimentally [12]. The redistribution of BeH bond of AB is demonstrated to be accomplished from
* Corresponding author. Department of Materials Science, Fudan University, Shanghai 200433, PR China. ** Corresponding author. Department of Chemistry, Anhui University, Hefei, Anhui 230601, PR China. E-mail addresses: [email protected] (K. Wang), [email protected] (X.-b. Yu). https://doi.org/10.1016/j.ijhydene.2018.01.002 0360-3199/© 2018 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved. Please cite this article in press as: Wang K, et al., The mechanism of the chain-growth of ammoniaborane: A classic Lewis pairs catalysed by a Frustrated Lewis Pairs, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.01.002
2
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e9
borazine theoretically [13]. Mckee and Nutt hypothesized a pathway to explain the cyclization of borazine from designed NH2BHNHBH2 and BH2NH2, where both the stepwise and synergetic pathway are based on the unsaturated bonds of Be N [14]. Here we hypothesize it is a key problem for the continuous application to avoid the spontaneous chaotic dehydrogenation in the oligomerization from AB to borazine. A lot of methods have been tried to solve the problem including tuning the characters of the dehydrocoupled products on the basis of Brǿnsted and Lewis theory [15,16]. Frustrated Lewis pairs (FLPs) is discovered by Stephen and Erker based on the Lewis acid-base theory [17e19]. The cooperative reactivities of Lewis acid and Lewis Base in FLPs are usually based on the steric hindrance in one molecule (or in two molecular fragments). The ultra-high activity is usually applied to activate small molecules, such as splitting HeH bond heterolytically [20e22]. Based on such characters, FLPs can be an excellent catalyst for the activation of the BeH (or NeH) bond of AB, which is just the key step for the dehydrogenation of AB. Recently, a metal-free Frustrated Lewis Pairs (FLPs) show success in catalysing the cyclization of three molecular NH2MeBH3 to form a cyclic triborazane under a quite mild condition in Aldridge's experiments (showed in Scheme 1) [23,24]. From the cyclized products (3 or 3′ in Scheme 1), it is possible to realize reversible formation of AB [13]. It is a clue for us to design a similar pathway for the generation of pure borazine from AB for the continuous regeneration, with avoiding the strong BeN coupling in the cyclization [25]. In this paper, we obtain the mechanism of the BN chain-growth and cyclization pathway from AB catalysed by dimethylxanthene-derived FLPs on the basis of Aldridge's
experiments [23]. We also analyse the possible reversibility of each step in the process of cyclization.
Computational methods The detailed mechanism of the cyclization catalysed by FLPs is predicted by using the gradient-corrected density functional theory with dunning basis set (M062x/cc-pVTZ) [26,27]. The reactions containing ions were calculated under the theoretical level of M062x/6e311þþG(d,p). In order to simplify the model to accelerate the calculation, we substitute the eCH3 and eC6F5 in the experimental structure [22] to eH and eCF3 respectively. The optimized structures and the corresponding coordinates are provided in the supporting information (Section S1 and S3). We explain the complete catalytic process from the first BeH activation of AB to the formation of CTB including the possible four processes of BeN chain-growth. We confirmed all the transition states by applying the intrinsic reaction coordination (IRC) method. The parallel calculations performed by Møller-plesset perturbation theory (mp2/6e311þþG(d,p)). The comparison of the total energy for each compound has been listed in the supporting information (Section S2 of SI). The results from both the two methods appear the same tendencies, which indicate our calculations are reasonable. The designed pathways are showed in Fig. 1. The enthalpies of each reaction under the same theoretical level are listed in supporting information (Table S1). We list the electronic enthalpies, Gibbs Free energies and Zero point energies (ZPE) of the stationary points and transition states including all the imaginary frequency of each transition state (Table S3).
H3BNH2Me (1 equiv.) 295 K
H3BNH2Me (1 equiv.) 328 K
Ph2P
-H2
O
O B(C6F5)2
2
Ph2P H
H B H
N
Me
O
B(C6F5)2
Ph2P
H
H
H
H B H
H
N
B
Me
H 2-(H)(BH2NH2Me)
B(C6F5)2 H N
H
H Me
2-(H)(BH2NHMeBH2NH2Me) H3BNH2Me (1 equiv.), 295 K -2H2
Me H
B N
Me
N B H 3'
B
H
N Me
2 (1 mol%) 328 K -3H2
H H
Me B
H N
B
H
H
H N B N Me H Me H H 3
+ O Ph2P
B(C6F5)2 2
Scheme 1 e The cyclization from Methyl substituted amine borane (NH2MeBH3) to cyclic triborazane (3) catalysed by dimethylxanthene-based FLPs (2) in Aldridge's experiments [23]. Please cite this article in press as: Wang K, et al., The mechanism of the chain-growth of ammoniaborane: A classic Lewis pairs catalysed by a Frustrated Lewis Pairs, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.01.002
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e9
3
Fig. 1 e The reversible dehydrogenation of NH3BH3 (AB) catalysed by dimethylxanthene-derived FLPs (P1). The numbers of B/ P/N are labelled (in red) on the structures of P1, P2, P4 and P7. (For interpretation of the references to color/colour in this figure legend, the reader is referred to the Web version of this article.)
Results and discussion Hydrogenation of dimethylxanthene-derived FLP The dimethylxanthene-derived FLP (P1) is optimized in Fig. 2. We also simulate the process of the heterolytic cleavage of hydrogen caused by the zwitterionic character of P1 showed in Fig. 2. The Natural Bond Orbital (NBO) charge of boron (B1 atom, the atomic labels are showed in Fig. 1) and phosphorus (P2 atom) are 0.65 and 0.33 e, which are dramatically polarized after the hydrogenation where the charge is 0.21 and 0.84 e, respectively. An NBO analysis of the interaction between hydrogen and P1 atom confirms that the electron transfer occurs through simultaneous s(H2) / B1 atom and P2 atom/ s*(H2) donations in a push-pull manner (Fig. 2). The energy barrier of hydrogenation is 15.9 kcal/mol, where the reaction is slightly endothermic with the enthalpy of 2.1 kcal/mol. The potential energy is showed in Fig. 3. We consider the hydride affinity (HA) of the Lewis acid (B1 atom in the group of -B(PH2C13H7O)(CF3)2) and proton affinity (PA) of Lewis base (P1 atom in the group of e PH2(B(CF3)2C13H7O)), which is important for the property of
catalysis. Considering the strong interaction between ions, we calculate the Gibbs free energies for HA and PA with a dispersion basis set of M062X/6e311þþG(d,p) [24,25]. As the results, HA of the Lewis acid is 116 kcal/mol, which appears more excellent spontaneity to interact with the hydride when comparing with Karkamkar's results from experiment [28] where the HA of BH3 group is 74.2 kcal/mol. In the similar systems, only the Lewis acid of BN appears a higher HA value in Dixon's calculation (160.3 kcal/mol [29]). Similarly, the electrophilicity of -P(Me)2 group of FLPs is evaluated by the value of PA, which is 201.7 kcal/mol. Caused by the electrondonating group of methyl, the PA value is larger than that of PH3 (188.7 kcal/mol [30]), which is almost close to that of NH3 (204 kcal/mol [28]). HA and PA of FLPs appear stronger acidity and basicity than that of traditional Lewis pairs (BH3/PH3). So it should be an excellent catalyst for the dehydrogenation of AB. Additionally, the Wiberg bond index (WBI) of B1eH and P2eH is 0.91 and 0.94, respectively. The slightly higher WBI indicates P2eH bond is more covalence than that of B1eH, which is also consistent with the lower formation Gibbs free energy of P2eH bond. It's also worth to note that the dehydrogenation from P0 is triggered by the activation of B1eH. But the interaction
Please cite this article in press as: Wang K, et al., The mechanism of the chain-growth of ammoniaborane: A classic Lewis pairs catalysed by a Frustrated Lewis Pairs, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.01.002
4
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e9
Fig. 2 e The optimized structure of P1 and the mechanism of its hydrogenation. Yellow: P; Red: O; Green: F; Pink: B; Grey: C; purple: H. NLMOs of the transition states TS0 are boxed. (For interpretation of the references to color/colour in this figure legend, the reader is referred to the Web version of this article.)
