- Email: [email protected]
Improved hydrogen sorption kinetics in Mg modified by chosen catalysts
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 4 4 ( 2 0 1 9 ) 8 3 1 5 e8 3 2 4
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/he
Improved hydrogen sorption kinetics in Mg modified by chosen catalysts Jiri Cermak a,b,*,1, Lubomir Kral a, Pavla Roupcova a,c a
Institute of Physics of Materials AS CR, v.v.i., Zizkova 22, CZ-61662 Brno, EU, Czech Republic CEITEC-Institute of Physics of Materials, AS CR, Zizkova 22, CZ-61662 Brno, EU, Czech Republic c CEITEC-Brno University of Technology, Purkynova 123, CZ-61662 Brno, EU, Czech Republic b
article info
abstract
Article history:
Hydrogen sorption characteristics were investigated in Mg modified by ten chosen cata-
Received 4 October 2018
lyzing additives X ¼ Mg2Si, Mg2Ge, Mg17Al12, Mg5Ga2, NaCl, LiCl, NaF, LiF and by two
Received in revised form
combinations Ni þ Mg17Al12 and Ni þ Mg2Si. Amorphous carbon (CB) was used as an anti-
1 February 2019
sticking component. It was found that optimum content of both X and CB was between 10
Accepted 12 February 2019
and 14 wt %. To assess the effect of X itself upon the sorption kinetics, concentration of
Available online 7 March 2019
both X and CB was fixed to about 12 wt %. The shortest desorption times t09 (time to
Keywords:
NaCl. Careful measurements of equilibrium hydrogen pressure, peq, at 623 K showed an
desorption of 90 pct of stored hydrogen) showed alloys with X ¼ NaF, Mg2Si, Mg2Ge and Hydrogen storage
increase in peq for added chlorides, fluorides, Mg2Ge and also for Mg2Si. TPD experiments
Mg alloys
revealed desorption peaks for X ¼ NaCl, LiF shifted down to temperatures as low as 597 K.
Carbon black
© 2019 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
Catalysis
Introduction Mg-based alloys are promising materials for hydrogen storage (HS) applications [1e6]. The principal HS phase in alloys of this type is MgH2 hydride. This phase shows high HS capacity (7.6 wt % H2) and excellent cycling behavior. The substantial drawback is its high thermodynamic stability, which implies high operation temperatures. Much research work was devoted to decrease the high stability of MgH2 (see, e.g., in Refs. [7e19]), i.e. to increase hydrogen pressure in equilibrium, peq, above the value for pure Mg hydride. This would decrease working temperatures of HS reservoirs down to some 373 K, which is a desired temperature region for HS applications (see US Department of energy target parameters [4]). However, this
proved to be an extraordinary difficult task. After a considerable research effort, only moderate success was reached in this respect [6]. Promising new techniques of HS materials preparation such as dual tuning of plasma milled samples [18,19] did not lead to any significant increase in equilibrium hydrogen pressure. Much expressive progress in the last years was reported in the field of improvement of the hydrogen sorption kinetics in Mg-based HS materials [4e6]. Of course, the dynamics and kinetics are closely interconnected, but it seems that influencing the obstacles that inhibit the sorption kinetics is much easier task then to affect effectively the generalized driving forces of the sorption process. Main ways how to improve the kinetics are the catalysis of partial sorption steps [20e26] and nanosizing [6,27,28].
* Corresponding author. Institute of Physics of Materials AS CR, v.v.i., Zizkova 22, CZ-61662 Brno, EU, Czech Republic. E-mail address: [email protected] (J. Cermak). 1 Permanent address: Institute of Physics of Materials v. v. i., Academy of Sciences of the Czech Republic, Zizkova 22, CZ-61662 Brno, Czech Republic. https://doi.org/10.1016/j.ijhydene.2019.02.084 0360-3199/© 2019 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
8316
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 4 4 ( 2 0 1 9 ) 8 3 1 5 e8 3 2 4
In the present work, HS characteristics of ball-milled (BM) blend of Mg with chosen additives were investigated. Five types of additives were tested: (i) Mg-based intermetallics formed by two elements from 13th group (Mg2Ge, Mg2Si), and by (ii) two elements of 14th group (Mg5Ga2, Mg17Al12). (iii) Two combinations Ni þ Mg17Al12 and Ni þ Mg2Si, (iv) two simple fluorides with Naþ and Liþ (NaF, LiF) and (v) two chlorides with the same cations (NaCl, LiCl) were also tested. Expectation of beneficial effect of the intermetallics with elements from 13th and 14th group upon the absorption/desorption kinetics (A/D) kinetics was based on promising results published in Refs. [18,29e32], synergy of those elements and Ni could be expected with regard to results from Refs. [33,34] and improvement of A/D kinetic by anions F, Cl and/or cations Naþ, Liþ might be expected from results published in Refs. [30,31,35e40]. Besides the above mentioned additives, also carbon black (CB) was introduced into the BM blends as antisticking agens [6,7] and also, as components that may take part in hydrogen sorption [41,42].
Experimental Samples were made from pure components purchased by Sigma-Aldrich. Mg was in the form of pieces (purity: 3N8), other metallic components e Ge(5N), Si(6N), Ga(4N), Al(5N), Ni(3N), halides e LiCl, NaF, NaCl, LiF (purity of all halides >2N) and CB (2N) were purchased in form of fine powders. Intermetallics Mg2Si and Mg2Ge were prepared by mechanical alloying in form of powders. Intermetalic compounds Mg17Al12 and Mg5Ga2 were induction melted in pure Ar(4N) and grinded into form of splinters. All the additives were mixed with splinters of Mg and ball milled in hydrogen atmosphere (6N) using Fritsch Pulverisette6 ball-mill. Two HS samples were prepared with admixture of fine powder of Ni. The mass ratio of the milling balls to the milled blend was about 60 and the milling cycle e 10 min milling/50 min cooling e was repeated 90 times. Hydrogen absorption under pressure p ¼ 2.5 MPa, desorption into a fixed volume with hydrogen pressure always well below the peq(T), and temperature programmed desorption (TPD) was carried out using Sievertstype gas sorption analyzer PCTePro Setaram Instrumentation at temperatures between 575 K and 658 K. All manipulations of the milled blend inclusive the filling the cuvette were done in the glove box in protective Ar atmosphere. Cuvette with the sample was transported in a small container filled by Ar and inserted into the sorption apparatus. Nominal chemical composition of ball milled batches is listed in Table 1. Phase composition of samples was obtained by XRD EMPYREAN device using CoKa radiation and the results were interpreted (Rietveld analysis) with the HighScore Plus Software and ICSD databases (Inorganic Crystal Structure Database Fachinformationszentrum Karlsruhe, Germany, and the U.S. Department of Commerce of the United States, version 1.9.8.). Accuracy of phase composition was about 2 wt %. Morphology of milled samples, average chemical composition and chemical maps were observed by SEM TESCAN LYRA3 equipped with X-max80 EDS in the area approximately 300 500 mm containing about 102 grains. Accuracy of average concentration of substitution elements (except for Li) was
within 1.5 wt%. Content of amorphous phase, cAM, (labeled below in the text as AM) and concentration of carbon, cC were fixed to value cCB, known from initial composition of the ball milling charge (see Table 1). Lithium concentration, cLi, was calculated from stoichiometric ratio. In LiF-containing samples, the value cLi ¼ 0.36537 cF and, for cLi in LiCl-containing samples the value cLi ¼ 0.1957 cCl was taken, where concentrations cF and cCl were obtained by SEM.
