- Email: [email protected]
Improved dehydrogenation cycle performance of the 1.1MgH2-2LiNH2-0.1LiBH4 system by addition of LaNi4.5Mn0.5 alloy
JOURNAL OF RARE EARTHS, Vol. 33, No. 7, July 2015, P. 783
Improved dehydrogenation cycle performance of the 1.1MgH2-2LiNH2-0.1LiBH4 system by addition of LaNi4.5Mn0.5 alloy ZHAO Wang (赵 汪), JIANG Lijun (蒋利军)*, WU Yuanfang (武媛方), YE Jianhua (叶建华), YUAN Baolong (袁宝龙), LI Zhinian (李志念), LIU Xiaopeng (刘晓鹏), WANG Shumao (王树茂) (Energy Materials and Technology Research Institute, General Research Institute for Nonferrous Metals, Beijing 100088, China) Received 8 December 2014; revised 13 March 2015
Abstract: The isothermal desorption kinetics of the 1.1MgH2-2LiNH2-0.1LiBH4 system were improved by addition of LaNi4.5Mn0.5 alloy. The hydrogen desorption peak temperature of the sample containing LaNi4.5Mn0.5 reduced by approximately 5 K and the activation energy reduced by 9%. The results of isothermal dehydrogenation kinetics analysis implied that the isothermal desorption process at initial stage was controlled by the phase boundary mechanism. Moreover, the cycle performance of the materials was extended. The growth and agglomeration of the sample particles caused the deterioration of kinetics during de-/hydrogenation cycles, and then resulted in an incomplete desorption/absorption reaction which were responsible for the capacity fading. The cracking and pulverization of LaNi4.5Mn0.5 alloy had an obvious effect on preventing the composites aggregating, and the fine alloy particles could enhance the catalytic effect of the alloy, thus effectively offsetting part of the deterioration of kinetics caused by particles growth. Keywords: hydrogen storage materials; Li-Mg-B-N-H hydride; LaNi4.5Mn0.5; cycle performance; rare earths
Hydrogen is a green and efficient energy carrier, which is abundant and environmentally benign. One of key issues that limit hydrogen energy application is efficient storage of hydrogen[1]. Nowadays, many efforts have been made for development of new hydrogen storage materials worldwide[2–8]. Among the studied materials, the Li-Mg-N-H system comprised of LiH and Mg(NH2)2 or LiNH2 and MgH2, exhibits good reversibility at moderate operation temperatures and a relatively high capacity of 5.6 wt.%. It is therefore regarded as a promising candidate for on-board application[9]. This system can dehydrogenate under 0.1 MPa at 363 K by thermodynamic calculation[9]. However, a rather high kinetics barrier of hydrogen sorption has been identified, which restricts its application. A variety of dopants, KH, KOH, KF, lithium halides, RbF, NaOH, LaH3, VCl3, TiCl3, TaN, TiN, Li3AlH6, Ca(BH4)2, Mg(BH4)2, LiBH4, were investigated to improve the kinetic properties[10–26]. The alkali-metal compounds have been proved to be one of the most effective additives[10–18]. The hydrogen storage properties of Mg(NH2)2-2LiH were significantly enhanced by adding a small amount of KH[10–12]. The dehydrogenation peak temperature was lowered from 459 K for the Mg(NH2)2-2LiH sample to 405 K for the Mg(NH2)21.9LiH-0.1KH sample[10]. However, the positive effects of K-based additives disappear when the hydrogen release and uptake of the KF-added Mg(NH2)2- 2LiH samples are performed at higher temperatures (>473 K)[14,15].
It was found that metal borohydrides, such as Mg(BH4)2, Ca(BH4)2 and LiBH4, were another sort of effective dopants for improving the reaction kinetics of both dehydrogenation and hydrogenation of the Li-Mg-N-H system[23–26]. Yang et al. demonstrated that by starting with the mixture MgH2+LiNH2+LiBH4, the desorption reaction could proceed at lower temperatures and with significantly enhanced kinetics[26]. Via an in- depth exploration of the ternary mixture, a self-catalyzing reaction has been found, which improved the low temperature kinetics of the reversible reaction between Li2Mg(NH)2 and 2LiH+Mg(NH2)2[27]. The effect of the stoichiometry on the hydrogen storage property of this ternary complex hydride system has also been investigated[28,29]. To improve the hydrogen storage behavior of the Li-Mg-B-N-H system, various nanosized metals (Ni, Co, Fe, Cu, Mn) were added to this system. It was found that the additives of Co and Ni lowered the hydrogen releasing temperature at least 75–100 K in the major hydrogen decomposition step, while other additives acted as catalysts and increased the rate at which hydrogen was released[30]. With the addition of ZrCoH3, significant improvements in the hydrogen absorption/desorption properties of 2LiNH2-1.1MgH2-0.1LiBH4 composite have been achieved[31,32]. However, the researches on cycle performance of the Li-Mg-B-N-H system were rarely reported. As men-
Foundation item: Project supported by High-Tech Research and Development Program of China (2012AA051503) * Corresponding author: JIANG Lijun (E-mail: [email protected]; Tel.: +86-10-82241240) DOI: 10.1016/S1002-0721(14)60485-3
784
tioned above, the hydrogen storage properties of the LiMg-N-H system could be improved by adding LaH3[19]. Here, we studied the effect of the LaNi4.5Mn0.5 alloy on hydrogen sorption properties of the 2LiNH2-1.1MgH20.1LiBH4 composite. LaNi4.5Mn0.5 alloy is hard, brittle and easy to pulverize. The cycle performance and mechanism for hydrogen capacity fading was investigated.
1 Experimental The starting materials LiNH2 and LiBH4 (95% purity, Sigma-Aldrich) were used as received without purification. The MgH2 was home-made by ball-milling Mg powder under 4 MPa H2 pressure for 60 h, and then it was rehydrogenated for three times to obtain high-purity MgH2 (>95% purity). The LaNi4.5Mn0.5 intermetallic compound was re-melted three times by magnetic levitation melting under an argon atmosphere. The alloy was annealed in evacuated quartz tubes at 1273 K for 8 h, and then quenched in the water. The annealed alloy was hydrogenated for pulverising to –500 mesh powder under 5 MPa H2 pressure at ambient temperature. 3 g mixture of MgH2, LiNH2 and LiBH4 with molar ratio of 1.1:2:0.1 was put into a stainless steel vial for highenergy ball-milling under 4 MPa H2 pressure using a Spex-8000 apparatus. The weight ratio of stainless steel ball to powder was 20:1 and the total milling time was 36 h. The composite with an addition of 10 wt.% LaNi4.5Mn0.5 was prepared by high-energy ball milling LaNi4.5Mn0.5 hydride powder and the as-milled Li-Mg-B-N-H powder for 12 h. In order to reduce the amorphous phase produced by ball-milling, the stainless steel balls were replaced with zirconia balls and the weight ratio of ball to powder was 2:1. In case of the alloy with partially amorphous phase, the p-c isotherms differ significantly from the alloy without amorphous phase with an enhancement of solid state solubility and a lowering of the plateau pressure[33]. Isothermal dehydrogenation kinetics were measured using a Sieverts-type apparatus. Approximately 0.5 g of sample was used in each measurement. Differential scanning calorimetry (DSC) measurements were performed on a Netzsch DSC 409 unit. About 40 mg of samples were heated at a heating rate of 5 K/min under Ar atmosphere. The gas generated in the heating process was monitored by a mass spectrometer (MS). XRD analysis was carried out by using an X’pert Pro MPD diffractometer with Cu Kα radiation at 40 kV and 40 mA. The IR absorption spectrum was collected in diffuse reflectance infrared Fourier transform mode at a resolution of 4 cm–1, and the sample was mixed with paraffine in an appropriate proportion in order to isolate air. The morphology and element distributions of the composites
JOURNAL OF RARE EARTHS, Vol. 33, No. 7, July 2015
were measured by a field emission scanning electron microscope (SEM, Hitachi model S4800). All handling procedures in this work were performed in a glove-box filled with purified argon to keep the H2O and O2 levels below 1 ppm.