Fig. 3 e The schematic potential energy surface of the first B-N chain-growth step. The possible (de)hydrogenation of P1/0 is boxed on the left. The corresponding chemical structures are listed in Fig. 1. The optimized structures are listed in the supporting information (Section S1).
between the FLP and AB is more interesting and more complicated.
The first step of BN chain growth There are three steps of the BeN chain growth (including the cyclization). The potential energy surface of the first step is showed in Fig. 3. The corresponding structures and the typical bond lengths are showed in Fig. 1 and in the supporting information (Section S1). AB and FLP first bind together to form an intermediate of [P1 þ AB], where the formation of this intermediate ([P1 þ AB] / P1 þ AB) reduces the total potential energy of 18.2 kcal/mol. In [P1 þ AB], AB links to the FLP
through the Hd of BeH bonds of AB as a structure of B/H/B1. It is a typical s-type interaction between AB and FLP. The Gibbs free energy of heterolytic splitting BeH bond of AB is 213.1 kcal/mol by Dixon's calculation under the theoretical level of CCSD(T)/aVTZ [30] (NH3BH3 / H þ [NH3BH2]þ). With the same methods, the splitting of NeH bond as the hypothetic reaction of NH3BH3 / Hþ þ [NH2BH3]- requires a Gibbs free energy is 355.5 kcal/mol [30]. With the interaction between FLPs and AB, it’’s different from the traditional study that the hydrogen is first formed from AB [3,5]. It is a complete heterolytic splitting of AB in the combination with FLPs, where the electrons transfer completely in the process. However, the decomposition of AB is after the formation of a dihydrogen
Please cite this article in press as: Wang K, et al., The mechanism of the chain-growth of ammoniaborane: A classic Lewis pairs catalysed by a Frustrated Lewis Pairs, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.01.002
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e9
bond, where the bond splitting is not a complete heterolytic cleavage and the electrons are partly transferring in the process. From [P1 þ AB], it is a synergetic step in TS1 where a dH(B) of AB moves toward to B1 atom of FLP, while the other fragment of [NH3BH2]þ is getting close to P2 atom through B3 atom of AB (the atomic labels are showed in Fig. 1 at the meantime with an energy barrier of 29.6 kcal/mol. It is a hydrogen migration from a traditional Lewis acid (B3 atom of AB) to a Frustrated Lewis acid (B1 atom of FLP) in TS1. The hydrogen anion generated from B3eH (the hydrogen atom connected with B3 atom) of AB moves to B1 atom (sp2 hybridization), which also polarizes the molecular FLP and increases the electron density on P2 atom (q(P2) ¼ 0.33 in P1, q(P2) ¼ 0.51 in TS1). This further enhances the interaction between N4 and B3 in P2¡f. There is an isomerization to form a more stable conformation (from P2¡f to P2) which is corresponding to the rotation of BeP bond (show in Figure S1), where the charges are reallocated in order to decrease the molecular polarity. The transformation of P2¡f/P2 reduces the chemical potential energy of 2.6 kcal/mol. The bond length of B3eN4 is shortened from 1.65 A in AB to 1.57 A in P2, where the zwitterionic character (B1/N) of P2 is enhanced for the further NeH activation, which is a necessary precursor of the next dehydrogenation for the continuous chain-growth processes. The distance of N4eH is slightly lengthened from 1.01 A (in AB) to 1.02 A (in P2). The averaged NBO charge of H(N4) (the hydrogen atom connected with N4 atom) increased from 0.40 e in molecular AB to 0.44 e in P2, which indicates the terminal N4eH bond is activated before dehydrogenation. The bond strength is evaluated by the Wiberg bond index (WBI) in P2, where WBI(N4eH) is 0.77, WBI(P2eH) is 0.95 and WBI(B1eH) is 0.91. The dehydrogenation of P2 starts from the activation of terminal N4eH bond, which is different from that of P0. Compared with WBI(P2eH) (0.94) and WBI(B1eH) (0.93) in P1, the reactivity of the new formed FLPs (between dþH(N4) and dH(B1) in P2¡f) is improved than that of the old one (dþH(P2) and dH(B1) in P1). Therefore, the terminal N4eH bond is activated to generate dþH(N4), which gets close to dH(B1) to generate hydrogen intramolecularly. The energy barrier between TS2 and P2 is 34.5 kcal/mol, about twice larger than the dehydrogenation barrier from P0 to P1, which is close to the barrier of the first dehydrogenation of AB (BH3NH3 / BH2NH2 þ H2), where the barrier of the reaction is 35.5 kcal/mol under the level of CCSD(T)/aVTZ [9]. The
5
heterolytic splitting of N4eH is the initial step in the dehydrogenation with Gibbs free energy of 13.18 kcal/mol under the level of M062X/6e311þþG(d,p), which is obtained by the similar designed reaction of Z-NH/[ZN] þ Hþ, where Z suggests the corresponding fragment of P2. The donoracceptor interaction of TS2 indicates the electrons are donated (i) from LP (Lone pairs) of N4 to p*(B1eH2) orbital, (ii) from p(B1eH2) orbital to N4 LP* orbital (Fig. 4). This interaction indicates the dehydrogenation is after the formation of the multi-center (3 centers with 2 electrons) bond of B1eH2, which is quite similar with the catalysis with transition metals [25]. And this is also consistent with the charge distribution of TS2. Additionally, our attempts to find the intermolecular dihydrogen bond to activate NeH bond between P2 and AB (or between P2 and the neighbouring P2) are both failed. We predict the dihydrogen bond is formed intramolecularly after terminal NeH activation. After the first dehydrogenation, there are two isomers (showed in Fig. 1) as P3 and P3¡1 with an energetic gap of 30.0 kcal/mol. A concept of ‘frustration’ indicates the strength between Lewis acid and Lewis base in FLPs [31]. We will use the gap between the FLP (P3) and the possible traditional Lewis pairs (P3¡1) to evaluate the ability of catalyst. Here the frustration of P3 is 30.0 kcal/mol. However, the further chaingrowth started from the structure of P3. In the 1st step of the BeN chain-growth, the reaction free energies of P1/P2¡f (5.0 kcal/mol) and P2f / P3þH2 (15.9 kcal/mol) both appear exothermic properties and excellent spontaneities.
The second and third steps of BN chain growth The next two steps of chain growth including the dehydrogenation both appear similar statements as Fig. 3 (or Fig. 5) shows. The reorientation of a new molecular AB to bind with the corresponding previous product (P3 or P5) decreases the chemical potential energies of 16.2 kcal/mol and 8.8 kcal/mol, where the B5eH (or B7eH) bond of the new AB starts to be activated in the corresponding intermediate of [P3þAB] and [P5þAB] (showed in Scheme 2). Then the molecular AB heterolytically splits into H(B)d and [NH3BH2]dþ to interact with FLP with the formation of a new B1eH bond and N4eB5 (or N6eB7) bond synergetically, which represents the second (or the third) step of BeN chain growth. The energy barriers of [P3þAB] / TS3 and [P5þAB] / TS5 are 7.4 kcal/mol and 16.8 kcal/mol, respectively. The relative
Fig. 4 e NLMOs of the multi-center bond of the transition states TS2. Please cite this article in press as: Wang K, et al., The mechanism of the chain-growth of ammoniaborane: A classic Lewis pairs catalysed by a Frustrated Lewis Pairs, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.01.002
6
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e9
Fig. 5 e The schematic potential energy surface of the second (upside) and third step (downside) in the process of B-N chain growth. The corresponding chemical structures are listed in Fig. 1. The optimized structures are listed in the supporting information. The optimized structures are listed in the supporting information (Section S1).