Results and discussion Characterization of samples Morphology of all samples after the BM, its change after hydrogen absorption (A) and hydrogen desorption steps (D) follows very similar evolution scheme. It can be illustrated by the sample Ge e see supplementary material, Fig. S1. The structure of Ge after BM is showed in Fig. S1a, where relatively smooth grains of size typically between 5 and 50 mm can be seen. These grains, composed principally of solid solution (Mg), transform during A to hydride phase and their appearance becomes corny e Fig. S1b. After D, when the hydrided phase(s) transformed back to hydride-free components, the corny structure persists and, moreover, additional fragmentation and transversal cracks of great grains appear e Fig. S1c. Fine particles of size about 1 mm (for Ge e bright submicron grains, see in Fig. S1a) were identified as impurities with traces of Ti, Al, W and Fe. Concentration of these elements was lower than 1 wt % and therefore, it was neglected in SEM analyses. XRD patterns were measured for each sample in BM, A and D states. The obtained data could be classified qualitatively into two considerably different groups. The first typical behavior can be described by the scheme: cAM in BM y cAM in C y cAM in D with cAM ¼ cCB (cAM stands for concentration of amorphous phase from XRD analysis, cCB is known concentration of carbon black). This type of behavior was observed in all alloys except for chloride-containing samples. The samples modified by chlorides (LiCl and NaCl) on the other hand, formed the other group where the content of AM measured by XRD was very high, especially in BM and D. The typical XRD patterns are exemplified in supplementary material e see Fig. S2 for Ge and NaCl. No carbides were detected in any sample; the components, which content was below 1% were neglected.
Structure after BM and A/D cycle Results of XRD and SEM observations of phase and chemical composition in states BM, A and D can be shortly presented as follows:
Sample Si After BM (indicated phases: 79(Mg)-4Mg2Si-2Si-2aMgH2-13AM; elements found: 84 Mg-3Si-13C) all in wt. %. Si was present both as regularly dispersed fine particles of Mg2Si and as scattered elemental Si in the form of particles of size up to 1 mm. Separate XRD analysis of ball milled blend Mg þ Si showed that the BM produced Mg2Si compound only up to about 21%. This explained observed differences between
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 4 4 ( 2 0 1 9 ) 8 3 1 5 e8 3 2 4
8317
Table 1 e Experimental samples. CB e carbon black. Sample label Si Ge Al Ga1 Ga2 Ga3 NiSi NiAl LiF Na NaF1 NaF2 NaF3 NaF4 LiCl NaCl a
Nominal phase composition of batch (wt. %) 76.36 75.28 76.58 81.61 75.81 68.19 71.87 72.37 74.04 90.91 85.60 74.81 60.00 70.67 75.94 74.61
Mge10.25 Mg2Si e 13.39 CB Mge12.61 Mg2Ge e 12.11 CB Mge10.40 Mg17Al12 e 13.02 CB Mge10.91 Mg5Ga2 e 7.48 CB Mge10.26 Mg5Ga2 e 13.93 CB Mge9.09 Mg5Ga2 e 22.72 CB Mge3.26 Ni e 12.47 Mg2Si e 12.40 CB Mge2.53 Ni e 12.84 Mg17Al12 e 12.31 CB Mge13.18 LiF e 12.78 CB Mge9.09 NaH Mge7.14 NaF e 7.26 CB Mge12.41 NaF e 12.78 CB Mge19.99 NaF e 20.01 CB Mge22.13 NaF e 7.20 CB Mge11.57 LiCl e 12.49 CB Mge12.99 NaCl 12.40 CB
Chemical composition (wt. %) 81.79 Mge4.82 Si e 13.39C 80.32 Mge7.57 Ge e 12.11C 82.48 Mge4.5 Al e 13.02C 86.72 Mge5.80 Ga 7.48C 80.61 Mge5.46 Ga e 13.93C 72.44 Mge4.84 Ga e 22.72C 79,76 Mge3,26 Ni e 4,58 Si e 12,40 C 79,82 Mge2,53 Ni e 5,39 Al e 12,31 C 74.04 Mge3.35 Li e 9.65 F e 12.78C 91.26e8.74 Naa 85.60 Mge3.91 Na e 3.23 Fe7.26C 74.81 Mge6.81 Na e 5.60 Fe12.78C 60.00 Mge10.95 Na e 9.04 Fe20.01C 70.67 Mge12.13 Na e 10.00 Fe7.20C 75,94 Mge1,89 Li e 9,68 Cl e 12,49 C 74.60 Mge5.20 Na 7.80 CL e 12.40C
In the initial state before sorption experiments (after decay of NaH).
nominal and measured phase composition in BM of sample Si. Presence of aMgH2 was a consequence of BM in H2. Chemical analysis agreed reasonably (within experimental errors) with the nominal values (Table 1). Carbon was distributed regularly. After A (78aMgH2-9Mg2Si-13AM; 78 Mg-3Si-13C-6H), lowpressure alpha hydride phase appeared, all elemental Si reacted with Mg and formed particles Mg2Si (size up to 2 mm). In D state (76(Mg)-11Mg2Si-13AM; 84 Mg-3Si-13C), particles Mg2Si (size up to 2 mm) persisted.
Sample Ge After BM (75(Mg)-12Mg2Ge-13AM; 81 Mg-7Ge-12C), Mg2Ge particles (~5 mm) were observed, C was uniformly distributed, chemical concentration agreed with nominal composition. After A (70aMgH2-11Mg2Ge-7(Mg)-12AM; 78 Mg-5Ge-11C-6H), Mg2Ge particles became less distinctly bounded. After D (75(Mg)-13Mg2Ge-12AM; 83 Mg-5Ge-12C), the original state was almost exactly restored.
Sample Al After BM (83(Mg)-3Mg1.95Al0.05-1MgH2-13AM; 82 Mg-5Al-13C), no Mg17Al12 phase was detected. It seemed that it was decayed during BM and the products entered (Mg) phase. Al was distributed approximately uniformly; slightly higher Al concentration could be detected in Mg1.95Al0.05 particles (size between 1 and 10 mm). After A (74aMgH2-13(Mg)-13AM; 79 Mg3Al-12C-6H), low-pressure hydride aMgH2 appeared, presence of (Mg) solid solution was a consequence of incomplete absorption. After D (86(Mg)-1MgH2-13AM; 82 Mg-5Al-13C), the state observed after BM was restored.
Sample Ga2 After BM (86(Mg)-14AM; 86 Mg-7Ga-7C), all components were uniformly scattered. Mg5Ga2 phase decomposed during BM and the decay products entered the AM and partly also the (Mg). After A (86bMgH2-4Mg5Ga2-3Mg2Ga-7AM; 85 Mg-2Ga-7C6H), one of high pressure phases was formed (b). The phase b was observed also in ball milled alloys Mg-M (M ¼ Ti, Nb, V) [9,17,43]. Loci (~1m) with significantly higher concentration of Ga were observed, which were identified as two MgmGan
phases. They precipitated from AM. After D (85((Mg)-8Mg5Ga27AM; 88 Mg-5Ga-7C), Mg2Ga disappeared whereas Mg5Ga2 particles persisted.
Sample NiSi After BM (77(Mg)-4Mg2Si-1Ni-2Si-5aMgH2-12AM; 80 Mg-4Ni4Si-12C), Si particles (~2 mm) that anti-coincided with Mg were observed. Mg2Si was not completely formed (similarly as in sample Si above). After A (76aMgH2-1(Mg)-10Mg2Si-1Ni-12AM, 77 Mg-2Ni-3Si-12C-6H), Formation of Mg2Si phase was completed (particles ~2e3 mm). After D (75(Mg)-12Mg2Si-13AM; 81 Mg-3Ni-4Si-12C), particles of Mg2Si were restored.