2 Results and discussion 2.1 Effects of LaNi4.5Mn0.5 on dehydrogenation kinetics and thermal effects Desorption isotherms were first measured for comparison of the effects induced by the addition of LaNi4.5Mn0.5. Fig. 1 presents hydrogen desorption kinetics curves under 0.1 MPa at 423 K of the Li-Mg-B- N-H system with and without LaNi4.5Mn0.5 addition. The dehydrogenation curves indicate that the desorption kinetics are enhanced by mixing 10 wt.% LaNi4.5Mn0.5 alloy. It has been found that the hydrogen release capacity is 3.12 wt.% and 3.62 wt.% within 120 min for the Li-MgB-N-H sample and the sample with LaNi4.5Mn0.5 addition respectively. However, the capacity loss for the sample added with LaNi4.5Mn0.5 is exhibited, due to the larger density and the smaller capacity of LaNi4.5Mn0.5, compared with the pristine sample of Li-Mg-B-N-H. Fig. 2(a) shows the DSC curves of the Li-Mg-B-N-H and 10 wt.% LaNi4.5Mn0.5 samples which were measured at a heating rate of 5 K/min from 303 to 623 K under argon atmosphere. The exothermic peaks at about 423 K coincide with the metathesis reaction that was observed by Luo et al.[34], which is very close to the results reported by Hu et al.[32]. The broad endothermic peaks in the temperature range of 463–483 K are obviously formed by the superposition of two peaks, which correspond to the two hydrogen desorption reactions reported by Yang et al.[27] in Eqs. (1) and (2) respectively. 2Li4BN3H10+3MgH2→3Li2Mg(NH)2+2LiBH4+6H2 (1) Mg(NH2)2+2LiH→Li2Mg(NH)2+2H2 (2) Li4BN3H10 should be formed by the reaction of LiNH2
Fig. 1 Hydrogen desorption kinetics curves under 0.1 MPa at 423 K of the Li-Mg-B-N-H sample and the 10 wt.% LaNi4.5Mn0.5 added sample
ZHAO Wang et al., Improved dehydrogenation cycle performance of the 1.1MgH2-2LiNH2-0.1LiBH4 system by …
785
alloy lowered desorption barrier, which is responsible for the reduction in the hydrogen desorption temperature and the enhancement in the hydrogen desorption kinetics. As we can see from the results of dehydrogenation kinetics and DSC-MS, the intermetallic compound LaNi4.5Mn0.5 plays a role of the catalyst to lower the kinetics barrier for dehydrogenation of the Li-Mg-B-N-H, reduces the dehydrogenation temperature and accelerates the rate of dehydrogenation effectively. 2.2 Isothermal dehydrogenation kinetics analysis
Fig. 2 DSC curves of the Li-Mg-B-N-H sample and the 10 wt.% LaNi4.5Mn0.5 added sample (a) and MS monitoring of H2 and NH3 generated during temperature programmed desorption (b) (The samples were heated at a heating rate of 5 K/min from 303 to 623 K)
and LiBH4 during sample preparation[27]. The peak temperature of the 10 wt.% LaNi4.5Mn0.5 sample is 466.7 K which is 4.8 K lower than that of the Li-Mg-B-N-H sample. It can be found that the positions of the H2 signal peaks correspond well with that of endothermic peaks. Compared to the pristine sample, the hydrogen desorption peak temperature of the sample with addition of LaNi4.5Mn0.5 reduces by approximately 5 K. The amount of NH3 release is extremely small compared to the amount of H2 release throughout the heating process. Differential scanning calorimetry (DSC) was also conducted to study the dehydrogenation activation energy of the Li-Mg-B-N-H and 10 wt.% LaNi4.5Mn0.5 added samples. The two Kissinger plots were drawn to estimate the activation energies (Fig. 3)[35]. The peak temperature data were obtained from the first peak. The activation energies by the linear fitting were found to be 81.0 kJ/mol for the Li-Mg-B-N-H sample and 73.8 kJ/mol for the 10 wt.% LaNi4.5Mn0.5 added sample, in other words, about 9 % reduction in the activation energy by the addition of LaNi4.5Mn0.5. This means that the addition of LaNi4.5Mn0.5
Fig. 3 Kissinger plots of the Li-Mg-B-N-H sample and the 10 wt.% LaNi4.5Mn0.5 sample
The isothermal dehydrogenation curves of the 10 wt.% LaNi4.5Mn0.5 added sample under 0.1 MPa H2 were further measured at 400–423 K, and the results are shown in Fig. 4. The dehydrogenation is dramatically decelerated as the temperature is decreased, and the hydrogen capacity is reduced. For understanding the details of dehydrogenation reaction kinetics, the isothermal data were analyzed in depth by using the Johnson-Mehl-Avrami (JMA) equation[36]. In practical application, the equation is generally rearranged as ln[–ln(1–α)]=n ln(t)+n ln(k) (3) In Eq. (3), α is the fraction reacted at time t, k is the rate constant, and n is the Avrami exponent. Fig. 5 shows the JMA plots for the isothermal dehydrogenation of the 10 wt.% LaNi4.5Mn0.5 added sample at various temperatures. There is a good linear relationship between ln[–ln(1–α)]
Fig. 4 Isothermal dehydrogenation curves under 0.1 MPa H2 of the 10 wt.% LaNi4.5Mn0.5 sample at various temperatures
Fig. 5 JMA plots for the isothermal dehydrogenation of the 10 wt.% LaNi4.5Mn0.5 sample at various temperatures
786
JOURNAL OF RARE EARTHS, Vol. 33, No. 7, July 2015
and ln(t) in the range of α=0–0.3, indicating that the dehydrogenation behaviors can be reasonably described by the JMA equation. Hancock et al. reported that this method applies not only to the common nucleation and growth equations (n=1, 2, 3), but also to non-integral values of n[37]. The certain values of Avrami exponent n represent the different nucleation and growth processes. For example, in phase-boundary reaction controlled mechanism, n is typically between 1.07–1.24. The values of the Avrami exponent (n) obtained by fitting the data are in the range of 0.9–1.1 for the dehydrogenation for 10 wt.% LaNi4.5Mn0.5 added sample, which is close to the value of n in interface controlled reactions. This reflects that the rate controlling step for the desorption process should be interface reaction. The reaction modeling equations (g(α)) related to the phase-boundary mechanism were applied to examine the mechanism underlying the dehydrogenation process. The expression of g(α) for phase-boundary mechanism is g(α)=1–(1–α)1/m=kt (4) The values of m are 1, 2 and 3, corresponding to the one-dimension, two-dimension and three-dimension phase-boundary mechanism, respectively. If the proposed mechanism was controlling the rates, then according to Eq. (4), a plot of g(α) versus time t should be linear. After fitting calculation, the model of one-dimension phase-boundary mechanism (m=1) versus time exhibits best linearity with linearity coefficient R2>0.99 in the range of α=0–0.4, as shown in Fig. 6. This implied that the isothermal desorption process at initial stage is controlled by the phase boundary mechanism. This result is in good agreement with the previous investigation reported by Chen et al.[38]. According to this mechanism, increasing the contact area of initial reactants will enhance the reaction rate. As a general rule, decreasing the particle size will increase the contact area. The addition of LaNi4.5Mn0.5 alloy that is hard could favor material pulverization during ball milling process and consequently quicken dehydrogenation kinetics.
2.3 Cycle performance of Li-Mg-B-N-H system with and without admixing of LaNi4.5Mn0.5 at 423 K Fig. 7 presents the capacities of 10 consecutive desorption cycles of the pristine sample and the 10 wt.% LaNi4.5Mn0.5 sample at 423 K. Compared to the desorption capacity of the first cycle, there is an evident decline in the second one for both the pristine sample and the 10 wt.% LaNi4.5Mn0.5 sample. The reason might be that the growth and/or agglomeration of the particle occurs after the first cycle which causes the degradation of the subsequent desorption kinetics and reduction in capacity within the stipulated time. In the next eight cycles, the capacity faded slowly, and the average decay rate is 0.063 wt.% per cycle for the Li-Mg-B-N-H sample, while the decay rate is 0.054 wt.% per cycle for the 10 wt.% LaNi4.5Mn0.5 sample. It has been proved that there is a slight improvement in the cycle performance by addition of LaNi4.5Mn0.5. More work is needed to understand the mechanism of the improvement. The deterioration of reaction kinetics after 10 cycles led to an incomplete des-/absorption reaction and consequently a capacity loss (this can be drawn from the hydrogen desorption kinetics curves which are not given in the paper). This phenomena is in good agreement with the findings reported by Hu et al.[32]. However, there may be some other reasons for capacity reduction, for example, NH3 formation[39], separation of the reactants by products and so on. This requires more in-depth studies to confirm them. In the first 10 cycles, the de-/hydrogenation time were 20 and 4 h, respectively. While shortening the time to 9 and 2 h in the next 10 cycles, it was found that the hydrogen capacity decayed faster (Fig. 8). The average decay rate of desorption capacities within 9 h increased from 0.051 wt.% to 0.064 wt.% per cycle. The short cyclic test time especially the hydrogenation time induces a more incomplete reaction, and brought about a larger decay rate.