[P5+AB]
[P3+AB]
O H2P H2B
1
2
B(CF3)2
3
Ph2P 3
4
NH2
+
5
H3B
O
2
1
BH2 4
6
NH3
H2 N
H2 B 5
B(C6H5)2 7
NH2 6
+ H3B
8
NH3
Scheme 2 e The atomic labels of the intermediate of [P3þAB] and [P5þAB].
lower barriers of second and third BeN chain growth processes than that of the first one suggest that B/N FLPs are more ‘frustrated’ than that of B/P. The NBO charges of B1/N4 (or B1/ N6) are 0.18/-1.06 in [P3þAB] and 0.16/-1.09 in [P5þAB] which also demonstrates the interaction between the B1/N is stronger compared with that of B1/P2 of [P1þAB] (0.17/0.26). In addition, the reactivity of the FLP is attenuated with elongation of the distance between Lewis acid and Lewis base. (The A and 4.58 A, distances of B1eN4 in P3 and B1eN6 in P5 are 3.40 respectively). So there are two factors effecting the activation
of BeH bond of molecular AB, which are the intensity of ‘frustration’ and the distance of the corresponding acid/base in P3 and P5. In addition, the chemical potential reduces 10.1 (/28.3) kcal/mol from [P3/5þAB] to P3/5. Similar with the isomerization of P2¡f / P2, the isomerizations of P4f / P4 and P6¡f / P6 decrease the potential energies of 89.4 kcal/mol and 3.2 kcal/mol, respectively. Among the three transformations, the largest change of the potential energy in the transformation from P4¡f to P4 appears the irreversibility of the reaction. The two structures
Please cite this article in press as: Wang K, et al., The mechanism of the chain-growth of ammoniaborane: A classic Lewis pairs catalysed by a Frustrated Lewis Pairs, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.01.002
7
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e9
Wiberg bond index (WBI), NBO charge, the activation energy of B1eH and the energy Table 1 e The typical bond length (A), barrier of the corresponding dehydrogenation from P2, P4 and P6. Bond length ( A) WBI Ea (kcal/mol)
B1eH(B1) NeH(N) H(N)dþ … dH(B1)a Atomic Charge (e) P2 P4 P6 a b
P2
P4
P6
1.21 1.02 (N4eH) 2.08
1.21 1.02 (N6eH) 2.24
1.21 1.02 (N8eH) 3.13
0.91 0.76 (N4eH)
P2
B1 0.21 0.21 0.20
N 0.87(N4) 0.87(N6) 0.89(N8)
H(B1) 0.06 0.05 0.04
H(N) 0.46/0.45/0.42 0.45/0.43/0.42 0.44/0.43/0.40
P4
P6
P2-TS2
P4-TS4
P6-TS6
0.94 0.78 (N6eH)
0.94 0.79 (N8eH)
34.7
35.2
45.1
37.7b
50.8
50.9
The dihydrogen bond between H(B1) and the corresponding terminal H(N) atoms. The energy ofthe corresponding dehydrogenation steps between H(N)dþ … dH(B1).
Fig. 6 e The schematic potential energy surface of the cyclization of AB (in the box) and possible fourth step of B-N chain growth. The optimized structures are listed in the supporting information. Part of hypothesized chemical structures not including in Fig. 1 are listed here.
are showed in Figure S1. A coplanar ‘spoon-like’ conformation of P2eB3eN4eB5eN6 in P4¡f transforms to a more stable ‘chairelike’ structure in P4, which is very close to the experimental results [21]. The distance of N4eB1 is shorten A (P4), while the WBI(N4eH) defrom 4.27 A (P4¡f) to 3.60 creases from 0.82 (P4¡f) to 0.78 (P4). However, there is no obvious difference of the conformation between the P6¡f and P6 in our calculation (Figure S1). The distance of N6eB1 is shorten from 4.24 A to 3.96 A, while the WBI(N6eH) decreases from 0.82 (P4¡f) to 0.79 (P4). The atomic charges of H(N6)/H(B1) are 0.43/-0.14 e in P4¡f and 0.43/-0.05e in P4. Similarly, the charges of H(N8)/H(B1) are 0.43/-0.05 e in P6¡f and 0.43/-0.01 e
in P6, which suggests that the isomerization weaken the bond strength of terminal N4eH (or N6eH) for the continuous dehydrogenation. Started from P4¡f (P6¡f), the terminal NeH bond of the BeN chain is activated for the dehydrogenation and the continuous chain growth. The barriers of the second and third dehydrogenation are 50.8 kcal/mol and 50.9 kcal/mol, respectively. The bond strength of N4eB5 (N6eB7) bond in P4 (P6) is lower than that of P2eB3 in P2 caused by reallocation of the electron density with the growing length of BN chain. Furthermore, the energy barrier (Ea) of dehydrogenation and the distance of H(N4/6)dþ … dH(B1) have positive correlation. The atomic
Please cite this article in press as: Wang K, et al., The mechanism of the chain-growth of ammoniaborane: A classic Lewis pairs catalysed by a Frustrated Lewis Pairs, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.01.002
8
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e9
charges of H(B1) and H(N) are slightly reduced with the elongation of BN chain. So it is gradually difficult to activate the terminal NeH bond with the chain growing (The relevant parameters are listed in Table 1). The ‘frustration's of P5 and P7 are 36.1 and 41.4 kcal/mol, respectively.
The cyclization of cyclotriborazylene (CTB) There are two different possible pathways after the formation of P7 (Fig. 6). First, there is probably a cyclization of the BN chains to form CTB only requiring a barrier of 3.8 kcal/mol as the experimental results. Secondly, we hypothesis the continuous BN chain growth based on P7. The fourth similar combination of P7 and AB to generate [P7þAB] from P7 and AB structure reduces the potential energy of 17.3 kcal/mol. Then the B1eH activation of AB is 22.3 kcal/mol catalysed by P7. Compared the two pathways in Fig. 6, there is no doubt that the cyclization to form CTB is the preferential pathway, which is consistent with the experiment [21]. As the theoretical results, the continuous BN chain growth is terminated possibly caused by the low barrier of the cyclization (3.7 kcal/mol) and the instability of the growing BeN chain-like structure. The reaction Gibbs free energy of P7 / P7 þCTB is 22.9 kcal/mol in the cyclization. Based on the initial structures (P1/P3/P5/P7) of each steps of BN chain growth (Table S2), we fit an equation to evaluate the relationship among the activation energy of BeH of ammoniaborane (Ea(BeH)), the atomic charge differences(C) and the distances (D) between the Lewis acid (B1 atom) and Lewis base (terminal N atom): Ea (kcal/mol) ¼ 3.962D-15.12C þ 16.17, where the charge difference (or the frustration between acid/ base) is the main factor to influence the heterolytic splitting of AB.
Conclusion In summary, we have explored the mechanisms for the heterolytic cleavage of BeH (and NeH) bond in AB to form a cyclic structure by dimethylxanthene-derived Lewis pair (P1), where the pure cyclized product could be a key intermediate for the regeneration of AB. In the cyclization from AB to CTB, the activations of BeH and terminal NeH bonds of AB play quite important roles in the whole process. The chain growth and cyclization of AB is catalysed by the FLPs. The terminal BeH bond of the saturated ammoniaborane is first activated by FLPs, which lead to the continuous dehydrogenation. From our study, it is a synergetic step of heterolytic cleavage of BeH bond in AB, where the energy barrier can be described by a function of FLP's structure. The activation of the terminal H(N) of each BN chain leads to a continuous chaingrowth, where the dehydrogenation is experienced from a B1eH2 adducts after the migration of dþH(N) towards to d H(B1). However, it's more and more difficult to activate the terminal NeH bond with the elongation of BeN chains. The cyclization to form CTB is the most possible pathway after three BN chain growth steps caused by its lowest energy barriers. Our findings provide a blueprint for the regeneration of AB after the dehydrogenation and broaden the applicability of this concept in relevant catalyst design.
Acknowledgements The work is supported by National Natural Science Foundation of China (NSFC 21701001, 51625102), Anhui Provincial Natural Science Foundation (1708085QB42), Natural Science Research Project of Anhui Province (KJ2016A032) and the Opening Project of State Key Laboratory of Science and Technology (Beijing Institute of Technology, KFJJ16-11M). We also acknowledge Professor John E. McGrady from Oxford for his help in the work.