Sample NiAl After BM (85(Mg)-1NiAl-1Ni-1aMgH2-12AM; 80 Mg-3Ni-5Al12C), the Mg17Al12 decayed and the residuals entered (Mg). After A (82aMgH2-1(Mg)-5Ni2Al3-12AM; 77 Mg-2Ni-3Al-12C6H), Mg was mostly bound in MgH2, Sporadic Mg particles were also observed. Al was scattered uniformly; no localized phase particles composed of Ni, Al and Mg were observed. After D (83(Mg)-4Ni2Al3-1NiAl-12AM; 80 Mg-3Ni-5Al-12C), new-formed phases NimAln replaced completely originally introduced phase Mg17Al12 in hydrogen-free sample.
Sample LiF After BM (75(Mg)-12LiF-13AM; 75 Mg-3Li-9F-13C), Small particles LiF (~3 mm) were observed on the surface of great Mg grains. C was scattered uniformly. After A (83bMgH2-2LiF2(Mg)-13AM; 68 Mg-4Li-10F-12C-6H), the identity of LiF phase persisted even after the hydrogen absorption. High pressure phase bMgH2 was formed by BM; (Mg) phase was a consequence of incomplete absorption. After D (75(Mg)-11LiF1MgH2-13AM; 73 Mg-4Li-10F-13C), the original state observed after BM was almost exactly restored.
Sample NaF2 After BM (80(Mg)-7NaMgF1.17H1.83e13AM; 75 Mg-6Na-6F-13C), NaF decomposed during BM and the decay products entered the (Mg) phase. A small fraction of ternary hydride was formed. All chemical components were scattered uniformly.
8318
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 4 4 ( 2 0 1 9 ) 8 3 1 5 e8 3 2 4
After A (84bMgH2-1NaMgH3-2(Mg)-13AM; 74 Mg-5Na-3F-12C6H), high pressure hydride bMgH2 was formed, ternary hydride changed its composition and F remained solved probably in AM and/or in 9Mg0 phases. After D (80(Mg)7NaMgF1.17H1.83e13AM; 79 Mg-4Na-4F-13C), the original state was approached, observed after BM.
Sample LiCl After BM (73(Mg)-11LiCl-4bMgH2-12AM; 76 Mg-2Li-10Cl-12C), both phase and chemical composition agreed approximately with nominal values; high pressure bMgH2 was formed. After A (72aMgH2-15Li-1(Mg)-12AM; 67 Mg-2Li-12Cl-12C-6H), LiCl was found completely decayed. Li precipitated in elemental form. After D (2(Mg)-98AM; 73 Mg-2Li-12Cl-12C), AM phase grew significantly. It was most likely due to interaction of Cl with Mg [43] that produced finally extremely fine (almost amorphous) phase. Positions of chlorides lines [43] are plotted in Fig. S2c. In case of fine structure they may broaden and superimpose over LiCl pattern. In results, they may be involved in overall AM phase.
Sample NaCl After BM (77(Mg)-8NaCl-3aMgH2-12AM; 75 Mg-5Na-8Cl-12C), NaCl phase was found partly decomposed; the decay products entered (Mg) and/or AM phase. After A (74aMgH2-9NaCl-5Mg12AM; 70 Mg-5Na-7Cl-12C-6H), NaCl phase remained also after hydrogen absorption. After D (10(Mg)-90AM; 74 Mg-5Na9Cl-12C), NaCl phase decayed and the transformation products entered the phase that behaves as AM. AM contained also prevailing fraction of Mg. Most likely, this was a result of fragmentation of MgH2 into very fine particles during interaction of Mg with Cl e reasoning is similar as in the case of sample LiCl.
Observed structure transformations e summary Chemical composition of all samples in states BM and D were reasonably equal to elements nominal concentrations in milling batches. On the basis of the present observation, the response of alloys structure during A/D cycle can be ranked into three distinct classes: Samples Ge, LiF and NaF form the first one. The phase composition of these samples in D is about the same as in A. Samples Si and NiSi can be also involved in this class, since the formation of Mg2Si compound is completed after the first A and then remains without changes. On the other hand, samples Al, NiAl and Ga in the second class went through a significant phase reformation leading to different phase composition in BM and D. Samples NaCl and LiCl in the last class showed also significant phase changes, but they tended to form amorphous phase already during BM. Great fraction of AM phase was found also in D.
Sorption kinetics Reproducibility A and D kinetics were studied in a repeated A/D cycling regime. The concentration-time curves, cH(t), were recorded after the stabilization of this process. Usually one or two A/D cycles were sufficient to reach the stable state, where the hydrogen storage capacity, cHmax, varied within about ±0.2% and reproducibility of time to 0.9 of cHmax, t09, was better
than 10 s for A and about 0.75 min for D e see in supplementary material - Fig. S3.
Optimum composition Measurements of cH(t) for varied values of cX and cCB (cX is concentration of additive X, concentration of the carbon black phase, cCB, is equal to concentration of carbon, cC) revealed that if the hydrogen sorption capacity, cHmax should be kept at about 6 wt % H2, and time t09 as short as possible, cX and cCB must be about 10e14 wt %. This is illustrated in Figs. S4 and S5 for alloys NaFx and Gax. It was observed, most likely due to alloying effects, that the sorption capacity might even exceed slightly the capacity of pure Mg (e.g., for sample Ga1 e Fig. S5). This led, however, to lowered sorption rate, which is general conclusion found for all studied alloys: the effect of X and CB upon the cHmax and t09 by desorption was significantly inverse. Increase in cHmax led to worse desorption kinetics.
Influence of X upon sorption kinetics Comparison of relative influence of additive X upon the sorption kinetics was carried out at temperature 623 K and with approximately equal content of CB and X. Measured sorption curves cH(t) are plotted in Fig. 1 for times up to 0.1 h compared with results for pure Mg. Hydrogen concentration cH(t) approached values about 6 wt % H2 in all cases after sufficiently long absorption time t. Very good sorption kinetics was found for sodium halides X ¼ NaF, NaCl and for alloys that were modified by Daltontype intermetallics X ¼ Mg2Si, Mg2Ge that proved to be stable during A/D cycle e see paragraph 3.2. Alloys Ga and alloy Na modified by pure sodium without halide element (introduced in the form of NaH hydride) showed relative slowest desorption kinetics. It can be seen that all additives shown in Fig. 1 improved the hydrogen sorption kinetics compared to hydrogen sorption in pure Mg. Worse desorption kinetics was observed with alloys modified by combination Ni þ Mg17Al12 and Ni þ Mg2Si e see Fig. 2. This was due most likely to unwanted interaction of Ni with other alloys components. Parasite phases of the type NimAln in alloy NiAl were even identified by XRD (see in paragraph Structure after BM and A/D cycle). Unwanted interaction between Mg, Ni and Si that must not be manifested directly in an explicit formation of unwished phases can be also expected [44].
Measurements of pressure-composition isotherms (PCT) Measurement of PCT curves was carried out with alloys Si, Al, Ga2, NaCl and LiF at temperatures between 573 K and 658 K. Other alloys, and also Mg e as a reference material e were studied at a single temperature 623 K. Typical results are illustrated in Fig. 3. Van't Hoff diagram is shown in Fig. 4, where the present results are compared with literature data for Mg. Our value of peq ¼ 0.54 ± 0.03 MPa obtained for Mg at temperature T ¼ 623 K agree quite well with other literature data for Mg plotted in Fig. 3. Entropies DS and enthalpies DH of hydride decomposition are listed in Tables 2e4. Values of DHcomp for alloys, where hydrogen equilibrium pressure during D was measured only at 623 K, were calculated using known (almost invariant [45] p. 86) value DS ¼ 0.135 ± 0.002 J/mol/K from Eq. (1).