Fig. 6 Time dependence of kinetic modeling equation g(α) (one-demension phase-boundary mechanism) for the dehydrogenation of the 10 wt.% LaNi4.5Mn0.5 added sample at various temperatures
Fig. 7 H2 desorption capacities of the Li-Mg-B-N-H sample and the 10 wt.% LaNi4.5Mn0.5 added sample (dehydrogenated under 0.1 MPa H2 pressure for 20 h and hydrogenated under 7.5 MPa H2 pressure for 4 h at 423 K)
ZHAO Wang et al., Improved dehydrogenation cycle performance of the 1.1MgH2-2LiNH2-0.1LiBH4 system by …
787
To shed light on the hydrogen capacity fading mechanism for the Li-Mg-B-N-H sample and the 10 wt.% LaNi4.5Mn0.5 added sample, the phase analysis, structure characterization and morphological observation were conducted. Fig. 9 presents the XRD patterns of the 10 wt.% LaNi4.5Mn0.5 added samples before and after dehydrogenation/hydrogenation cycles. For the ball-milled sample, the original materials, MgH2, LiNH2, LiBH4, cannot be identified for the amorphization behavior of the grains by ball-milling. The diffraction peak intensity of LaNi4.5Mn0.5 hydride significantly weakens. For the first desorption sample, only the Li2Mg(NH)2 and LaNi4.5Mn0.5 can be identified, and only Mg(NH2)2, LiH, LaNi4.5Mn0.5Hx for the first absorption sample, which manifest a good reversibility of dehydrogenation/hydrogenation of the sample added with 10 wt.% LaNi4.5Mn0.5. After 20 cycles, the reactant Mg(NH2)2 and LiH can be found in the dehydrogenated sample, and Li2Mg(NH)2 can be found in the hydrogenated sample as well. These findings provide powerful evidence for an incomplete desorption/absorption
reaction as mentioned above. No newly formed phase can be identified from these results in the sample with LaNi4.5Mn0.5 addition. It means that no new phase forms by the interaction between LaNi4.5Mn0.5 and Li-Mg-B-N-H composite. It is noted that the diffraction peak of LaNi4.5Mn0.5 shifts to the left slightly for the dehydrogenated sample (both the first and the twentieth) compared to the original LaNi4.5Mn0.5 alloy which suggests that there are residual H atoms in the alloy and the LaNi4.5Mn0.5 hydride does not dehydrogenate completely, because the hydrogen desorption pressure is 0.1 MPa instead of vacuum. Fig. 10 presents the FTIR spectra of the 10 wt.% LaNi4.5Mn0.5 added samples before and after dehydrogenation/hydrogenation cycles. For the ball-milled sample, the absorbance at 3325 and 3271 cm–1 suggests the existence of Mg(NH2)2 and the signature N–H vibration of starting material LiNH2 at 1563 cm–1 has also been found, which indicates an incomplete metathesis reaction[32]. In the hydrogenated samples (both the first and the twentieth) the absorbance at 1563 cm–1 disappears, and the signature N–H vibration of Mg(NH2)2 at 1570 cm–1 appears. In the dehydrogenated sample, the N–H vibration at 3168 cm–1 indicates the formation of the ternary imide Li2Mg(NH)2, and the absorbance at 3300 and 3241 cm–1 suggests the existence of the solid solution Li4BN3H10, which has been reported by Hu et al.[40]. However, the quaternary hydride Li4BN3H10 cannot be identified in the XRD patterns maybe owing to its low concentration. The existence of Li4BN3H10 with low melting point can be helpful to the growth and agglomeration of powder particles[41]. In the samples after twenty cycles, the absorbance at 3168 cm–1 corresponding to Li2Mg(NH)2 and 3325, 3271 cm–1 corresponding to Mg(NH2)2 become weaker compared to that in the samples before cycles. It can be inferred that the concentration of dehydrogenation products in desorption sample and hydrogenation products in absorption decreases, which signifies an incomplete desorption/absorption reaction within the limited time. Fig. 11 shows the morphologies of the 10 wt.%
Fig. 9 XRD patterns of the 10 wt.% LaNi4.5Mn0.5 added samples before and after dehydrogenation/hydrogenation cycles
Fig. 10 FTIR spectra of the 10 wt.% LaNi4.5Mn0.5 samples before and after dehydrogenation/hydrogenation cycles
Fig. 8 H2 desorption capacities within 9 h of the 10 wt.% LaNi4.5Mn0.5 added sample in 20 desorption cycles (dehydrogenated under 0.1 MPa H2 for 20 h and hydrogenated under 7.5 MPa H2 for 4 h in the first 10 cycles at 423 K, dehydrogenated for 9 h and hydrogenated for 2 h in the next 10 cycles)
2.4 Mechanism for hydrogen capacity fading
788
JOURNAL OF RARE EARTHS, Vol. 33, No. 7, July 2015
Fig. 11 SEM (top) and back scattered electron (bottom) images of the 10 wt.% LaNi4.5Mn0.5 added samples (a, c) As-milled; (b, d) After cycles
LaNi4.5Mn0.5 added samples before and after dehydrogenation/hydrogenation cycles. The particles are relatively uniform in size and some agglomerations can be found in the as-milled sample (Fig. 11(a), top). For the sample after cycles, the initially separated particles are remarkably aggregated and the particle size is augmented (Fig. 11(b), top). As mentioned above, for the heterogeneous solid-state of Li-Mg- N-H system, the reaction rate is controlled by the interface reaction in the early stage of the reaction. And in the later stage, it is controlled by the mass transport through the imide layer[38]. Neglecting degree of mixing of reacting species for the samples before and after cycles, the increscent particle size is responsible for the deterioration of reaction kinetics. From the back scattered electron images (Fig. 11 bottom), the particle size of LaNi4.5Mn0.5 alloy is about 1–2 μm, and the alloy particles were dehiscent and pulverized after 20 cycles. Many small alloy particles are released, which can increase the contact area between the alloy and base material and enhance the catalytic effect of the alloy, and then effectively offset part of the deterioration of kinetics caused by particles growth. For this reason, there is an improvement in the cycle performance by addition of LaNi4.5Mn0.5.
3 Conclusions The Li-Mg-B-N-H samples with and without addition
of LaNi4.5Mn0.5 were characterized by Sievert’s method for the hydrogen desorption, DSC-MS, XRD, FTIR, and SEM. The LaNi4.5Mn0.5 alloy could enhance the desorption kinetics and reduce the dehydrogenation peak temperature of Li-Mg-B-N-H system. The intermetallic compound LaNi4.5Mn0.5 played a role of the catalyst to lower the kinetics barrier for dehydrogenation of the Li-MgB-N-H. The results of isothermal dehydrogenation kinetics analysis implied that the isothermal desorption process at initial stage was controlled by the phase boundary mechanism. The increscent size and the agglomeration of the sample particles caused the deterioration of kinetics, and then the incomplete desorption/absorption reaction within limited time was a crucial reason for capacity fading. The LaNi4.5Mn0.5 alloy particles were dehiscent and pulverized after cycles as we expected. The pulverization increased the dispersion degree of the alloy, which could enhance the catalytic effect of the alloy, and then partly offset the deterioration of kinetics caused by particles growth and agglomeration. For this reason, there was an improvement in the cycle performance by addition of LaNi4.5Mn0.5 alloy.
References: [1] Schlapbach L, Züttel A. Hydrogen-storage materials for mobile applications. Nature, 2001, 414(6861): 353.