Appendix A. Supplementary data Supplementary data related to this article can be found at https://doi.org/10.1016/j.ijhydene.2018.01.002.
references
[1] Schlapbach L, Zu¨ttel A. Hydrogen-storage materials for mobile applications. Nature 2001:353e8. [2] He T, Pachfule P, Wu H, Xu Q, Chen P. Hydrogen carriers. Nat Rev Mater 2016:1e17. 16059. [3] Chua YS, Chen P, Wu GP, Xiong ZT. Development of amidoboranes for hydrogen storage. Chem Commun 2011;47:5116e29. [4] Xiong ZT, Yong CK, Wu GT, Chen P, Shaw W, Karkamkar A, et al. High-capacity hydrogen storage in lithium and sodium amidoboranes. Nat Mater 2008;7(2):138e41. [5] Wang K, Zhang J-G, Man T-T, Wu M, Chen C-C. Recent process and development of metal Aminoborane. Chem Asian J 2013;8:1076e89. [6] Kobayashi T, Gupta S, Caporini AM, Pecharsky KV, Pruski M. Mechanism of solid-state thermolysis of ammonia borane: a 15N NMR study using fast magic-Angle spinning and dynamic nuclear polarization. J Phys Chem C 2014;118(34): 19548e55. [7] Al-Kukhun A, Hwang HT, Varma A. Mechanistic studies of ammonia borane dehydrogenation. Int J Hydrogen Energy 2013;38:169e79. [8] Wang P. Solid-state thermolysis of ammonia borane and related materials for high-capacity hydrogen storage. Dalton Trans 2012;41:4296e302. [9] Nguyen TM, Nguyen VS, Matus HM, Gopakumar G, Dixon DA. Molecular mechanism for H2 release from BH3NH3, including the catalytic role of the Lewis acid BH3. J Phys Chem A 2007;111:679e90. [10] Fijalkowski KJ, Jurczakowski R, Kozminski W, Grochala W. Insights from impedance spectroscopy into the mechanism of thermal decomposition of M(NH2BH3), M ¼ H, Li, Na, Li0.5Na0.5, hydrogenstores. Phys Chem Chem Phys 2012;14:5778e84. [11] Smythe NC, Gordon JC. Ammonia borane as a hydrogen carrier: dehydrogenation and regeneration. Eur J Inorg Chem 2010;2010(4):499e656. [12] Sutton AD, Burrell AK, Dixon DA, Garner EB, Gordon JC, Nakagawa T, et al. Regeneration of ammonia borane spent fuel by direct reaction with hydrazine and liquid ammonia. Science 2011;331:1426e9. [13] Davis BL, Rekken BD, Michalczyk R, Garner EB, Dixon DA, Kalviri H, et al. Lewis base assisted B-H bond redistribution in borazine and polyborazylene. Chem Comm 2013;49(80):9095e7.
Please cite this article in press as: Wang K, et al., The mechanism of the chain-growth of ammoniaborane: A classic Lewis pairs catalysed by a Frustrated Lewis Pairs, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.01.002
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e9
[14] Nutt WR, McKee LM. Theoretical study of reaction pathways to borazine. Inorg Chem 2007;46:7633e45. [15] Morris RH. Bronsted-lowry acid strength of metal hydride and dihydrogen complexes. Chem Rev 2016;116(15):8588e654. [16] Longobardi LE, Zatsepin P, Korol R, Liu L, Grimme S, Stephan DW. Reactions of boron-derived radicals with nucleophiles. J Am Chem Soc 2017;139(1):426e35. [17] Stephan DW. The broadening reach of frustrated Lewis pair chemistry. Science 2016;354(6317):1248e50. [18] Welch GC, Juan RRS, Masuda JD, Stephan DW. Reversible, metal-free hydrogen activation. Science 2006;34:1124e6. [19] Erker G, Stephan DW. Frustrated Lewis pairs I: uncovering and understanding. Springer: Verlag Berlin Heidelberg; 2013. [20] Mummadi S, Unruh DK, Zhao J, Li S, Krempner C. “Inverse” frustrated Lewis pairseactivation of dihydrogen with organosuperbases and moderate to weak Lewis acids. J Am Chem Soc 2016;138(10):3286e9. [21] Metters OJ, Forrest SJ, Sparkes HA, Manners I, Wass DF. Small molecule activation by intermolecular Zr(IV)-Phosphine frustrated Lewis pairs. J Am Chem Soc 2016;138(6):1994e2003. [22] Li H, Aquino AJ, Cordes DB, Hung-Low F, Hase WL, Krempner C. A zwitterionic carbanion frustrated by boranesdihydrogen cleavage with weak Lewis acids via an “inverse” frustrated Lewis pair approach. J Am Chem Soc 2013;135(43):16066e9. [23] Mo ZB, Pit A, Campos J, Kolychev EL, Aldridge S. Catalytic B-N dehydrogenation using frustrated Lewis pairs: evidence for a chain-growth coupling mechanism. J Am Chem Soc 2016;138(10):3306e9.
9
[24] Lui MW, Paisle NR, Robert McDonald, ferguson MJ, rivard E. Metal-free dehydrogenation of amine-boranes by tunable NHeterocyclic iminoboranes. Chem Eur J 2016;22(6):2134e45. [25] Bhunya S, Malakar T, Ganguly G, Paul A. Combining protons and hydrides by homogeneous catalysis for controlling the release of hydrogen from ammonia-borane: present status and challenges. ACS Catal 2016;6(11):7907e34. [26] Truhlar DG, Zhao Y. 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 Accounts 2006;120(1e3):215e41. [27] Kendall RA, Thorn H. Dunning. Electron affinities of the firstrow atoms revisited. Systematic basis sets and wave functions. J Chem Phys 1992;96:6796e806. [28] Autrey T, Bowden M, Karkamkar A. Control of hydrogen release and uptake in amineborane molecular complexes: thermodynamics of ammonia borane, ammonium borohydride, and the diammoniate of diborane. Faraday Discuss 2011;151:157e69. [29] Matus HM, Grant JD, Nguyen MT, Dixon DA. Fundamental thermochemical properties of ammonia borane and dehydrogenated derivatives (BNHn, n ¼ 0-6). J Phys Chem C 2009;113:16553e60. [30] William JL. Modern inorganic chemistry. second ed. New York: McGraw-Hill; 1991, ISBN 0-07-112651-1. [31] Rokob TA, Hamza A, Stirling A, Soos T, Papai I. Turning frustration into bond activation: a theoretical mechanistic study on heterolytic hydrogen splitting by frustrated Lewis pairs. Angew Chem Int Ed 2008;47(13):2435e8.
Please cite this article in press as: Wang K, et al., The mechanism of the chain-growth of ammoniaborane: A classic Lewis pairs catalysed by a Frustrated Lewis Pairs, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.01.002
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/he
The mechanism of the chain-growth of ammoniaborane: A classic Lewis pairs catalysed by a Frustrated Lewis Pairs Kun Wang a,b,c,**, Long-jiu Cheng b, Jian-guo Zhang c, Xue-bin Yu a,* a
Department of Materials Science, Fudan University, Shanghai 200433, PR China Department of Chemistry, Anhui University, Hefei, Anhui 230601, PR China c State Key Laboratory of Explosion Science and Technology, Beijing Institute of Technology, Beijing 100081, PR China b
article info
abstract
Article history:
Ammonia borane (AB) is an outstanding hydrogen-storage material, which has been gained
Received 20 October 2017
many attentions. It can be regenerated from polyborazylene based on the experimental
Received in revised form
study, but the uncontrollable BeN dehydro-coupling in the polymerization from AB to
30 December 2017
borazine is still a problem for the continuous application of AB. Frustrated Lewis Pairs
Accepted 1 January 2018
(FLPs) is a landmark discovery to dissociate small molecules by activating strong nonpolar
Available online xxx
bonds reversibly, which is recently applied to directly polymerize three molecular AB to a cyclotriborazylene (CTB, B3N3H12, which transforms into borazine after releasing 3 equiv.
Keywords:
H2). Here we analyse how a dimethylxanthene-derived FLPs catalyse AB to CTB. The
Frustrated Lewis Pairs (FLPs)
cyclization experiences three steps by our calculation: the connection between FLPs and AB
Metal-free hydrogenation
(1), the growth of BeN chain (2) and an intermolecular dehydrogenation between terminal
Hydrogen-storage material
N-Hdþ and
Density functional theory
nation of AB and broadens the applicability of this concept in the relevant catalyst design.