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 4 4 ( 2 0 1 9 ) 8 3 1 5 e8 3 2 4
8319
Fig. 1 e Sorption curves cH(t) measured at 623 K for cCB ¼ 12e14 wt% CB. (a) e A, (b) e D. 1: Al, 2: Si, 3: Ga2, 4: LiF, 5: Na, 6: NaF2, 7: Ge, 8: NaCl, 9: LiCl. Sorption curves for pure Mg measured in the present work are also plotted for comparison.
DHcomp
DS lnpeq : ¼ R,T R
(1)
In Eq. (1), R stands for gas constant, T is temperature, DS is entropy of hydride decomposition and peq is equilibrium hydrogen pressure during D at temperature T. Calculated
Fig. 2 e Influence of Ni co-alloying together with Mg17Al12 and Mg2Si (alloys NiAl and NiSi) upon sorption curves cH(t): comparison with alloys without Ni (alloys Si and Al); T ¼ 623 K.
values of DHcomp are listed in Table 2 also in cases where DH was directly measured in PCT experiments. It follows from comparison of DH and DHcomp that Eq. (1) gives quite reasonable results.
Fig. 3 e Example of measured PCT curves. Alloy NaCl.
8320
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 4 4 ( 2 0 1 9 ) 8 3 1 5 e8 3 2 4
Table 3 e Enthalpy of hydride decay in alloys NaFx calculated from Eq. (1) using measured values of peq at T ¼ 623 K and average value of DS ¼ 0.135 ± 0.002 J/mol H2/K. sample
NaF1 NaF2 NaF3 NaF4
peq
DHcomp
MPa
kJ/mol H2
0.600 0.573 0.567 0.604
± 0.005 ± 0.003 ± 0.005 ± 0.003
74.8 75.1 75.1 74.8
± 3.0 ± 3.0 ± 3.0 ± 3.0
Table 4 e Enthalpy of hydride decay in alloys Gax calculated from Eq. (1) using measured values of peq at T ¼ 623 K and average value of DS ¼ 0.135 ± 0.002 J/mol H2/K. sample
Ga1 Ga2 Ga3
Fig. 4 e Van't Hoff diagram. Comparison of present results for desorption with literature data known for Mg (1 e [44], 2 e [46], 3 e [47], 4 e [48]).
The calculated values of DHcomp are equal in frame of errors with measured values DH and with known literature values for pure MgH2 hydride. This means that the modification of Mg by X did not lead to any significant destabilization of thermodynamic stability of its hydride MgH2. The slight decrease in DHcomp observed in some cases may represent decrease in temperature of the start of hydrogen desorption by several degrees only. To quantify this temperature shift more precisely, temperature programmed desorption was studied.
Temperature programmed desorption (TPD) Fully hydrogen charged samples (2.5 MPa of H2 at temperature 653 K) were desorbed in temperature-increase regime using
peq
DHcomp
MPa
kJ/mol H2
0.55 ± 0.04 0.56 ± 0.04 0.55 ± 0.04
75.3 ± 2.0 75.5 ± 3.0 75.3 ± 2.0
the rate 10 K/min. Relative hydrogen concentration a (a ¼ cH/ cHmax) was registered in dependence on time of desorption, t. The numerically obtained time derivative da/dt is plotted in Fig. 5 as a function of increasing temperature T(t). The temperature of desorption peak Tpeak identifies the temperature of the most intensive hydrogen desorption and therefore, it can be defined to be the temperature of desorption in modified Mg. Contrary to the fact that measured enthalpies DH were almost un-affected by X, it can be clearly seen that the desorption temperature Tpeak depends on X quite convincingly. The lowest desorption temperature showed Mg modified by Mg2Si, the highest Tpeak was obtained for NiAl a NiSi. These extreme values correspond quite reasonably with observed kinetic behavior of alloys Si on one hand and NiAl and NiSi on the other (see Fig. 2). It is also obvious in Fig. 5 that the peak temperature, Tpeak, measured for alloys NiAl and NiSi was even higher than Tpeak in Mg. All other alloys showed decrease in hydrogen desorption temperatures.
Table 2 e Enthalpy and entropy of hydride decay. Mg and alloys with approximately equal concentration of additive and carbon (cX ≅ cCB ≅ 12e14 wt %) are listed. material
Mg Mg Mg Mg Si Al Ga2 NaCl LiF Ge NaF2 LiCl
DH
DS
DHcomp*)
kJ/mol H2 74.1 76.6 74.2 74.5 77.2 77.6 76.9 77.6 75.3
kJ/mol H2/K
kJ/mol H2
± 2.9
0.133 ± 0.005 0.137 0.134
75.6 ± 3.0 75.0 ± 3.0 75.0 ± 3.0
± 5.8 ± 1.5 ± 3.2 ± 3.0
0.138 0.140 0.138 0.140 0.135
± 0.009 ± 0.002 ± 0.005 ± 0.004
75.2 74.4 75.5 74.3 75.0 74.8 75.1 74.2
± 3.0 ± 3.0 ± 3.0 ± 3.0 ± 3.0 ± 3.0 ± 3.0 ± 3.0
remark
[44], 100e1000 mm Hg [46], 587 Ke662 K [47] [48] This work This work This work This work This work This work This work This work
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 4 4 ( 2 0 1 9 ) 8 3 1 5 e8 3 2 4
Fig. 5 e TPD results.
8321
Fig. 6 e Equilibrium hydrogen pressure, peq, measured during desorption at 623 K, in dependence on desorption temperature Tpeak, measured by TPD using the rate 10 K/ min. Shaded area e values of peq calculated from literature data on enthalpy and entropy of hydride decay (Table 2).
Effect of studied additives upon HS parameters To assess the influence of studied X's upon the parameters that determine the HS behavior of studied alloys, it is effectual to combine results of kinetic (t09), PCT (peq) at 623 K and TPD (Tpeak) measurements. In Fig. 6, peq is plotted in dependence on Tpeak. The shaded area covers known literature values of peq at 623 K [44,46e48]. Unfortunately, no details on Tpeak are known to literature data on peq for Mg from Refs. [44,46e48]. However, our point peq ¼ 0.54 ± 0.05 MPa measured at 623 K in pure Mg at Tpeak ¼ 631 K (plotted in Fig. 6) may serve as a reference point to compare results measured with experimental alloys with pure Mg. It can be seen that our value of peq falls well into the interval of literature data. The alloys lying above the shaded region, show decrease in the stability of MgH2. It is obvious that alloys Si and NaCl show the most beneficial hydrogen storage performance: The lowest stability of hydride phase (i.e., the highest peq), which is lower than that by pure Mg, and, at the same time, the lowest hydrogen desorption temperature Tpeak. Results of kinetic study and TPD measurements are compared in Fig. 7. It is obvious that the alloys that slightly destabilize MgH2, led also to acceleration of hydrogen sorption. The worst additives with the longest time t09 on the other hand, were X ¼ MgmGan, and combinations Ni þ Mg2Si and Ni þ Mg17Al12. To assess the results showed in Figs. 6 and 7 in perspective of structure changes during A/D (see par. 3.2), it can be concluded that the best HS performance (high peq and, at the same time, low Tpeak and t09) show alloys, which phase composition seems to be invariant during hydrogen sorption (Ge, LiF, NaF, Si), and/or alloys that tend to form easily significant fraction of AM (NaCl, LiCl). Alloy Al is seemingly a special case in these scheme. However, if it is considered that X ¼ Mg17Al12 is,
in fact, mashed in ball mill into Mg and Al (see par. 3.2.), it is obvious that the sorption experiment was not conducted with alloy Al but with alloy Al’ with phase composition (Mg)-Al-CB that can be considered invariant during subsequent A/D cycling. This idea can explain that Al’ show acceptable HS performance (Figs. 6 and 7).