ZHAO Wang et al., Improved dehydrogenation cycle performance of the 1.1MgH2-2LiNH2-0.1LiBH4 system by … [2] Oelerich W, Klassen T, Bormann R. Metal oxides as catalysts for improved hydrogen sorption in nanocrystalline Mg-based materials. J. Alloys Compd., 2001, 315(1): 237. [3] Liu Y F, Liang C, Zhou H, Gao M X, Pan H G, Wang Q D. A novel catalyst precursor K2TiF6 with remarkable synergetic effects of K, Ti and F together on reversible hydrogen storage of NaAlH4. Chem. Commun., 2011, 47(6): 1740. [4] Guo Y H, Yu X B, Gao L, Xia G L, Guo Z P, Liu H K. Significantly improved dehydrogenation of LiBH4 destabilized by TiF3. Energy & Environmental Science, 2010, 3(4): 464. [5] Xia G L, Leng H Y, Xu N X, Li Z L, Wu Z, Du J L, Yu X B. Enhanced hydrogen storage properties of LiBH4-MgH2 composite by the catalytic effect of MoCl3. Int. J. Hydrogen Energy, 2011, 36(12): 7128. [6] Liu J, Li Q. Hydriding/dehydriding behaviors of La2Mg1710 wt.% Ni composite prepared by mechanical milling. J. Rare Earths, 2013, 31(1): 73. [7] Chen P, Xiong Z, Luo J, Lin J, Tan K L. Interaction of hydrogen with metal nitrides and imides. Nature, 2002, 420(6913): 302. [8] Xiong Z T, Wu G T, Hu JJ, Chen P. Ternary imides for hydrogen storage. Adv. Mater., 2004, 16(17): 1522. [9] Xiong Z T, Hu JJ, Wu G T, Chen P, Luo W F, Cross K, Wang J. Thermodynamic and kinetic investigations of the hydrogen storage in the Li-Mg-N-H system. J. Alloys Compd., 2005, 398(1): 235. [10] Wang J H, Liu T, Wu G T, Li W, Liu Y F, Araujo C M, Scheicher R H, Blomqvist A, Ahuja R, Xiong Z T, Yang P, Gao M X, Pan H G, Chen P. Potassium-modified Mg(NH2)2/2LiH system for hydrogen storage. Angew. Chem. Int. Ed., 2009, 48(32): 5828. [11] Li C, Liu Y F, Pang Y P, Gu Y J, Gao M X, Pan H G. Compositional effects on the hydrogen storage properties of Mg(NH2)2- 2LiH-xKH and the activity of KH during dehydrogenation reactions. Dalton Transactions, 2014, 43(6): 2369. [12] Li C, Liu Y F, Ma R J, Zhang X, Li Y, Gao M X, Pan H G. Superior dehydrogenation/hydrogenation kinetics and long- term cycling performance of K and Rb cocatalyzed Mg(NH2)2‐2LiH system. Appl. Mater. Interfaces, 2014, 6(19): 17024. [13] Liang C, Liu Y F, Gao M X, Pan H G. Understanding the role of K in the significantly improved hydrogen storage properties of a KOH-doped Li-Mg-N-H system. J. Mater. Chem. A, 2013, 1(16): 5031. [14] Liu Y F, Li C, Li B, Gao M X, Pan H G. Metathesis reaction-induced significant improvement in hydrogen storage properties of the KF-added Mg(NH2)2−2LiH system. J. Phys. Chem. C, 2013, 117(2): 866. [15] Li C, Liu Y F, Yang Y J, Gao M X, Pan H G. High-temperature failure behaviour and mechanism of K-based additives in Li-Mg-N-H hydrogen storage systems. J. Mater. Chem. A, 2014, 2(20): 7345. [16] Li B, Liu Y F, Li C, Gao M X, Pan H G. In situ formation of lithium fast-ion conductors and improved hydrogen desorption properties of the LiNH2-MgH2 system with the addition of lithium halides. J. Mater. Chem. A, 2014, 2(9):
789
3155. [17] Li C, Liu Y F, Gu Y J, Gao M X, Pan H G. Improved hydrogen-storage thermodynamics and kinetics for an RbF-doped Mg(NH2)2-2LiH system. Chemistry–An Asian Journal, 2013, 8(9): 2136. [18] Liang C, Liu Y F, Wei Z J, Jiang Y, Wu F, Gao M X, Pan H G. Enhanced dehydrogenation/hydrogenation kinetics of the Mg(NH2)2-2LiH system with NaOH additive. Int. J. Hydrogen Energy, 2011, 36(3): 2137. [19] Zhu X L, Han S M, Zhao X, Li Y, Liu B Z. Effect of lanthanum hydride on microstructures and hydrogen storage performances of 2LiNH2-MgH2 system. J. Rare Earths, 2014, 32(5): 429. [20] Shahi R R, Yadav T P, Shaz M A, Srivastva O N. Studies on dehydrogenation characteristic of Mg(NH2)2/LiH mixture admixed with vanadium and vanadium based catalysts (V, V2O5 and VCl3). Int. J. Hydrogen Energy, 2010, 35(1): 238. [21] Ma L P, Wang P, Dai H B, Cheng H M. Catalytically enhanced dehydrogenation of Li-Mg-N-H hydrogen storage material by transition metal nitrides. J. Alloys Compd., 2009, 468(1): L21. [22] Cao H J, Zhang Y, Wang J H, Xiong Z T, Wu G T, Qiu J S, Chen P. Effects of Al-based additives on the hydrogen storage performance of the Mg(NH2)2-2LiH system. Dalton Transactions, 2013, 42(15): 5524. [23] Li B, Liu Y F, Gu J, Gu Y J, Gao M X, Pan H G. Mechanistic investigations on significantly improved hydrogen storage performance of the Ca(BH4)2-added 2LiNH2/MgH2 system. Int. J. Hydrogen Energy, 2013, 38(12): 5030. [24] Li B, Liu Y F, Gu J, Gao M X, Pan H G. Synergetic effects of in situ formed CaH2 and LiBH4 on hydrogen storage properties of the Li-Mg-N-H system. Chemistry–An Asian Journal, 2013, 8(2): 374. [25] Pan H G, Shi S B, Liu Y F, Li B, Yang Y J, Gao M X. Improved hydrogen storage kinetics of the Li-Mg-N-H system by addition of Mg(BH4)2. Dalton Transactions, 2013, 42(11): 3802. [26] Yang J, Sudik A, Siegel D J, Halliday D, Drews A, Carter III R O, Wolverton C, Lewis G J, Sachtler J WA, Low J J, Faheem S A, Lesch D A, Ozolins V. Hydrogen storage properties of 2LiNH2+LiBH4+MgH2. J. Alloys Compd., 2007, 446: 345. [27] Yang J, Sudik A, Siegel D J, Halliday D, Drews A, Carter III R O, Wolverton C, Lewis G J, Sachtler J WA, Low J J, Faheem S A, Lesch D A, Ozolins V. A self-catalyzing hydrogen-storage material. Angew. Chem. Int. Ed., 2008, 47(5): 882. [28] Sudik A, Yang J, Halliday D, Wolverton C. Hydrogen storage properties in (LiNH2)2-LiBH4-(MgH2)x mixtures (x=0.0–1.0). J. Phys. Chem. C, 2008, 112(11): 4384. [29] Hu J J, Fichtner M, Chen P. Investigation on the properties of the mixture consisting of Mg(NH2)2, LiH, and LiBH4 as a hydrogen storage material. Chem. Mater., 2008, 20(22): 7089. [30] Srinivasan S S, Niemann M U, Hattrick-Simpers J R, McGrath K, Sharma P C, Goswami D Y, Stefanakos E K. Effects of nano additives on hydrogen storage behavior of
790
[31]
[32]
[33] [34]
[35] [36]
JOURNAL OF RARE EARTHS, Vol. 33, No. 7, July 2015 the multinary complex hydride LiBH4/LiNH2/MgH2. Int. J. Hydrogen Energy, 2010, 35(18): 9646. Zhang X G, Li Z N, Lv F, Li H L, Mi J, Wang S M, Liu X P, Jiang L J. Improved hydrogen storage performance of the LiNH2- MgH2-LiBH4 system by addition of ZrCo hydride. Int. J. Hydrogen Energy, 2010, 35(15): 7809. Hu J J, Pohl A, Wang S M, Rothe J, Fichtner M. Additive effects of LiBH4 and ZrCoH3 on the hydrogen sorption of the Li-Mg-N-H hydrogen storage system. J. Phys. Chem. C, 2012, 116(38): 20246. Zaluski L, Zaluska A, Ström-Olsen J O. Nanocrystalline metal hydrides. J. Alloys Compd., 1997, 253: 70. Luo W F, Rönnebro E. Towards a viable hydrogen storage system for transportation application. J. Alloys Compd., 2005, 404: 392. Kissinger H E. Reaction kinetics in differential thermal analysis. Anal. Chem., 1957, 29(11): 1702. Avrami M. Kinetics of phase change II transformationtime relations for random distribution of nuclei. J. Chem. Phys., 1940, 8(2): 212.
[37] Hancock J D, Sharp J H. Method of comparing solid-state kinetic data and its application to the decomposition of kaolinite, brucite, and BaCO3. J. Am. Ceram. Soc., 1972, 55(2): 74. [38] Chen P, Xiong Z T, Yang L F, Wu G T, Luo W F. Mechanistic investigations on the heterogeneous solid-state reaction of magnesium amides and lithium hydrides. J. Phys. Chem. B, 2006, 110(29): 14221. [39] Luo W F, Stewart K. Characterization of NH3 formation in desorption of Li-Mg-N-H storage system. J. Alloys Compd., 2007, 440(1): 357. [40] Hu J J, Liu Y F, Wu G T, Xiong Z T, Chua Y S, Chen P. Improvement of hydrogen storage properties of the Li-Mg-N- H system by addition of LiBH4. Chem. Mater., 2008, 20(13): 4398. [41] Meisner G P, Scullin M L, Balogh M P, Pinkerton F E, Meyer M S. Hydrogen release from mixtures of lithium borohydride and lithium amide: a phase diagram study. J. Phys. Chem. B, 2006, 110(9): 4186.