Reaction kinetic study
d
HeB (3), which gives us a clue of a possible reversibility of the dehydroge-
© 2018 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
Introduction Hydrogen energy resource is a kind of clean energy where hydrogen is stored in condensed materials by physisorption or chemisorption as atomic or ionic states [1,2]. Boron-nitrogen compounds always appear potential application in the field of the solid state hydrogen-storage materials. Ammonia borane (AB), with a molecular formula of H3NBH3, was a star compound in the area of hydrogen storage materials owing to its favourable stability and high hydrogen content (up to 19.5 wt%) [3e5]. Lots of people studied the decomposition
pathway of AB and its derivate [6e8]. Dixon's results show e BH3 group catalyse the dehydrogenation of AB very efficiently and decrease its energy gap of 6.1 kcal/mol after the formation of BH3NH3eBH3 [9]. Based on the research of Shimoda and Fijalkowski, the major mechanism of decomposition is evolution of hydrogen by N-Hdþ … dHeB directly [10]. The strong dihydrogen bond also gives AB much higher volumetric density due to the classic intramolecular interaction between Lewis acid and Lewis base [11]. AB is proved to be regenerated from polyborazylene experimentally [12]. The redistribution of BeH bond of AB is demonstrated to be accomplished from
* Corresponding author. Department of Materials Science, Fudan University, Shanghai 200433, PR China. ** Corresponding author. Department of Chemistry, Anhui University, Hefei, Anhui 230601, PR China. E-mail addresses: [email protected] (K. Wang), [email protected] (X.-b. Yu). https://doi.org/10.1016/j.ijhydene.2018.01.002 0360-3199/© 2018 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved. Please cite this article in press as: Wang K, et al., The mechanism of the chain-growth of ammoniaborane: A classic Lewis pairs catalysed by a Frustrated Lewis Pairs, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.01.002
2
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e9
borazine theoretically [13]. Mckee and Nutt hypothesized a pathway to explain the cyclization of borazine from designed NH2BHNHBH2 and BH2NH2, where both the stepwise and synergetic pathway are based on the unsaturated bonds of Be N [14]. Here we hypothesize it is a key problem for the continuous application to avoid the spontaneous chaotic dehydrogenation in the oligomerization from AB to borazine. A lot of methods have been tried to solve the problem including tuning the characters of the dehydrocoupled products on the basis of Brǿnsted and Lewis theory [15,16]. Frustrated Lewis pairs (FLPs) is discovered by Stephen and Erker based on the Lewis acid-base theory [17e19]. The cooperative reactivities of Lewis acid and Lewis Base in FLPs are usually based on the steric hindrance in one molecule (or in two molecular fragments). The ultra-high activity is usually applied to activate small molecules, such as splitting HeH bond heterolytically [20e22]. Based on such characters, FLPs can be an excellent catalyst for the activation of the BeH (or NeH) bond of AB, which is just the key step for the dehydrogenation of AB. Recently, a metal-free Frustrated Lewis Pairs (FLPs) show success in catalysing the cyclization of three molecular NH2MeBH3 to form a cyclic triborazane under a quite mild condition in Aldridge's experiments (showed in Scheme 1) [23,24]. From the cyclized products (3 or 3′ in Scheme 1), it is possible to realize reversible formation of AB [13]. It is a clue for us to design a similar pathway for the generation of pure borazine from AB for the continuous regeneration, with avoiding the strong BeN coupling in the cyclization [25]. In this paper, we obtain the mechanism of the BN chain-growth and cyclization pathway from AB catalysed by dimethylxanthene-derived FLPs on the basis of Aldridge's
experiments [23]. We also analyse the possible reversibility of each step in the process of cyclization.
Computational methods The detailed mechanism of the cyclization catalysed by FLPs is predicted by using the gradient-corrected density functional theory with dunning basis set (M062x/cc-pVTZ) [26,27]. The reactions containing ions were calculated under the theoretical level of M062x/6e311þþG(d,p). In order to simplify the model to accelerate the calculation, we substitute the eCH3 and eC6F5 in the experimental structure [22] to eH and eCF3 respectively. The optimized structures and the corresponding coordinates are provided in the supporting information (Section S1 and S3). We explain the complete catalytic process from the first BeH activation of AB to the formation of CTB including the possible four processes of BeN chain-growth. We confirmed all the transition states by applying the intrinsic reaction coordination (IRC) method. The parallel calculations performed by Møller-plesset perturbation theory (mp2/6e311þþG(d,p)). The comparison of the total energy for each compound has been listed in the supporting information (Section S2 of SI). The results from both the two methods appear the same tendencies, which indicate our calculations are reasonable. The designed pathways are showed in Fig. 1. The enthalpies of each reaction under the same theoretical level are listed in supporting information (Table S1). We list the electronic enthalpies, Gibbs Free energies and Zero point energies (ZPE) of the stationary points and transition states including all the imaginary frequency of each transition state (Table S3).
H3BNH2Me (1 equiv.) 295 K
H3BNH2Me (1 equiv.) 328 K
Ph2P
-H2
O
O B(C6F5)2
2
Ph2P H
H B H
N
Me
O
B(C6F5)2
Ph2P
H
H
H
H B H
H
N
B
Me
H 2-(H)(BH2NH2Me)
B(C6F5)2 H N
H
H Me
2-(H)(BH2NHMeBH2NH2Me) H3BNH2Me (1 equiv.), 295 K -2H2
Me H
B N
Me
N B H 3'
B
H
N Me
2 (1 mol%) 328 K -3H2
H H
Me B
H N
B
H
H
H N B N Me H Me H H 3
+ O Ph2P
B(C6F5)2 2
Scheme 1 e The cyclization from Methyl substituted amine borane (NH2MeBH3) to cyclic triborazane (3) catalysed by dimethylxanthene-based FLPs (2) in Aldridge's experiments [23]. Please cite this article in press as: Wang K, et al., The mechanism of the chain-growth of ammoniaborane: A classic Lewis pairs catalysed by a Frustrated Lewis Pairs, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.01.002
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e9
3
Fig. 1 e The reversible dehydrogenation of NH3BH3 (AB) catalysed by dimethylxanthene-derived FLPs (P1). The numbers of B/ P/N are labelled (in red) on the structures of P1, P2, P4 and P7. (For interpretation of the references to color/colour in this figure legend, the reader is referred to the Web version of this article.)
Results and discussion Hydrogenation of dimethylxanthene-derived FLP The dimethylxanthene-derived FLP (P1) is optimized in Fig. 2. We also simulate the process of the heterolytic cleavage of hydrogen caused by the zwitterionic character of P1 showed in Fig. 2. The Natural Bond Orbital (NBO) charge of boron (B1 atom, the atomic labels are showed in Fig. 1) and phosphorus (P2 atom) are 0.65 and 0.33 e, which are dramatically polarized after the hydrogenation where the charge is 0.21 and 0.84 e, respectively. An NBO analysis of the interaction between hydrogen and P1 atom confirms that the electron transfer occurs through simultaneous s(H2) / B1 atom and P2 atom/ s*(H2) donations in a push-pull manner (Fig. 2). The energy barrier of hydrogenation is 15.9 kcal/mol, where the reaction is slightly endothermic with the enthalpy of 2.1 kcal/mol. The potential energy is showed in Fig. 3. We consider the hydride affinity (HA) of the Lewis acid (B1 atom in the group of -B(PH2C13H7O)(CF3)2) and proton affinity (PA) of Lewis base (P1 atom in the group of e PH2(B(CF3)2C13H7O)), which is important for the property of
catalysis. Considering the strong interaction between ions, we calculate the Gibbs free energies for HA and PA with a dispersion basis set of M062X/6e311þþG(d,p) [24,25]. As the results, HA of the Lewis acid is 116 kcal/mol, which appears more excellent spontaneity to interact with the hydride when comparing with Karkamkar's results from experiment [28] where the HA of BH3 group is 74.2 kcal/mol. In the similar systems, only the Lewis acid of BN appears a higher HA value in Dixon's calculation (160.3 kcal/mol [29]). Similarly, the electrophilicity of -P(Me)2 group of FLPs is evaluated by the value of PA, which is 201.7 kcal/mol. Caused by the electrondonating group of methyl, the PA value is larger than that of PH3 (188.7 kcal/mol [30]), which is almost close to that of NH3 (204 kcal/mol [28]). HA and PA of FLPs appear stronger acidity and basicity than that of traditional Lewis pairs (BH3/PH3). So it should be an excellent catalyst for the dehydrogenation of AB. Additionally, the Wiberg bond index (WBI) of B1eH and P2eH is 0.91 and 0.94, respectively. The slightly higher WBI indicates P2eH bond is more covalence than that of B1eH, which is also consistent with the lower formation Gibbs free energy of P2eH bond. It's also worth to note that the dehydrogenation from P0 is triggered by the activation of B1eH. But the interaction
Please cite this article in press as: Wang K, et al., The mechanism of the chain-growth of ammoniaborane: A classic Lewis pairs catalysed by a Frustrated Lewis Pairs, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.01.002
4
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e9
Fig. 2 e The optimized structure of P1 and the mechanism of its hydrogenation. Yellow: P; Red: O; Green: F; Pink: B; Grey: C; purple: H. NLMOs of the transition states TS0 are boxed. (For interpretation of the references to color/colour in this figure legend, the reader is referred to the Web version of this article.)