Fig. 7 e Hydrogen desorption rate at 623 K scaled by time t09 in dependence on peak temperature Tpeak measured by TPD.
8322
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 4 4 ( 2 0 1 9 ) 8 3 1 5 e8 3 2 4
Fig. 8 e Comparison of hydrogen sorption kinetic curves with literature (broken lines): Mg-Al-C-Nb2O5 [50]. Mg-Sb-C [16], Mg-M-Li [30], MgH2-VO2 [49]. (a) e A, (b) e D. Thick solid lines in (a, b) e NaF2, thick dotted lines in (b) e Si, Ge, NaCl).
Comparison with literature data Contrary to the fact that a great number of studies were devoted to the hydrogen sorption kinetics in Mg-based alloys, it is not easy to compare adequately the data sets. This is because of diverse test temperatures, various sample preparation and different experimental details. This may make difficult the comparison of desorption data. Comparison of desorption kinetics is strongly conditioned by hydrogen pressure p. The high value of p may considerably slow-down the hydrogen desorption in the vicinity of peq. Therefore, it is advisable to make comparison of desorption data sets at higher temperature, with p well below peq. In Fig. 8, the present sorption data are compared with literature [16,30,49,50] for the temperature 623 K, where the hydrogen pressure during desorption p 0.03 MPa was well bellow equilibrium pressure p y 0.7 MPa. All the shown literature data were obtained under similar conditions as the present results. For the sake of simplicity, the literature data are plotted in the form of broken lines going from origin and through knots in 0.9cHmax and in cHmax. Individual literature kinetic curves cH(t) in Fig. 8 refer to different content of C, Nb2O5 and different grain size [50], to different content of Sb [16], and to different concentration of added elements (Ag, Al, Zn, Y) [50], or VO2 [49] (for details on alloys labeling e see the original papers). It is obvious that the present data represent improvement in hydrogen sorption kinetics compared to pure Mg (see in Fig. 8a for A and in Fig. 8b for D [16]) and to other available literature results [16,49,50] irrespective of composition and other experimental parameters. Just to note: the admixture of transition metal oxides (VO2, Nb2O5) and compounds containing Li are considered to be the most efficient modifier of HS Mg-based alloys [3,5]. It is easy to find that the hydrogen
desorption kinetics observed in the present work is better than kinetics found in Mg(In)-MgF2 and in Mg85In5Al5Ti5 [18,19] and also much better than that reported in Ref. [10] for Mg alloys modified by yttrium.
Conclusion Several chosen additives X were prejudged as potentially good catalysts for hydrogen sorption kinetics in mixture Mg-X-C. The aim of the present work was to find a composition that would improve the HS kinetics of Mg. It was found that admixture of NaCl, Mg2Si, and LiF most effectively increase the hydrogen equilibrium pressure at 623 K (slightly destabilize MgH2) and decrease t09 and Tpeak. NaCl and LiF shift the peak temperature to the lowest temperature Tpeak ¼ 597 K. On the other hand, Mg modified by MgmGan and by combinations of Ni þ Mg17Al12 and Ni þ Mg2Si showed bad HS performance at temperature 623 K.
Acknowledgements This work was supported by project of Czech Science Foundation No. 17-21683S and by Central European Institute of Technology, CEITEC-Institute of Physics of Materials, AS CR Brno, Czech Republic.
Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.ijhydene.2019.02.084.
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 4 4 ( 2 0 1 9 ) 8 3 1 5 e8 3 2 4
references
[1] Huot J. Metal hydrides. In: Hirscher M, editor. Handbook of hydrogen storage. Weinheim: WILEY-VCH Verlag GmbH &Co. KGaA; 2010. p. 81e110. [2] Crivelo JC, Denis RV, Dornheim M, Felderhoff M, Grant DM, Huot J, Jensen TR, deJongh P, Latroche M, Walker GS, Webb CJ, Yartys VA. Mg-based compounds for hydrogen storage. Appl Phys 2016;85:122e39. [3] Khafidz NZK, Yaakob Z, Lim KL, Timmiati SN. The kinetics of lightweight solid-state hydrogen storage materials: a review. Int J Hydrog Energy 2016;41:13131e51. [4] Barthelemy H, Weber M, Barbier F. Hydrogen storage: recent improvements and industrial perspectives. Int J Hydrog Energy 2017;42:7254e62. [5] Ren J, Musyoka NM, Langmi HW, Mathe M, Liao S. Current research trends and perspectives on materials-based hydrogen storage solutions: a critical review. Int J Hydrog Energy 2017;42:289e311. [6] Zhang J, Yan S, Qu H. Recent progress in magnesium hydride modified through catalysis and nanoconfinement. Int J Hydrog Energy 2018;43:1545e65. [7] Wang H, Lin HJ, Cai WT, Ouyang LY, Zhu M. Tuning kinetics and thermodynamics of hydrogen storage in light metal element based systems e a review of recent progress. J Alloy Compd 2016;658:280e300. [8] Zhang X, Leng Z, Gao M, Hu J, Du F, Yao J, Pan H, Liu Y. Enhanced hydrogen storage properties of MgH2 catalyzed with carbon-supported nanocrystalline TiO2. J Power Sources 2018;398. 183-92. [9] Korablov D, Besenbacher F, Jensen TR. Kinetics and thermodynamics of hydrogenation-dehydrogenation for Mg25%TM (TM ¼ Ti, Nb or V) composites synthesized by reactive ball milling in hydrogen. Int J Hydrog Energy 2018;43:16804e14. [10] Yang T, Li Q, Liu N, Liang C, Yin F, Zhang Y. Improved hydrogen absorption and desorption kinetics of magnesiumbased alloy via addition of yttrium. J Power Sources 2018;378:636e45. [11] Yahya MS, Ismail M. Synergistic catalytic effect of SrTiO3 and Ni on the hydrogen storage properties of MgH2. Int J Hydrog Energy 2018;43:6244e55. [12] Yahya MS, Sulaiman NN, Mustafa NS, Yap FAH, Ismail M. Improvement of hydrogen storage properties in MgH2 catalysed by K2NbF7. Int J Hydrog Energy 2018;43:14532e40. [13] Xu E, Li H, You X, Bu C, Zhang L, Wang Q, Zhao Z. Influence of micro-amount O2 or N2 on the hydrogenation/ dehydrogenation kinetics of hydrogen-storage material MgH2. Int J Hydrog Energy 2017;42:8057e62. [14] Floriano R, Deledda S, Hauback BC, Leiva DR, Botta WJ. Iron and niobium based additives in magnesium hydride: microstructure and hydrogen storage properties. Int J Hydrog Energy 2017;42:6810e9. [15] Xie L, Li J, Zhang T, Song L, Kou H. Microstructure and hydrogen storage properties of Mg-Ni-Ce alloys with a longperiod stacking ordered phase. J Power Sources 2017;338:91e102. [16] Cermak J, Kral L, Roupcova P. Improvement of hydrogen storage kinetics in ball-milled magnesium doped with antimony. Int J Hydrog Energy 2017;42:6144e51. [17] Korablov D, Besenbacher F, Jensen TR. Ternary compounds in the magnesium-titanium hydrogen storage system. Int J Hydrog Energy 2014;39:9700e8. [18] Ouyang LZ, Cao ZJ, Wang H, Liu JW, Sun DL, Zhang QA, Zhu M. Enhanced dehydriding thermodynamics and kinetics in Mg(In)eMgF2 composite directly synthesized by plasma milling. J Alloy Compd 2014;586:113e7.