Improved dehydrogenation cycle performance of the 1.1MgH2-2LiNH2-0.1LiBH4 system by addition of LaNi4.5Mn0.5 alloy ZHAO Wang (赵 汪), JIANG Lijun (蒋利军)*, WU Yuanfang (武媛方), YE Jianhua (叶建华), YUAN Baolong (袁宝龙), LI Zhinian (李志念), LIU Xiaopeng (刘晓鹏), WANG Shumao (王树茂) (Energy Materials and Technology Research Institute, General Research Institute for Nonferrous Metals, Beijing 100088, China) Received 8 December 2014; revised 13 March 2015
Abstract: The isothermal desorption kinetics of the 1.1MgH2-2LiNH2-0.1LiBH4 system were improved by addition of LaNi4.5Mn0.5 alloy. The hydrogen desorption peak temperature of the sample containing LaNi4.5Mn0.5 reduced by approximately 5 K and the activation energy reduced by 9%. The results of isothermal dehydrogenation kinetics analysis implied that the isothermal desorption process at initial stage was controlled by the phase boundary mechanism. Moreover, the cycle performance of the materials was extended. The growth and agglomeration of the sample particles caused the deterioration of kinetics during de-/hydrogenation cycles, and then resulted in an incomplete desorption/absorption reaction which were responsible for the capacity fading. The cracking and pulverization of LaNi4.5Mn0.5 alloy had an obvious effect on preventing the composites aggregating, and the fine alloy particles could enhance the catalytic effect of the alloy, thus effectively offsetting part of the deterioration of kinetics caused by particles growth. Keywords: hydrogen storage materials; Li-Mg-B-N-H hydride; LaNi4.5Mn0.5; cycle performance; rare earths
Hydrogen is a green and efficient energy carrier, which is abundant and environmentally benign. One of key issues that limit hydrogen energy application is efficient storage of hydrogen[1]. Nowadays, many efforts have been made for development of new hydrogen storage materials worldwide[2–8]. Among the studied materials, the Li-Mg-N-H system comprised of LiH and Mg(NH2)2 or LiNH2 and MgH2, exhibits good reversibility at moderate operation temperatures and a relatively high capacity of 5.6 wt.%. It is therefore regarded as a promising candidate for on-board application[9]. This system can dehydrogenate under 0.1 MPa at 363 K by thermodynamic calculation[9]. However, a rather high kinetics barrier of hydrogen sorption has been identified, which restricts its application. A variety of dopants, KH, KOH, KF, lithium halides, RbF, NaOH, LaH3, VCl3, TiCl3, TaN, TiN, Li3AlH6, Ca(BH4)2, Mg(BH4)2, LiBH4, were investigated to improve the kinetic properties[10–26]. The alkali-metal compounds have been proved to be one of the most effective additives[10–18]. The hydrogen storage properties of Mg(NH2)2-2LiH were significantly enhanced by adding a small amount of KH[10–12]. The dehydrogenation peak temperature was lowered from 459 K for the Mg(NH2)2-2LiH sample to 405 K for the Mg(NH2)21.9LiH-0.1KH sample[10]. However, the positive effects of K-based additives disappear when the hydrogen release and uptake of the KF-added Mg(NH2)2- 2LiH samples are performed at higher temperatures (>473 K)[14,15].
It was found that metal borohydrides, such as Mg(BH4)2, Ca(BH4)2 and LiBH4, were another sort of effective dopants for improving the reaction kinetics of both dehydrogenation and hydrogenation of the Li-Mg-N-H system[23–26]. Yang et al. demonstrated that by starting with the mixture MgH2+LiNH2+LiBH4, the desorption reaction could proceed at lower temperatures and with significantly enhanced kinetics[26]. Via an in- depth exploration of the ternary mixture, a self-catalyzing reaction has been found, which improved the low temperature kinetics of the reversible reaction between Li2Mg(NH)2 and 2LiH+Mg(NH2)2[27]. The effect of the stoichiometry on the hydrogen storage property of this ternary complex hydride system has also been investigated[28,29]. To improve the hydrogen storage behavior of the Li-Mg-B-N-H system, various nanosized metals (Ni, Co, Fe, Cu, Mn) were added to this system. It was found that the additives of Co and Ni lowered the hydrogen releasing temperature at least 75–100 K in the major hydrogen decomposition step, while other additives acted as catalysts and increased the rate at which hydrogen was released[30]. With the addition of ZrCoH3, significant improvements in the hydrogen absorption/desorption properties of 2LiNH2-1.1MgH2-0.1LiBH4 composite have been achieved[31,32]. However, the researches on cycle performance of the Li-Mg-B-N-H system were rarely reported. As men-
Foundation item: Project supported by High-Tech Research and Development Program of China (2012AA051503) * Corresponding author: JIANG Lijun (E-mail: [email protected]; Tel.: +86-10-82241240) DOI: 10.1016/S1002-0721(14)60485-3
784
tioned above, the hydrogen storage properties of the LiMg-N-H system could be improved by adding LaH3[19]. Here, we studied the effect of the LaNi4.5Mn0.5 alloy on hydrogen sorption properties of the 2LiNH2-1.1MgH20.1LiBH4 composite. LaNi4.5Mn0.5 alloy is hard, brittle and easy to pulverize. The cycle performance and mechanism for hydrogen capacity fading was investigated.
1 Experimental The starting materials LiNH2 and LiBH4 (95% purity, Sigma-Aldrich) were used as received without purification. The MgH2 was home-made by ball-milling Mg powder under 4 MPa H2 pressure for 60 h, and then it was rehydrogenated for three times to obtain high-purity MgH2 (>95% purity). The LaNi4.5Mn0.5 intermetallic compound was re-melted three times by magnetic levitation melting under an argon atmosphere. The alloy was annealed in evacuated quartz tubes at 1273 K for 8 h, and then quenched in the water. The annealed alloy was hydrogenated for pulverising to –500 mesh powder under 5 MPa H2 pressure at ambient temperature. 3 g mixture of MgH2, LiNH2 and LiBH4 with molar ratio of 1.1:2:0.1 was put into a stainless steel vial for highenergy ball-milling under 4 MPa H2 pressure using a Spex-8000 apparatus. The weight ratio of stainless steel ball to powder was 20:1 and the total milling time was 36 h. The composite with an addition of 10 wt.% LaNi4.5Mn0.5 was prepared by high-energy ball milling LaNi4.5Mn0.5 hydride powder and the as-milled Li-Mg-B-N-H powder for 12 h. In order to reduce the amorphous phase produced by ball-milling, the stainless steel balls were replaced with zirconia balls and the weight ratio of ball to powder was 2:1. In case of the alloy with partially amorphous phase, the p-c isotherms differ significantly from the alloy without amorphous phase with an enhancement of solid state solubility and a lowering of the plateau pressure[33]. Isothermal dehydrogenation kinetics were measured using a Sieverts-type apparatus. Approximately 0.5 g of sample was used in each measurement. Differential scanning calorimetry (DSC) measurements were performed on a Netzsch DSC 409 unit. About 40 mg of samples were heated at a heating rate of 5 K/min under Ar atmosphere. The gas generated in the heating process was monitored by a mass spectrometer (MS). XRD analysis was carried out by using an X’pert Pro MPD diffractometer with Cu Kα radiation at 40 kV and 40 mA. The IR absorption spectrum was collected in diffuse reflectance infrared Fourier transform mode at a resolution of 4 cm–1, and the sample was mixed with paraffine in an appropriate proportion in order to isolate air. The morphology and element distributions of the composites
JOURNAL OF RARE EARTHS, Vol. 33, No. 7, July 2015
were measured by a field emission scanning electron microscope (SEM, Hitachi model S4800). All handling procedures in this work were performed in a glove-box filled with purified argon to keep the H2O and O2 levels below 1 ppm.