Fig. 3 e The schematic potential energy surface of the first B-N chain-growth step. The possible (de)hydrogenation of P1/0 is boxed on the left. The corresponding chemical structures are listed in Fig. 1. The optimized structures are listed in the supporting information (Section S1).
between the FLP and AB is more interesting and more complicated.
The first step of BN chain growth There are three steps of the BeN chain growth (including the cyclization). The potential energy surface of the first step is showed in Fig. 3. The corresponding structures and the typical bond lengths are showed in Fig. 1 and in the supporting information (Section S1). AB and FLP first bind together to form an intermediate of [P1 þ AB], where the formation of this intermediate ([P1 þ AB] / P1 þ AB) reduces the total potential energy of 18.2 kcal/mol. In [P1 þ AB], AB links to the FLP
through the Hd of BeH bonds of AB as a structure of B/H/B1. It is a typical s-type interaction between AB and FLP. The Gibbs free energy of heterolytic splitting BeH bond of AB is 213.1 kcal/mol by Dixon's calculation under the theoretical level of CCSD(T)/aVTZ [30] (NH3BH3 / H þ [NH3BH2]þ). With the same methods, the splitting of NeH bond as the hypothetic reaction of NH3BH3 / Hþ þ [NH2BH3]- requires a Gibbs free energy is 355.5 kcal/mol [30]. With the interaction between FLPs and AB, it’’s different from the traditional study that the hydrogen is first formed from AB [3,5]. It is a complete heterolytic splitting of AB in the combination with FLPs, where the electrons transfer completely in the process. However, the decomposition of AB is after the formation of a dihydrogen
Please cite this article in press as: Wang K, et al., The mechanism of the chain-growth of ammoniaborane: A classic Lewis pairs catalysed by a Frustrated Lewis Pairs, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.01.002
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e9
bond, where the bond splitting is not a complete heterolytic cleavage and the electrons are partly transferring in the process. From [P1 þ AB], it is a synergetic step in TS1 where a dH(B) of AB moves toward to B1 atom of FLP, while the other fragment of [NH3BH2]þ is getting close to P2 atom through B3 atom of AB (the atomic labels are showed in Fig. 1 at the meantime with an energy barrier of 29.6 kcal/mol. It is a hydrogen migration from a traditional Lewis acid (B3 atom of AB) to a Frustrated Lewis acid (B1 atom of FLP) in TS1. The hydrogen anion generated from B3eH (the hydrogen atom connected with B3 atom) of AB moves to B1 atom (sp2 hybridization), which also polarizes the molecular FLP and increases the electron density on P2 atom (q(P2) ¼ 0.33 in P1, q(P2) ¼ 0.51 in TS1). This further enhances the interaction between N4 and B3 in P2¡f. There is an isomerization to form a more stable conformation (from P2¡f to P2) which is corresponding to the rotation of BeP bond (show in Figure S1), where the charges are reallocated in order to decrease the molecular polarity. The transformation of P2¡f/P2 reduces the chemical potential energy of 2.6 kcal/mol. The bond length of B3eN4 is shortened from 1.65 A in AB to 1.57 A in P2, where the zwitterionic character (B1/N) of P2 is enhanced for the further NeH activation, which is a necessary precursor of the next dehydrogenation for the continuous chain-growth processes. The distance of N4eH is slightly lengthened from 1.01 A (in AB) to 1.02 A (in P2). The averaged NBO charge of H(N4) (the hydrogen atom connected with N4 atom) increased from 0.40 e in molecular AB to 0.44 e in P2, which indicates the terminal N4eH bond is activated before dehydrogenation. The bond strength is evaluated by the Wiberg bond index (WBI) in P2, where WBI(N4eH) is 0.77, WBI(P2eH) is 0.95 and WBI(B1eH) is 0.91. The dehydrogenation of P2 starts from the activation of terminal N4eH bond, which is different from that of P0. Compared with WBI(P2eH) (0.94) and WBI(B1eH) (0.93) in P1, the reactivity of the new formed FLPs (between dþH(N4) and dH(B1) in P2¡f) is improved than that of the old one (dþH(P2) and dH(B1) in P1). Therefore, the terminal N4eH bond is activated to generate dþH(N4), which gets close to dH(B1) to generate hydrogen intramolecularly. The energy barrier between TS2 and P2 is 34.5 kcal/mol, about twice larger than the dehydrogenation barrier from P0 to P1, which is close to the barrier of the first dehydrogenation of AB (BH3NH3 / BH2NH2 þ H2), where the barrier of the reaction is 35.5 kcal/mol under the level of CCSD(T)/aVTZ [9]. The
5
heterolytic splitting of N4eH is the initial step in the dehydrogenation with Gibbs free energy of 13.18 kcal/mol under the level of M062X/6e311þþG(d,p), which is obtained by the similar designed reaction of Z-NH/[ZN] þ Hþ, where Z suggests the corresponding fragment of P2. The donoracceptor interaction of TS2 indicates the electrons are donated (i) from LP (Lone pairs) of N4 to p*(B1eH2) orbital, (ii) from p(B1eH2) orbital to N4 LP* orbital (Fig. 4). This interaction indicates the dehydrogenation is after the formation of the multi-center (3 centers with 2 electrons) bond of B1eH2, which is quite similar with the catalysis with transition metals [25]. And this is also consistent with the charge distribution of TS2. Additionally, our attempts to find the intermolecular dihydrogen bond to activate NeH bond between P2 and AB (or between P2 and the neighbouring P2) are both failed. We predict the dihydrogen bond is formed intramolecularly after terminal NeH activation. After the first dehydrogenation, there are two isomers (showed in Fig. 1) as P3 and P3¡1 with an energetic gap of 30.0 kcal/mol. A concept of ‘frustration’ indicates the strength between Lewis acid and Lewis base in FLPs [31]. We will use the gap between the FLP (P3) and the possible traditional Lewis pairs (P3¡1) to evaluate the ability of catalyst. Here the frustration of P3 is 30.0 kcal/mol. However, the further chaingrowth started from the structure of P3. In the 1st step of the BeN chain-growth, the reaction free energies of P1/P2¡f (5.0 kcal/mol) and P2f / P3þH2 (15.9 kcal/mol) both appear exothermic properties and excellent spontaneities.