8323
[19] Cao Z, Ouyang L, Wu Y, Wang H, Liu J, Fang F, Sun D, Zhang Q, Zhu M. Dual-tuning effects of In, Al, and Ti on the thermodynamics and kinetics of Mg85In5Al5Ti5 alloy synthesized by plasma milling. J Alloy Compd 2015;623:354e8. [20] Valentoni A, Mulas G, Enzo S, Garroni S. Remarkable hydrogen storage properties of MgH2 doped with VNbO5. Phys Chem Chem Phys 2018;20:4100e8. n M, Al-Ajmi F. Discovering a new [21] El-Eskandarany MS, Banya MgH2 metastable phase. RSC Adv 2018;8:15667e74. [22] Pukazhselvan D, Nasani N, Correia P, Carbo-Argibay E, OteroIrurueta G, Stroppa DG, Fagg DP. Evolution of reduced Ti containing phase(s) in MgH2/TiO2 system and its effect on the hydrogen storage behavior of MgH2. J Power Sources 2017;362:174e83. [23] Yahya MS, Ismail M. Catalytic effect of SrTiO3 on the hydrogen storage behaviour of MgH2. J Energy Chemistry 2019;28:46e53. https://doi.org/10.1016/j.jechem.2017.10.020. [24] Ismail M, Mustafa NS, Juahir N, Yap FAH. Catalytic effect of CeCl3 on the hydrogen storage properties of MgH2. Mater Chem Phys 2016;170:77e82. [25] Sulaiman NN, Juahir N, Mustafa NS, Yap FAH, Ismail M. Improved hydrogen storage properties of MgH2 catalyzed with K2NiF6. J Energy Chemistry 2016;25:832e9. [26] Yahya MS, Ismail M. Improvement of hydrogen storage properties of MgH2 catalyzed by K2NbF7 and multiwall carbon nanotube. J Phys Chem C 2018;122:11222e33. [27] Li Q, Lin Q, Chou K-C, Jiang L. A study on the hydridingdehydriding kinetics of Mg1.9Al0.1Ni. J Mater Sci 2001;39:61e5. [28] Yhu Y, Liu Z, Yang Y, Gu H, Li L, Cai M. Hydrogen storage properties of MgeNieC system hydrogen storage materials prepared by hydriding combustion synthesis and mechanical milling. Int J Hydrog Energy 2010;35:6350e5. [29] Cermak J, Kral L. Alloying of Mg/Mg2Ni eutectic by chosen non-hydride forming elements: relation between segregation of the third element and hydride storage capacity. J Power Sources 2012;197:116e20. [30] Suarez-Alcantara K, Palacios-Lazcano AF, Funatsu T, Cabanas-Moreno JG. Mg-M-LiH alloys prepared by mechanical milling and their hydrogen storage characteristics. Int J Hydrog energy 2015;40:17344e53. [31] Yap FAH, Yahya MS, Ismail M. Enhancement of hydrogen storage properties in 4MgH2-Na3AlH6 composite catalyzed by TiF3. Int J Hydrog Energy 2017;42:21096e104. [32] Zhang H, Fu L, Xuan W, Qin H, Ji Z. Ni-Al films induced surface modification of La2Mg17 alloy leading to improved dehydrogenation properties. J Power Sources 2018;385:27e31. [33] Cermak J, Kral L. Hydrogen diffusion in MgeH and MgeNieH alloys. Acta Mater 2008;56:2677e86. [34] Ding X, Chen R, Jin Y, Chen X, Guo J, Su Y, Ding H, Fu H. Activation mechanism and dehydrogenation behavior in bulk hypo/hypereutectic Mg-Ni alloy. J Power Sources 2018;374:158e65. [35] Vajo JJ, Mertens F, Ahn CC, Bowman jr RC, Fultz B. Altering hydrogen storage properties by hydride destabilization through alloy formation: LiH and MgH2 destabilized with Si. J Phys Chem B 2004;108:13977e83. [36] Liu H, Wu C, Zhou H, Chen T, Liu Y, Wang X, Dong Z, Ge H, Li S, Yan M. Synergistically thermodynamic and kinetic tailoring of the hydrogen desorption properties of MgH2 by co-addition of AlH3 and CeF3. RSC Adv 2015;5:22091e6. [37] Lyci G, Surendra KS. Structural stability of metal hydrides, alanates and borohydrates of alkali and alkali-earth elements: a review. Int J Hydrog Energy 2010;35:5454e70. [38] Zhang B, Wu Y. Microstructural evolution and improved hydrogen storage properties for the Li3N-MgH2 system by addition of LiNH2 during the hydrogenation/ dehydrogenation. Int J Hydrog Energy 2015;40:9298e305.
8324
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 4 4 ( 2 0 1 9 ) 8 3 1 5 e8 3 2 4
[39] Oumellal Y, Courty M, Rougier A, Nazri GA, Aymard L. Electrochemical reactivity of magnesium hydride toward lithium: new synthesis route of nano-particles suitable for hydrogen storage. Int J Hydrog Energy 2014;39:5852e7. [40] Sulaiman NN, Ismail M. Effects of TiF3 addition on the hydrogen storage properties of 4MgH2 þ Cd composite. Int J Hydrog Energy 2018:1e9. accepted, https://doi.org/10.1016/j. ijhydene.2018.03.095. [41] Cermak J, Kral L. Hydrogen absorption in carbon allotropes and talc. Kovove Mater Metallic Mater 2018;56:75e80. [42] Zhou S, Zhang X, Li T, Wang N, Chen H, Zhang T, Yu H, Niu H, Liu D. Nano-confined magnesium for hydrogen storage from reactive milling with anthracite carbon as milling aid. Int J Hydrog Energy 2014;39:13628e33. [43] Vejeeston P, Ravindran P, Hauback BC, Fjellvag H, Kjekshus A, Furuseth S, Hanfland M. Phys Rev B 2006;73:224102. [44] Kennelley JA, Varwig JW, Myers HW. Magnesium-hydrogen relationships. J Phys Chem 1960;64:703e4. [45] Dornheim M. Tailoring reaction enthalpies of hydrides. In: Hirscher M, editor. Handbook of hydrogen storage.
[46] [47] [48]
[49]
[50]
Weinheim: WILEY-VCH Verlag GmbH &Co. KGaA; 2010. p. 187e214. Stampfer JF, Holley CE, Suttle FJ. The magnesium-hydrogen system. J Am Chem Soc 1960;82:3504e8. Schwarz RB. Hydrogen storage in Magnesium-based alloys. MRS Bull 1999;11:40e4. Gerard N, Ono S. Hydride formation and decomposition kinetics. In: Schlapbach L, editor. Hydrogen in intermetallic compounds II. Berlin, Heidelberg, New York, London, Paris, Tokyo, Hong Kong, Barcelona, Budapest: Springer-Verlag; 1992. p. 165e96. Milosevic S, Kurko S, Pasquini L, Matovic L, Vujasin R, Novakovic N, Grbovic Novakovic J. Fast hydrogen sorption from MgH2-VO2(B) composite materials. J Power Sources 2016;307:481e8. Hardian R, Pistida C, Chaudhary A-L, Capurso G, Gizer G, Cao H, Milanese C, Girella A, Santoru A, Yigit D, Dieringa H, Kainer KU, Klassen T, Dornheim M. Waste Mg-Al based alloys for hydrogen storage. Int J Hydrog Energy 2018;43:16738e48.