2 Results and discussion 2.1 Effects of LaNi4.5Mn0.5 on dehydrogenation kinetics and thermal effects Desorption isotherms were first measured for comparison of the effects induced by the addition of LaNi4.5Mn0.5. Fig. 1 presents hydrogen desorption kinetics curves under 0.1 MPa at 423 K of the Li-Mg-B- N-H system with and without LaNi4.5Mn0.5 addition. The dehydrogenation curves indicate that the desorption kinetics are enhanced by mixing 10 wt.% LaNi4.5Mn0.5 alloy. It has been found that the hydrogen release capacity is 3.12 wt.% and 3.62 wt.% within 120 min for the Li-MgB-N-H sample and the sample with LaNi4.5Mn0.5 addition respectively. However, the capacity loss for the sample added with LaNi4.5Mn0.5 is exhibited, due to the larger density and the smaller capacity of LaNi4.5Mn0.5, compared with the pristine sample of Li-Mg-B-N-H. Fig. 2(a) shows the DSC curves of the Li-Mg-B-N-H and 10 wt.% LaNi4.5Mn0.5 samples which were measured at a heating rate of 5 K/min from 303 to 623 K under argon atmosphere. The exothermic peaks at about 423 K coincide with the metathesis reaction that was observed by Luo et al.[34], which is very close to the results reported by Hu et al.[32]. The broad endothermic peaks in the temperature range of 463–483 K are obviously formed by the superposition of two peaks, which correspond to the two hydrogen desorption reactions reported by Yang et al.[27] in Eqs. (1) and (2) respectively. 2Li4BN3H10+3MgH2→3Li2Mg(NH)2+2LiBH4+6H2 (1) Mg(NH2)2+2LiH→Li2Mg(NH)2+2H2 (2) Li4BN3H10 should be formed by the reaction of LiNH2
Fig. 1 Hydrogen desorption kinetics curves under 0.1 MPa at 423 K of the Li-Mg-B-N-H sample and the 10 wt.% LaNi4.5Mn0.5 added sample
ZHAO Wang et al., Improved dehydrogenation cycle performance of the 1.1MgH2-2LiNH2-0.1LiBH4 system by …
785
alloy lowered desorption barrier, which is responsible for the reduction in the hydrogen desorption temperature and the enhancement in the hydrogen desorption kinetics. As we can see from the results of dehydrogenation kinetics and DSC-MS, the intermetallic compound LaNi4.5Mn0.5 plays a role of the catalyst to lower the kinetics barrier for dehydrogenation of the Li-Mg-B-N-H, reduces the dehydrogenation temperature and accelerates the rate of dehydrogenation effectively. 2.2 Isothermal dehydrogenation kinetics analysis
Fig. 2 DSC curves of the Li-Mg-B-N-H sample and the 10 wt.% LaNi4.5Mn0.5 added sample (a) and MS monitoring of H2 and NH3 generated during temperature programmed desorption (b) (The samples were heated at a heating rate of 5 K/min from 303 to 623 K)
and LiBH4 during sample preparation[27]. The peak temperature of the 10 wt.% LaNi4.5Mn0.5 sample is 466.7 K which is 4.8 K lower than that of the Li-Mg-B-N-H sample. It can be found that the positions of the H2 signal peaks correspond well with that of endothermic peaks. Compared to the pristine sample, the hydrogen desorption peak temperature of the sample with addition of LaNi4.5Mn0.5 reduces by approximately 5 K. The amount of NH3 release is extremely small compared to the amount of H2 release throughout the heating process. Differential scanning calorimetry (DSC) was also conducted to study the dehydrogenation activation energy of the Li-Mg-B-N-H and 10 wt.% LaNi4.5Mn0.5 added samples. The two Kissinger plots were drawn to estimate the activation energies (Fig. 3)[35]. The peak temperature data were obtained from the first peak. The activation energies by the linear fitting were found to be 81.0 kJ/mol for the Li-Mg-B-N-H sample and 73.8 kJ/mol for the 10 wt.% LaNi4.5Mn0.5 added sample, in other words, about 9 % reduction in the activation energy by the addition of LaNi4.5Mn0.5. This means that the addition of LaNi4.5Mn0.5
Fig. 3 Kissinger plots of the Li-Mg-B-N-H sample and the 10 wt.% LaNi4.5Mn0.5 sample
The isothermal dehydrogenation curves of the 10 wt.% LaNi4.5Mn0.5 added sample under 0.1 MPa H2 were further measured at 400–423 K, and the results are shown in Fig. 4. The dehydrogenation is dramatically decelerated as the temperature is decreased, and the hydrogen capacity is reduced. For understanding the details of dehydrogenation reaction kinetics, the isothermal data were analyzed in depth by using the Johnson-Mehl-Avrami (JMA) equation[36]. In practical application, the equation is generally rearranged as ln[–ln(1–α)]=n ln(t)+n ln(k) (3) In Eq. (3), α is the fraction reacted at time t, k is the rate constant, and n is the Avrami exponent. Fig. 5 shows the JMA plots for the isothermal dehydrogenation of the 10 wt.% LaNi4.5Mn0.5 added sample at various temperatures. There is a good linear relationship between ln[–ln(1–α)]
Fig. 4 Isothermal dehydrogenation curves under 0.1 MPa H2 of the 10 wt.% LaNi4.5Mn0.5 sample at various temperatures
Fig. 5 JMA plots for the isothermal dehydrogenation of the 10 wt.% LaNi4.5Mn0.5 sample at various temperatures
786
JOURNAL OF RARE EARTHS, Vol. 33, No. 7, July 2015
and ln(t) in the range of α=0–0.3, indicating that the dehydrogenation behaviors can be reasonably described by the JMA equation. Hancock et al. reported that this method applies not only to the common nucleation and growth equations (n=1, 2, 3), but also to non-integral values of n[37]. The certain values of Avrami exponent n represent the different nucleation and growth processes. For example, in phase-boundary reaction controlled mechanism, n is typically between 1.07–1.24. The values of the Avrami exponent (n) obtained by fitting the data are in the range of 0.9–1.1 for the dehydrogenation for 10 wt.% LaNi4.5Mn0.5 added sample, which is close to the value of n in interface controlled reactions. This reflects that the rate controlling step for the desorption process should be interface reaction. The reaction modeling equations (g(α)) related to the phase-boundary mechanism were applied to examine the mechanism underlying the dehydrogenation process. The expression of g(α) for phase-boundary mechanism is g(α)=1–(1–α)1/m=kt (4) The values of m are 1, 2 and 3, corresponding to the one-dimension, two-dimension and three-dimension phase-boundary mechanism, respectively. If the proposed mechanism was controlling the rates, then according to Eq. (4), a plot of g(α) versus time t should be linear. After fitting calculation, the model of one-dimension phase-boundary mechanism (m=1) versus time exhibits best linearity with linearity coefficient R2>0.99 in the range of α=0–0.4, as shown in Fig. 6. This implied that the isothermal desorption process at initial stage is controlled by the phase boundary mechanism. This result is in good agreement with the previous investigation reported by Chen et al.[38]. According to this mechanism, increasing the contact area of initial reactants will enhance the reaction rate. As a general rule, decreasing the particle size will increase the contact area. The addition of LaNi4.5Mn0.5 alloy that is hard could favor material pulverization during ball milling process and consequently quicken dehydrogenation kinetics.
2.3 Cycle performance of Li-Mg-B-N-H system with and without admixing of LaNi4.5Mn0.5 at 423 K Fig. 7 presents the capacities of 10 consecutive desorption cycles of the pristine sample and the 10 wt.% LaNi4.5Mn0.5 sample at 423 K. Compared to the desorption capacity of the first cycle, there is an evident decline in the second one for both the pristine sample and the 10 wt.% LaNi4.5Mn0.5 sample. The reason might be that the growth and/or agglomeration of the particle occurs after the first cycle which causes the degradation of the subsequent desorption kinetics and reduction in capacity within the stipulated time. In the next eight cycles, the capacity faded slowly, and the average decay rate is 0.063 wt.% per cycle for the Li-Mg-B-N-H sample, while the decay rate is 0.054 wt.% per cycle for the 10 wt.% LaNi4.5Mn0.5 sample. It has been proved that there is a slight improvement in the cycle performance by addition of LaNi4.5Mn0.5. More work is needed to understand the mechanism of the improvement. The deterioration of reaction kinetics after 10 cycles led to an incomplete des-/absorption reaction and consequently a capacity loss (this can be drawn from the hydrogen desorption kinetics curves which are not given in the paper). This phenomena is in good agreement with the findings reported by Hu et al.[32]. However, there may be some other reasons for capacity reduction, for example, NH3 formation[39], separation of the reactants by products and so on. This requires more in-depth studies to confirm them. In the first 10 cycles, the de-/hydrogenation time were 20 and 4 h, respectively. While shortening the time to 9 and 2 h in the next 10 cycles, it was found that the hydrogen capacity decayed faster (Fig. 8). The average decay rate of desorption capacities within 9 h increased from 0.051 wt.% to 0.064 wt.% per cycle. The short cyclic test time especially the hydrogenation time induces a more incomplete reaction, and brought about a larger decay rate.