The second and third steps of BN chain growth The next two steps of chain growth including the dehydrogenation both appear similar statements as Fig. 3 (or Fig. 5) shows. The reorientation of a new molecular AB to bind with the corresponding previous product (P3 or P5) decreases the chemical potential energies of 16.2 kcal/mol and 8.8 kcal/mol, where the B5eH (or B7eH) bond of the new AB starts to be activated in the corresponding intermediate of [P3þAB] and [P5þAB] (showed in Scheme 2). Then the molecular AB heterolytically splits into H(B)d and [NH3BH2]dþ to interact with FLP with the formation of a new B1eH bond and N4eB5 (or N6eB7) bond synergetically, which represents the second (or the third) step of BeN chain growth. The energy barriers of [P3þAB] / TS3 and [P5þAB] / TS5 are 7.4 kcal/mol and 16.8 kcal/mol, respectively. The relative
Fig. 4 e NLMOs of the multi-center bond of the transition states TS2. Please cite this article in press as: Wang K, et al., The mechanism of the chain-growth of ammoniaborane: A classic Lewis pairs catalysed by a Frustrated Lewis Pairs, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.01.002
6
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e9
Fig. 5 e The schematic potential energy surface of the second (upside) and third step (downside) in the process of B-N chain growth. The corresponding chemical structures are listed in Fig. 1. The optimized structures are listed in the supporting information. The optimized structures are listed in the supporting information (Section S1).
[P5+AB]
[P3+AB]
O H2P H2B
1
2
B(CF3)2
3
Ph2P 3
4
NH2
+
5
H3B
O
2
1
BH2 4
6
NH3
H2 N
H2 B 5
B(C6H5)2 7
NH2 6
+ H3B
8
NH3
Scheme 2 e The atomic labels of the intermediate of [P3þAB] and [P5þAB].
lower barriers of second and third BeN chain growth processes than that of the first one suggest that B/N FLPs are more ‘frustrated’ than that of B/P. The NBO charges of B1/N4 (or B1/ N6) are 0.18/-1.06 in [P3þAB] and 0.16/-1.09 in [P5þAB] which also demonstrates the interaction between the B1/N is stronger compared with that of B1/P2 of [P1þAB] (0.17/0.26). In addition, the reactivity of the FLP is attenuated with elongation of the distance between Lewis acid and Lewis base. (The A and 4.58 A, distances of B1eN4 in P3 and B1eN6 in P5 are 3.40 respectively). So there are two factors effecting the activation
of BeH bond of molecular AB, which are the intensity of ‘frustration’ and the distance of the corresponding acid/base in P3 and P5. In addition, the chemical potential reduces 10.1 (/28.3) kcal/mol from [P3/5þAB] to P3/5. Similar with the isomerization of P2¡f / P2, the isomerizations of P4f / P4 and P6¡f / P6 decrease the potential energies of 89.4 kcal/mol and 3.2 kcal/mol, respectively. Among the three transformations, the largest change of the potential energy in the transformation from P4¡f to P4 appears the irreversibility of the reaction. The two structures
Please cite this article in press as: Wang K, et al., The mechanism of the chain-growth of ammoniaborane: A classic Lewis pairs catalysed by a Frustrated Lewis Pairs, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.01.002
7
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e9
Wiberg bond index (WBI), NBO charge, the activation energy of B1eH and the energy Table 1 e The typical bond length (A), barrier of the corresponding dehydrogenation from P2, P4 and P6. Bond length ( A) WBI Ea (kcal/mol)
B1eH(B1) NeH(N) H(N)dþ … dH(B1)a Atomic Charge (e) P2 P4 P6 a b
P2
P4
P6
1.21 1.02 (N4eH) 2.08
1.21 1.02 (N6eH) 2.24
1.21 1.02 (N8eH) 3.13
0.91 0.76 (N4eH)
P2
B1 0.21 0.21 0.20
N 0.87(N4) 0.87(N6) 0.89(N8)
H(B1) 0.06 0.05 0.04
H(N) 0.46/0.45/0.42 0.45/0.43/0.42 0.44/0.43/0.40
P4
P6
P2-TS2
P4-TS4
P6-TS6
0.94 0.78 (N6eH)
0.94 0.79 (N8eH)
34.7
35.2
45.1
37.7b
50.8
50.9
The dihydrogen bond between H(B1) and the corresponding terminal H(N) atoms. The energy ofthe corresponding dehydrogenation steps between H(N)dþ … dH(B1).
Fig. 6 e The schematic potential energy surface of the cyclization of AB (in the box) and possible fourth step of B-N chain growth. The optimized structures are listed in the supporting information. Part of hypothesized chemical structures not including in Fig. 1 are listed here.
are showed in Figure S1. A coplanar ‘spoon-like’ conformation of P2eB3eN4eB5eN6 in P4¡f transforms to a more stable ‘chairelike’ structure in P4, which is very close to the experimental results [21]. The distance of N4eB1 is shorten A (P4), while the WBI(N4eH) defrom 4.27 A (P4¡f) to 3.60 creases from 0.82 (P4¡f) to 0.78 (P4). However, there is no obvious difference of the conformation between the P6¡f and P6 in our calculation (Figure S1). The distance of N6eB1 is shorten from 4.24 A to 3.96 A, while the WBI(N6eH) decreases from 0.82 (P4¡f) to 0.79 (P4). The atomic charges of H(N6)/H(B1) are 0.43/-0.14 e in P4¡f and 0.43/-0.05e in P4. Similarly, the charges of H(N8)/H(B1) are 0.43/-0.05 e in P6¡f and 0.43/-0.01 e
in P6, which suggests that the isomerization weaken the bond strength of terminal N4eH (or N6eH) for the continuous dehydrogenation. Started from P4¡f (P6¡f), the terminal NeH bond of the BeN chain is activated for the dehydrogenation and the continuous chain growth. The barriers of the second and third dehydrogenation are 50.8 kcal/mol and 50.9 kcal/mol, respectively. The bond strength of N4eB5 (N6eB7) bond in P4 (P6) is lower than that of P2eB3 in P2 caused by reallocation of the electron density with the growing length of BN chain. Furthermore, the energy barrier (Ea) of dehydrogenation and the distance of H(N4/6)dþ … dH(B1) have positive correlation. The atomic
Please cite this article in press as: Wang K, et al., The mechanism of the chain-growth of ammoniaborane: A classic Lewis pairs catalysed by a Frustrated Lewis Pairs, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.01.002
8
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e9
charges of H(B1) and H(N) are slightly reduced with the elongation of BN chain. So it is gradually difficult to activate the terminal NeH bond with the chain growing (The relevant parameters are listed in Table 1). The ‘frustration's of P5 and P7 are 36.1 and 41.4 kcal/mol, respectively.
The cyclization of cyclotriborazylene (CTB) There are two different possible pathways after the formation of P7 (Fig. 6). First, there is probably a cyclization of the BN chains to form CTB only requiring a barrier of 3.8 kcal/mol as the experimental results. Secondly, we hypothesis the continuous BN chain growth based on P7. The fourth similar combination of P7 and AB to generate [P7þAB] from P7 and AB structure reduces the potential energy of 17.3 kcal/mol. Then the B1eH activation of AB is 22.3 kcal/mol catalysed by P7. Compared the two pathways in Fig. 6, there is no doubt that the cyclization to form CTB is the preferential pathway, which is consistent with the experiment [21]. As the theoretical results, the continuous BN chain growth is terminated possibly caused by the low barrier of the cyclization (3.7 kcal/mol) and the instability of the growing BeN chain-like structure. The reaction Gibbs free energy of P7 / P7 þCTB is 22.9 kcal/mol in the cyclization. Based on the initial structures (P1/P3/P5/P7) of each steps of BN chain growth (Table S2), we fit an equation to evaluate the relationship among the activation energy of BeH of ammoniaborane (Ea(BeH)), the atomic charge differences(C) and the distances (D) between the Lewis acid (B1 atom) and Lewis base (terminal N atom): Ea (kcal/mol) ¼ 3.962D-15.12C þ 16.17, where the charge difference (or the frustration between acid/ base) is the main factor to influence the heterolytic splitting of AB.