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/he
Improved hydrogen sorption kinetics in Mg modified by chosen catalysts Jiri Cermak a,b,*,1, Lubomir Kral a, Pavla Roupcova a,c a
Institute of Physics of Materials AS CR, v.v.i., Zizkova 22, CZ-61662 Brno, EU, Czech Republic CEITEC-Institute of Physics of Materials, AS CR, Zizkova 22, CZ-61662 Brno, EU, Czech Republic c CEITEC-Brno University of Technology, Purkynova 123, CZ-61662 Brno, EU, Czech Republic b
article info
abstract
Article history:
Hydrogen sorption characteristics were investigated in Mg modified by ten chosen cata-
Received 4 October 2018
lyzing additives X ¼ Mg2Si, Mg2Ge, Mg17Al12, Mg5Ga2, NaCl, LiCl, NaF, LiF and by two
Received in revised form
combinations Ni þ Mg17Al12 and Ni þ Mg2Si. Amorphous carbon (CB) was used as an anti-
1 February 2019
sticking component. It was found that optimum content of both X and CB was between 10
Accepted 12 February 2019
and 14 wt %. To assess the effect of X itself upon the sorption kinetics, concentration of
Available online 7 March 2019
both X and CB was fixed to about 12 wt %. The shortest desorption times t09 (time to
Keywords:
NaCl. Careful measurements of equilibrium hydrogen pressure, peq, at 623 K showed an
desorption of 90 pct of stored hydrogen) showed alloys with X ¼ NaF, Mg2Si, Mg2Ge and Hydrogen storage
increase in peq for added chlorides, fluorides, Mg2Ge and also for Mg2Si. TPD experiments
Mg alloys
revealed desorption peaks for X ¼ NaCl, LiF shifted down to temperatures as low as 597 K.
Carbon black
© 2019 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
Catalysis
Introduction Mg-based alloys are promising materials for hydrogen storage (HS) applications [1e6]. The principal HS phase in alloys of this type is MgH2 hydride. This phase shows high HS capacity (7.6 wt % H2) and excellent cycling behavior. The substantial drawback is its high thermodynamic stability, which implies high operation temperatures. Much research work was devoted to decrease the high stability of MgH2 (see, e.g., in Refs. [7e19]), i.e. to increase hydrogen pressure in equilibrium, peq, above the value for pure Mg hydride. This would decrease working temperatures of HS reservoirs down to some 373 K, which is a desired temperature region for HS applications (see US Department of energy target parameters [4]). However, this
proved to be an extraordinary difficult task. After a considerable research effort, only moderate success was reached in this respect [6]. Promising new techniques of HS materials preparation such as dual tuning of plasma milled samples [18,19] did not lead to any significant increase in equilibrium hydrogen pressure. Much expressive progress in the last years was reported in the field of improvement of the hydrogen sorption kinetics in Mg-based HS materials [4e6]. Of course, the dynamics and kinetics are closely interconnected, but it seems that influencing the obstacles that inhibit the sorption kinetics is much easier task then to affect effectively the generalized driving forces of the sorption process. Main ways how to improve the kinetics are the catalysis of partial sorption steps [20e26] and nanosizing [6,27,28].
* Corresponding author. Institute of Physics of Materials AS CR, v.v.i., Zizkova 22, CZ-61662 Brno, EU, Czech Republic. E-mail address: [email protected] (J. Cermak). 1 Permanent address: Institute of Physics of Materials v. v. i., Academy of Sciences of the Czech Republic, Zizkova 22, CZ-61662 Brno, Czech Republic. https://doi.org/10.1016/j.ijhydene.2019.02.084 0360-3199/© 2019 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.
8316
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 4 4 ( 2 0 1 9 ) 8 3 1 5 e8 3 2 4
In the present work, HS characteristics of ball-milled (BM) blend of Mg with chosen additives were investigated. Five types of additives were tested: (i) Mg-based intermetallics formed by two elements from 13th group (Mg2Ge, Mg2Si), and by (ii) two elements of 14th group (Mg5Ga2, Mg17Al12). (iii) Two combinations Ni þ Mg17Al12 and Ni þ Mg2Si, (iv) two simple fluorides with Naþ and Liþ (NaF, LiF) and (v) two chlorides with the same cations (NaCl, LiCl) were also tested. Expectation of beneficial effect of the intermetallics with elements from 13th and 14th group upon the absorption/desorption kinetics (A/D) kinetics was based on promising results published in Refs. [18,29e32], synergy of those elements and Ni could be expected with regard to results from Refs. [33,34] and improvement of A/D kinetic by anions F, Cl and/or cations Naþ, Liþ might be expected from results published in Refs. [30,31,35e40]. Besides the above mentioned additives, also carbon black (CB) was introduced into the BM blends as antisticking agens [6,7] and also, as components that may take part in hydrogen sorption [41,42].
Experimental Samples were made from pure components purchased by Sigma-Aldrich. Mg was in the form of pieces (purity: 3N8), other metallic components e Ge(5N), Si(6N), Ga(4N), Al(5N), Ni(3N), halides e LiCl, NaF, NaCl, LiF (purity of all halides >2N) and CB (2N) were purchased in form of fine powders. Intermetallics Mg2Si and Mg2Ge were prepared by mechanical alloying in form of powders. Intermetalic compounds Mg17Al12 and Mg5Ga2 were induction melted in pure Ar(4N) and grinded into form of splinters. All the additives were mixed with splinters of Mg and ball milled in hydrogen atmosphere (6N) using Fritsch Pulverisette6 ball-mill. Two HS samples were prepared with admixture of fine powder of Ni. The mass ratio of the milling balls to the milled blend was about 60 and the milling cycle e 10 min milling/50 min cooling e was repeated 90 times. Hydrogen absorption under pressure p ¼ 2.5 MPa, desorption into a fixed volume with hydrogen pressure always well below the peq(T), and temperature programmed desorption (TPD) was carried out using Sievertstype gas sorption analyzer PCTePro Setaram Instrumentation at temperatures between 575 K and 658 K. All manipulations of the milled blend inclusive the filling the cuvette were done in the glove box in protective Ar atmosphere. Cuvette with the sample was transported in a small container filled by Ar and inserted into the sorption apparatus. Nominal chemical composition of ball milled batches is listed in Table 1. Phase composition of samples was obtained by XRD EMPYREAN device using CoKa radiation and the results were interpreted (Rietveld analysis) with the HighScore Plus Software and ICSD databases (Inorganic Crystal Structure Database Fachinformationszentrum Karlsruhe, Germany, and the U.S. Department of Commerce of the United States, version 1.9.8.). Accuracy of phase composition was about 2 wt %. Morphology of milled samples, average chemical composition and chemical maps were observed by SEM TESCAN LYRA3 equipped with X-max80 EDS in the area approximately 300 500 mm containing about 102 grains. Accuracy of average concentration of substitution elements (except for Li) was
within 1.5 wt%. Content of amorphous phase, cAM, (labeled below in the text as AM) and concentration of carbon, cC were fixed to value cCB, known from initial composition of the ball milling charge (see Table 1). Lithium concentration, cLi, was calculated from stoichiometric ratio. In LiF-containing samples, the value cLi ¼ 0.36537 cF and, for cLi in LiCl-containing samples the value cLi ¼ 0.1957 cCl was taken, where concentrations cF and cCl were obtained by SEM.