Fig. 6 Time dependence of kinetic modeling equation g(α) (one-demension phase-boundary mechanism) for the dehydrogenation of the 10 wt.% LaNi4.5Mn0.5 added sample at various temperatures
Fig. 7 H2 desorption capacities of the Li-Mg-B-N-H sample and the 10 wt.% LaNi4.5Mn0.5 added sample (dehydrogenated under 0.1 MPa H2 pressure for 20 h and hydrogenated under 7.5 MPa H2 pressure for 4 h at 423 K)
ZHAO Wang et al., Improved dehydrogenation cycle performance of the 1.1MgH2-2LiNH2-0.1LiBH4 system by …
787
To shed light on the hydrogen capacity fading mechanism for the Li-Mg-B-N-H sample and the 10 wt.% LaNi4.5Mn0.5 added sample, the phase analysis, structure characterization and morphological observation were conducted. Fig. 9 presents the XRD patterns of the 10 wt.% LaNi4.5Mn0.5 added samples before and after dehydrogenation/hydrogenation cycles. For the ball-milled sample, the original materials, MgH2, LiNH2, LiBH4, cannot be identified for the amorphization behavior of the grains by ball-milling. The diffraction peak intensity of LaNi4.5Mn0.5 hydride significantly weakens. For the first desorption sample, only the Li2Mg(NH)2 and LaNi4.5Mn0.5 can be identified, and only Mg(NH2)2, LiH, LaNi4.5Mn0.5Hx for the first absorption sample, which manifest a good reversibility of dehydrogenation/hydrogenation of the sample added with 10 wt.% LaNi4.5Mn0.5. After 20 cycles, the reactant Mg(NH2)2 and LiH can be found in the dehydrogenated sample, and Li2Mg(NH)2 can be found in the hydrogenated sample as well. These findings provide powerful evidence for an incomplete desorption/absorption
reaction as mentioned above. No newly formed phase can be identified from these results in the sample with LaNi4.5Mn0.5 addition. It means that no new phase forms by the interaction between LaNi4.5Mn0.5 and Li-Mg-B-N-H composite. It is noted that the diffraction peak of LaNi4.5Mn0.5 shifts to the left slightly for the dehydrogenated sample (both the first and the twentieth) compared to the original LaNi4.5Mn0.5 alloy which suggests that there are residual H atoms in the alloy and the LaNi4.5Mn0.5 hydride does not dehydrogenate completely, because the hydrogen desorption pressure is 0.1 MPa instead of vacuum. Fig. 10 presents the FTIR spectra of the 10 wt.% LaNi4.5Mn0.5 added samples before and after dehydrogenation/hydrogenation cycles. For the ball-milled sample, the absorbance at 3325 and 3271 cm–1 suggests the existence of Mg(NH2)2 and the signature N–H vibration of starting material LiNH2 at 1563 cm–1 has also been found, which indicates an incomplete metathesis reaction[32]. In the hydrogenated samples (both the first and the twentieth) the absorbance at 1563 cm–1 disappears, and the signature N–H vibration of Mg(NH2)2 at 1570 cm–1 appears. In the dehydrogenated sample, the N–H vibration at 3168 cm–1 indicates the formation of the ternary imide Li2Mg(NH)2, and the absorbance at 3300 and 3241 cm–1 suggests the existence of the solid solution Li4BN3H10, which has been reported by Hu et al.[40]. However, the quaternary hydride Li4BN3H10 cannot be identified in the XRD patterns maybe owing to its low concentration. The existence of Li4BN3H10 with low melting point can be helpful to the growth and agglomeration of powder particles[41]. In the samples after twenty cycles, the absorbance at 3168 cm–1 corresponding to Li2Mg(NH)2 and 3325, 3271 cm–1 corresponding to Mg(NH2)2 become weaker compared to that in the samples before cycles. It can be inferred that the concentration of dehydrogenation products in desorption sample and hydrogenation products in absorption decreases, which signifies an incomplete desorption/absorption reaction within the limited time. Fig. 11 shows the morphologies of the 10 wt.%
Fig. 9 XRD patterns of the 10 wt.% LaNi4.5Mn0.5 added samples before and after dehydrogenation/hydrogenation cycles
Fig. 10 FTIR spectra of the 10 wt.% LaNi4.5Mn0.5 samples before and after dehydrogenation/hydrogenation cycles
Fig. 8 H2 desorption capacities within 9 h of the 10 wt.% LaNi4.5Mn0.5 added sample in 20 desorption cycles (dehydrogenated under 0.1 MPa H2 for 20 h and hydrogenated under 7.5 MPa H2 for 4 h in the first 10 cycles at 423 K, dehydrogenated for 9 h and hydrogenated for 2 h in the next 10 cycles)
2.4 Mechanism for hydrogen capacity fading
788
JOURNAL OF RARE EARTHS, Vol. 33, No. 7, July 2015
Fig. 11 SEM (top) and back scattered electron (bottom) images of the 10 wt.% LaNi4.5Mn0.5 added samples (a, c) As-milled; (b, d) After cycles
LaNi4.5Mn0.5 added samples before and after dehydrogenation/hydrogenation cycles. The particles are relatively uniform in size and some agglomerations can be found in the as-milled sample (Fig. 11(a), top). For the sample after cycles, the initially separated particles are remarkably aggregated and the particle size is augmented (Fig. 11(b), top). As mentioned above, for the heterogeneous solid-state of Li-Mg- N-H system, the reaction rate is controlled by the interface reaction in the early stage of the reaction. And in the later stage, it is controlled by the mass transport through the imide layer[38]. Neglecting degree of mixing of reacting species for the samples before and after cycles, the increscent particle size is responsible for the deterioration of reaction kinetics. From the back scattered electron images (Fig. 11 bottom), the particle size of LaNi4.5Mn0.5 alloy is about 1–2 μm, and the alloy particles were dehiscent and pulverized after 20 cycles. Many small alloy particles are released, which can increase the contact area between the alloy and base material and enhance the catalytic effect of the alloy, and then effectively offset part of the deterioration of kinetics caused by particles growth. For this reason, there is an improvement in the cycle performance by addition of LaNi4.5Mn0.5.
3 Conclusions The Li-Mg-B-N-H samples with and without addition
of LaNi4.5Mn0.5 were characterized by Sievert’s method for the hydrogen desorption, DSC-MS, XRD, FTIR, and SEM. The LaNi4.5Mn0.5 alloy could enhance the desorption kinetics and reduce the dehydrogenation peak temperature of Li-Mg-B-N-H system. The intermetallic compound LaNi4.5Mn0.5 played a role of the catalyst to lower the kinetics barrier for dehydrogenation of the Li-MgB-N-H. The results of isothermal dehydrogenation kinetics analysis implied that the isothermal desorption process at initial stage was controlled by the phase boundary mechanism. The increscent size and the agglomeration of the sample particles caused the deterioration of kinetics, and then the incomplete desorption/absorption reaction within limited time was a crucial reason for capacity fading. The LaNi4.5Mn0.5 alloy particles were dehiscent and pulverized after cycles as we expected. The pulverization increased the dispersion degree of the alloy, which could enhance the catalytic effect of the alloy, and then partly offset the deterioration of kinetics caused by particles growth and agglomeration. For this reason, there was an improvement in the cycle performance by addition of LaNi4.5Mn0.5 alloy.
References: [1] Schlapbach L, Züttel A. Hydrogen-storage materials for mobile applications. Nature, 2001, 414(6861): 353.