Conclusion In summary, we have explored the mechanisms for the heterolytic cleavage of BeH (and NeH) bond in AB to form a cyclic structure by dimethylxanthene-derived Lewis pair (P1), where the pure cyclized product could be a key intermediate for the regeneration of AB. In the cyclization from AB to CTB, the activations of BeH and terminal NeH bonds of AB play quite important roles in the whole process. The chain growth and cyclization of AB is catalysed by the FLPs. The terminal BeH bond of the saturated ammoniaborane is first activated by FLPs, which lead to the continuous dehydrogenation. From our study, it is a synergetic step of heterolytic cleavage of BeH bond in AB, where the energy barrier can be described by a function of FLP's structure. The activation of the terminal H(N) of each BN chain leads to a continuous chaingrowth, where the dehydrogenation is experienced from a B1eH2 adducts after the migration of dþH(N) towards to d H(B1). However, it's more and more difficult to activate the terminal NeH bond with the elongation of BeN chains. The cyclization to form CTB is the most possible pathway after three BN chain growth steps caused by its lowest energy barriers. Our findings provide a blueprint for the regeneration of AB after the dehydrogenation and broaden the applicability of this concept in relevant catalyst design.
Acknowledgements The work is supported by National Natural Science Foundation of China (NSFC 21701001, 51625102), Anhui Provincial Natural Science Foundation (1708085QB42), Natural Science Research Project of Anhui Province (KJ2016A032) and the Opening Project of State Key Laboratory of Science and Technology (Beijing Institute of Technology, KFJJ16-11M). We also acknowledge Professor John E. McGrady from Oxford for his help in the work.
Appendix A. Supplementary data Supplementary data related to this article can be found at https://doi.org/10.1016/j.ijhydene.2018.01.002.
references
[1] Schlapbach L, Zu¨ttel A. Hydrogen-storage materials for mobile applications. Nature 2001:353e8. [2] He T, Pachfule P, Wu H, Xu Q, Chen P. Hydrogen carriers. Nat Rev Mater 2016:1e17. 16059. [3] Chua YS, Chen P, Wu GP, Xiong ZT. Development of amidoboranes for hydrogen storage. Chem Commun 2011;47:5116e29. [4] Xiong ZT, Yong CK, Wu GT, Chen P, Shaw W, Karkamkar A, et al. High-capacity hydrogen storage in lithium and sodium amidoboranes. Nat Mater 2008;7(2):138e41. [5] Wang K, Zhang J-G, Man T-T, Wu M, Chen C-C. Recent process and development of metal Aminoborane. Chem Asian J 2013;8:1076e89. [6] Kobayashi T, Gupta S, Caporini AM, Pecharsky KV, Pruski M. Mechanism of solid-state thermolysis of ammonia borane: a 15N NMR study using fast magic-Angle spinning and dynamic nuclear polarization. J Phys Chem C 2014;118(34): 19548e55. [7] Al-Kukhun A, Hwang HT, Varma A. Mechanistic studies of ammonia borane dehydrogenation. Int J Hydrogen Energy 2013;38:169e79. [8] Wang P. Solid-state thermolysis of ammonia borane and related materials for high-capacity hydrogen storage. Dalton Trans 2012;41:4296e302. [9] Nguyen TM, Nguyen VS, Matus HM, Gopakumar G, Dixon DA. Molecular mechanism for H2 release from BH3NH3, including the catalytic role of the Lewis acid BH3. J Phys Chem A 2007;111:679e90. [10] Fijalkowski KJ, Jurczakowski R, Kozminski W, Grochala W. Insights from impedance spectroscopy into the mechanism of thermal decomposition of M(NH2BH3), M ¼ H, Li, Na, Li0.5Na0.5, hydrogenstores. Phys Chem Chem Phys 2012;14:5778e84. [11] Smythe NC, Gordon JC. Ammonia borane as a hydrogen carrier: dehydrogenation and regeneration. Eur J Inorg Chem 2010;2010(4):499e656. [12] Sutton AD, Burrell AK, Dixon DA, Garner EB, Gordon JC, Nakagawa T, et al. Regeneration of ammonia borane spent fuel by direct reaction with hydrazine and liquid ammonia. Science 2011;331:1426e9. [13] Davis BL, Rekken BD, Michalczyk R, Garner EB, Dixon DA, Kalviri H, et al. Lewis base assisted B-H bond redistribution in borazine and polyborazylene. Chem Comm 2013;49(80):9095e7.
Please cite this article in press as: Wang K, et al., The mechanism of the chain-growth of ammoniaborane: A classic Lewis pairs catalysed by a Frustrated Lewis Pairs, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.01.002
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y x x x ( 2 0 1 8 ) 1 e9
[14] Nutt WR, McKee LM. Theoretical study of reaction pathways to borazine. Inorg Chem 2007;46:7633e45. [15] Morris RH. Bronsted-lowry acid strength of metal hydride and dihydrogen complexes. Chem Rev 2016;116(15):8588e654. [16] Longobardi LE, Zatsepin P, Korol R, Liu L, Grimme S, Stephan DW. Reactions of boron-derived radicals with nucleophiles. J Am Chem Soc 2017;139(1):426e35. [17] Stephan DW. The broadening reach of frustrated Lewis pair chemistry. Science 2016;354(6317):1248e50. [18] Welch GC, Juan RRS, Masuda JD, Stephan DW. Reversible, metal-free hydrogen activation. Science 2006;34:1124e6. [19] Erker G, Stephan DW. Frustrated Lewis pairs I: uncovering and understanding. Springer: Verlag Berlin Heidelberg; 2013. [20] Mummadi S, Unruh DK, Zhao J, Li S, Krempner C. “Inverse” frustrated Lewis pairseactivation of dihydrogen with organosuperbases and moderate to weak Lewis acids. J Am Chem Soc 2016;138(10):3286e9. [21] Metters OJ, Forrest SJ, Sparkes HA, Manners I, Wass DF. Small molecule activation by intermolecular Zr(IV)-Phosphine frustrated Lewis pairs. J Am Chem Soc 2016;138(6):1994e2003. [22] Li H, Aquino AJ, Cordes DB, Hung-Low F, Hase WL, Krempner C. A zwitterionic carbanion frustrated by boranesdihydrogen cleavage with weak Lewis acids via an “inverse” frustrated Lewis pair approach. J Am Chem Soc 2013;135(43):16066e9. [23] Mo ZB, Pit A, Campos J, Kolychev EL, Aldridge S. Catalytic B-N dehydrogenation using frustrated Lewis pairs: evidence for a chain-growth coupling mechanism. J Am Chem Soc 2016;138(10):3306e9.
9
[24] Lui MW, Paisle NR, Robert McDonald, ferguson MJ, rivard E. Metal-free dehydrogenation of amine-boranes by tunable NHeterocyclic iminoboranes. Chem Eur J 2016;22(6):2134e45. [25] Bhunya S, Malakar T, Ganguly G, Paul A. Combining protons and hydrides by homogeneous catalysis for controlling the release of hydrogen from ammonia-borane: present status and challenges. ACS Catal 2016;6(11):7907e34. [26] Truhlar DG, Zhao Y. 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 Accounts 2006;120(1e3):215e41. [27] Kendall RA, Thorn H. Dunning. Electron affinities of the firstrow atoms revisited. Systematic basis sets and wave functions. J Chem Phys 1992;96:6796e806. [28] Autrey T, Bowden M, Karkamkar A. Control of hydrogen release and uptake in amineborane molecular complexes: thermodynamics of ammonia borane, ammonium borohydride, and the diammoniate of diborane. Faraday Discuss 2011;151:157e69. [29] Matus HM, Grant JD, Nguyen MT, Dixon DA. Fundamental thermochemical properties of ammonia borane and dehydrogenated derivatives (BNHn, n ¼ 0-6). J Phys Chem C 2009;113:16553e60. [30] William JL. Modern inorganic chemistry. second ed. New York: McGraw-Hill; 1991, ISBN 0-07-112651-1. [31] Rokob TA, Hamza A, Stirling A, Soos T, Papai I. Turning frustration into bond activation: a theoretical mechanistic study on heterolytic hydrogen splitting by frustrated Lewis pairs. Angew Chem Int Ed 2008;47(13):2435e8.
Please cite this article in press as: Wang K, et al., The mechanism of the chain-growth of ammoniaborane: A classic Lewis pairs catalysed by a Frustrated Lewis Pairs, International Journal of Hydrogen Energy (2018), https://doi.org/10.1016/j.ijhydene.2018.01.002