ZHAO Wang et al., Improved dehydrogenation cycle performance of the 1.1MgH2-2LiNH2-0.1LiBH4 system by … [2] Oelerich W, Klassen T, Bormann R. Metal oxides as catalysts for improved hydrogen sorption in nanocrystalline Mg-based materials. J. Alloys Compd., 2001, 315(1): 237. [3] Liu Y F, Liang C, Zhou H, Gao M X, Pan H G, Wang Q D. A novel catalyst precursor K2TiF6 with remarkable synergetic effects of K, Ti and F together on reversible hydrogen storage of NaAlH4. Chem. Commun., 2011, 47(6): 1740. [4] Guo Y H, Yu X B, Gao L, Xia G L, Guo Z P, Liu H K. Significantly improved dehydrogenation of LiBH4 destabilized by TiF3. Energy & Environmental Science, 2010, 3(4): 464. [5] Xia G L, Leng H Y, Xu N X, Li Z L, Wu Z, Du J L, Yu X B. Enhanced hydrogen storage properties of LiBH4-MgH2 composite by the catalytic effect of MoCl3. Int. J. Hydrogen Energy, 2011, 36(12): 7128. [6] Liu J, Li Q. Hydriding/dehydriding behaviors of La2Mg1710 wt.% Ni composite prepared by mechanical milling. J. Rare Earths, 2013, 31(1): 73. [7] Chen P, Xiong Z, Luo J, Lin J, Tan K L. Interaction of hydrogen with metal nitrides and imides. Nature, 2002, 420(6913): 302. [8] Xiong Z T, Wu G T, Hu JJ, Chen P. Ternary imides for hydrogen storage. Adv. Mater., 2004, 16(17): 1522. [9] Xiong Z T, Hu JJ, Wu G T, Chen P, Luo W F, Cross K, Wang J. Thermodynamic and kinetic investigations of the hydrogen storage in the Li-Mg-N-H system. J. Alloys Compd., 2005, 398(1): 235. [10] Wang J H, Liu T, Wu G T, Li W, Liu Y F, Araujo C M, Scheicher R H, Blomqvist A, Ahuja R, Xiong Z T, Yang P, Gao M X, Pan H G, Chen P. Potassium-modified Mg(NH2)2/2LiH system for hydrogen storage. Angew. Chem. Int. Ed., 2009, 48(32): 5828. [11] Li C, Liu Y F, Pang Y P, Gu Y J, Gao M X, Pan H G. Compositional effects on the hydrogen storage properties of Mg(NH2)2- 2LiH-xKH and the activity of KH during dehydrogenation reactions. Dalton Transactions, 2014, 43(6): 2369. [12] Li C, Liu Y F, Ma R J, Zhang X, Li Y, Gao M X, Pan H G. Superior dehydrogenation/hydrogenation kinetics and long- term cycling performance of K and Rb cocatalyzed Mg(NH2)2‐2LiH system. Appl. Mater. Interfaces, 2014, 6(19): 17024. [13] Liang C, Liu Y F, Gao M X, Pan H G. Understanding the role of K in the significantly improved hydrogen storage properties of a KOH-doped Li-Mg-N-H system. J. Mater. Chem. A, 2013, 1(16): 5031. [14] Liu Y F, Li C, Li B, Gao M X, Pan H G. Metathesis reaction-induced significant improvement in hydrogen storage properties of the KF-added Mg(NH2)2−2LiH system. J. Phys. Chem. C, 2013, 117(2): 866. [15] Li C, Liu Y F, Yang Y J, Gao M X, Pan H G. High-temperature failure behaviour and mechanism of K-based additives in Li-Mg-N-H hydrogen storage systems. J. Mater. Chem. A, 2014, 2(20): 7345. [16] Li B, Liu Y F, Li C, Gao M X, Pan H G. In situ formation of lithium fast-ion conductors and improved hydrogen desorption properties of the LiNH2-MgH2 system with the addition of lithium halides. J. Mater. Chem. A, 2014, 2(9):
789
3155. [17] Li C, Liu Y F, Gu Y J, Gao M X, Pan H G. Improved hydrogen-storage thermodynamics and kinetics for an RbF-doped Mg(NH2)2-2LiH system. Chemistry–An Asian Journal, 2013, 8(9): 2136. [18] Liang C, Liu Y F, Wei Z J, Jiang Y, Wu F, Gao M X, Pan H G. Enhanced dehydrogenation/hydrogenation kinetics of the Mg(NH2)2-2LiH system with NaOH additive. Int. J. Hydrogen Energy, 2011, 36(3): 2137. [19] Zhu X L, Han S M, Zhao X, Li Y, Liu B Z. Effect of lanthanum hydride on microstructures and hydrogen storage performances of 2LiNH2-MgH2 system. J. Rare Earths, 2014, 32(5): 429. [20] Shahi R R, Yadav T P, Shaz M A, Srivastva O N. Studies on dehydrogenation characteristic of Mg(NH2)2/LiH mixture admixed with vanadium and vanadium based catalysts (V, V2O5 and VCl3). Int. J. Hydrogen Energy, 2010, 35(1): 238. [21] Ma L P, Wang P, Dai H B, Cheng H M. Catalytically enhanced dehydrogenation of Li-Mg-N-H hydrogen storage material by transition metal nitrides. J. Alloys Compd., 2009, 468(1): L21. [22] Cao H J, Zhang Y, Wang J H, Xiong Z T, Wu G T, Qiu J S, Chen P. Effects of Al-based additives on the hydrogen storage performance of the Mg(NH2)2-2LiH system. Dalton Transactions, 2013, 42(15): 5524. [23] Li B, Liu Y F, Gu J, Gu Y J, Gao M X, Pan H G. Mechanistic investigations on significantly improved hydrogen storage performance of the Ca(BH4)2-added 2LiNH2/MgH2 system. Int. J. Hydrogen Energy, 2013, 38(12): 5030. [24] Li B, Liu Y F, Gu J, Gao M X, Pan H G. Synergetic effects of in situ formed CaH2 and LiBH4 on hydrogen storage properties of the Li-Mg-N-H system. Chemistry–An Asian Journal, 2013, 8(2): 374. [25] Pan H G, Shi S B, Liu Y F, Li B, Yang Y J, Gao M X. Improved hydrogen storage kinetics of the Li-Mg-N-H system by addition of Mg(BH4)2. Dalton Transactions, 2013, 42(11): 3802. [26] Yang J, Sudik A, Siegel D J, Halliday D, Drews A, Carter III R O, Wolverton C, Lewis G J, Sachtler J WA, Low J J, Faheem S A, Lesch D A, Ozolins V. Hydrogen storage properties of 2LiNH2+LiBH4+MgH2. J. Alloys Compd., 2007, 446: 345. [27] Yang J, Sudik A, Siegel D J, Halliday D, Drews A, Carter III R O, Wolverton C, Lewis G J, Sachtler J WA, Low J J, Faheem S A, Lesch D A, Ozolins V. A self-catalyzing hydrogen-storage material. Angew. Chem. Int. Ed., 2008, 47(5): 882. [28] Sudik A, Yang J, Halliday D, Wolverton C. Hydrogen storage properties in (LiNH2)2-LiBH4-(MgH2)x mixtures (x=0.0–1.0). J. Phys. Chem. C, 2008, 112(11): 4384. [29] Hu J J, Fichtner M, Chen P. Investigation on the properties of the mixture consisting of Mg(NH2)2, LiH, and LiBH4 as a hydrogen storage material. Chem. Mater., 2008, 20(22): 7089. [30] Srinivasan S S, Niemann M U, Hattrick-Simpers J R, McGrath K, Sharma P C, Goswami D Y, Stefanakos E K. Effects of nano additives on hydrogen storage behavior of
790
[31]
[32]
[33] [34]
[35] [36]
JOURNAL OF RARE EARTHS, Vol. 33, No. 7, July 2015 the multinary complex hydride LiBH4/LiNH2/MgH2. Int. J. Hydrogen Energy, 2010, 35(18): 9646. Zhang X G, Li Z N, Lv F, Li H L, Mi J, Wang S M, Liu X P, Jiang L J. Improved hydrogen storage performance of the LiNH2- MgH2-LiBH4 system by addition of ZrCo hydride. Int. J. Hydrogen Energy, 2010, 35(15): 7809. Hu J J, Pohl A, Wang S M, Rothe J, Fichtner M. Additive effects of LiBH4 and ZrCoH3 on the hydrogen sorption of the Li-Mg-N-H hydrogen storage system. J. Phys. Chem. C, 2012, 116(38): 20246. Zaluski L, Zaluska A, Ström-Olsen J O. Nanocrystalline metal hydrides. J. Alloys Compd., 1997, 253: 70. Luo W F, Rönnebro E. Towards a viable hydrogen storage system for transportation application. J. Alloys Compd., 2005, 404: 392. Kissinger H E. Reaction kinetics in differential thermal analysis. Anal. Chem., 1957, 29(11): 1702. Avrami M. Kinetics of phase change II transformationtime relations for random distribution of nuclei. J. Chem. Phys., 1940, 8(2): 212.
[37] Hancock J D, Sharp J H. Method of comparing solid-state kinetic data and its application to the decomposition of kaolinite, brucite, and BaCO3. J. Am. Ceram. Soc., 1972, 55(2): 74. [38] Chen P, Xiong Z T, Yang L F, Wu G T, Luo W F. Mechanistic investigations on the heterogeneous solid-state reaction of magnesium amides and lithium hydrides. J. Phys. Chem. B, 2006, 110(29): 14221. [39] Luo W F, Stewart K. Characterization of NH3 formation in desorption of Li-Mg-N-H storage system. J. Alloys Compd., 2007, 440(1): 357. [40] Hu J J, Liu Y F, Wu G T, Xiong Z T, Chua Y S, Chen P. Improvement of hydrogen storage properties of the Li-Mg-N- H system by addition of LiBH4. Chem. Mater., 2008, 20(13): 4398. [41] Meisner G P, Scullin M L, Balogh M P, Pinkerton F E, Meyer M S. Hydrogen release from mixtures of lithium borohydride and lithium amide: a phase diagram study. J. Phys. Chem. B, 2006, 110(9): 4186.