Results and discussion Characterization of samples Morphology of all samples after the BM, its change after hydrogen absorption (A) and hydrogen desorption steps (D) follows very similar evolution scheme. It can be illustrated by the sample Ge e see supplementary material, Fig. S1. The structure of Ge after BM is showed in Fig. S1a, where relatively smooth grains of size typically between 5 and 50 mm can be seen. These grains, composed principally of solid solution (Mg), transform during A to hydride phase and their appearance becomes corny e Fig. S1b. After D, when the hydrided phase(s) transformed back to hydride-free components, the corny structure persists and, moreover, additional fragmentation and transversal cracks of great grains appear e Fig. S1c. Fine particles of size about 1 mm (for Ge e bright submicron grains, see in Fig. S1a) were identified as impurities with traces of Ti, Al, W and Fe. Concentration of these elements was lower than 1 wt % and therefore, it was neglected in SEM analyses. XRD patterns were measured for each sample in BM, A and D states. The obtained data could be classified qualitatively into two considerably different groups. The first typical behavior can be described by the scheme: cAM in BM y cAM in C y cAM in D with cAM ¼ cCB (cAM stands for concentration of amorphous phase from XRD analysis, cCB is known concentration of carbon black). This type of behavior was observed in all alloys except for chloride-containing samples. The samples modified by chlorides (LiCl and NaCl) on the other hand, formed the other group where the content of AM measured by XRD was very high, especially in BM and D. The typical XRD patterns are exemplified in supplementary material e see Fig. S2 for Ge and NaCl. No carbides were detected in any sample; the components, which content was below 1% were neglected.
Structure after BM and A/D cycle Results of XRD and SEM observations of phase and chemical composition in states BM, A and D can be shortly presented as follows:
Sample Si After BM (indicated phases: 79(Mg)-4Mg2Si-2Si-2aMgH2-13AM; elements found: 84 Mg-3Si-13C) all in wt. %. Si was present both as regularly dispersed fine particles of Mg2Si and as scattered elemental Si in the form of particles of size up to 1 mm. Separate XRD analysis of ball milled blend Mg þ Si showed that the BM produced Mg2Si compound only up to about 21%. This explained observed differences between
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 4 4 ( 2 0 1 9 ) 8 3 1 5 e8 3 2 4
8317
Table 1 e Experimental samples. CB e carbon black. Sample label Si Ge Al Ga1 Ga2 Ga3 NiSi NiAl LiF Na NaF1 NaF2 NaF3 NaF4 LiCl NaCl a
Nominal phase composition of batch (wt. %) 76.36 75.28 76.58 81.61 75.81 68.19 71.87 72.37 74.04 90.91 85.60 74.81 60.00 70.67 75.94 74.61
Mge10.25 Mg2Si e 13.39 CB Mge12.61 Mg2Ge e 12.11 CB Mge10.40 Mg17Al12 e 13.02 CB Mge10.91 Mg5Ga2 e 7.48 CB Mge10.26 Mg5Ga2 e 13.93 CB Mge9.09 Mg5Ga2 e 22.72 CB Mge3.26 Ni e 12.47 Mg2Si e 12.40 CB Mge2.53 Ni e 12.84 Mg17Al12 e 12.31 CB Mge13.18 LiF e 12.78 CB Mge9.09 NaH Mge7.14 NaF e 7.26 CB Mge12.41 NaF e 12.78 CB Mge19.99 NaF e 20.01 CB Mge22.13 NaF e 7.20 CB Mge11.57 LiCl e 12.49 CB Mge12.99 NaCl 12.40 CB
Chemical composition (wt. %) 81.79 Mge4.82 Si e 13.39C 80.32 Mge7.57 Ge e 12.11C 82.48 Mge4.5 Al e 13.02C 86.72 Mge5.80 Ga 7.48C 80.61 Mge5.46 Ga e 13.93C 72.44 Mge4.84 Ga e 22.72C 79,76 Mge3,26 Ni e 4,58 Si e 12,40 C 79,82 Mge2,53 Ni e 5,39 Al e 12,31 C 74.04 Mge3.35 Li e 9.65 F e 12.78C 91.26e8.74 Naa 85.60 Mge3.91 Na e 3.23 Fe7.26C 74.81 Mge6.81 Na e 5.60 Fe12.78C 60.00 Mge10.95 Na e 9.04 Fe20.01C 70.67 Mge12.13 Na e 10.00 Fe7.20C 75,94 Mge1,89 Li e 9,68 Cl e 12,49 C 74.60 Mge5.20 Na 7.80 CL e 12.40C
In the initial state before sorption experiments (after decay of NaH).
nominal and measured phase composition in BM of sample Si. Presence of aMgH2 was a consequence of BM in H2. Chemical analysis agreed reasonably (within experimental errors) with the nominal values (Table 1). Carbon was distributed regularly. After A (78aMgH2-9Mg2Si-13AM; 78 Mg-3Si-13C-6H), lowpressure alpha hydride phase appeared, all elemental Si reacted with Mg and formed particles Mg2Si (size up to 2 mm). In D state (76(Mg)-11Mg2Si-13AM; 84 Mg-3Si-13C), particles Mg2Si (size up to 2 mm) persisted.
Sample Ge After BM (75(Mg)-12Mg2Ge-13AM; 81 Mg-7Ge-12C), Mg2Ge particles (~5 mm) were observed, C was uniformly distributed, chemical concentration agreed with nominal composition. After A (70aMgH2-11Mg2Ge-7(Mg)-12AM; 78 Mg-5Ge-11C-6H), Mg2Ge particles became less distinctly bounded. After D (75(Mg)-13Mg2Ge-12AM; 83 Mg-5Ge-12C), the original state was almost exactly restored.
Sample Al After BM (83(Mg)-3Mg1.95Al0.05-1MgH2-13AM; 82 Mg-5Al-13C), no Mg17Al12 phase was detected. It seemed that it was decayed during BM and the products entered (Mg) phase. Al was distributed approximately uniformly; slightly higher Al concentration could be detected in Mg1.95Al0.05 particles (size between 1 and 10 mm). After A (74aMgH2-13(Mg)-13AM; 79 Mg3Al-12C-6H), low-pressure hydride aMgH2 appeared, presence of (Mg) solid solution was a consequence of incomplete absorption. After D (86(Mg)-1MgH2-13AM; 82 Mg-5Al-13C), the state observed after BM was restored.
Sample Ga2 After BM (86(Mg)-14AM; 86 Mg-7Ga-7C), all components were uniformly scattered. Mg5Ga2 phase decomposed during BM and the decay products entered the AM and partly also the (Mg). After A (86bMgH2-4Mg5Ga2-3Mg2Ga-7AM; 85 Mg-2Ga-7C6H), one of high pressure phases was formed (b). The phase b was observed also in ball milled alloys Mg-M (M ¼ Ti, Nb, V) [9,17,43]. Loci (~1m) with significantly higher concentration of Ga were observed, which were identified as two MgmGan
phases. They precipitated from AM. After D (85((Mg)-8Mg5Ga27AM; 88 Mg-5Ga-7C), Mg2Ga disappeared whereas Mg5Ga2 particles persisted.
Sample NiSi After BM (77(Mg)-4Mg2Si-1Ni-2Si-5aMgH2-12AM; 80 Mg-4Ni4Si-12C), Si particles (~2 mm) that anti-coincided with Mg were observed. Mg2Si was not completely formed (similarly as in sample Si above). After A (76aMgH2-1(Mg)-10Mg2Si-1Ni-12AM, 77 Mg-2Ni-3Si-12C-6H), Formation of Mg2Si phase was completed (particles ~2e3 mm). After D (75(Mg)-12Mg2Si-13AM; 81 Mg-3Ni-4Si-12C), particles of Mg2Si were restored.
Sample NiAl After BM (85(Mg)-1NiAl-1Ni-1aMgH2-12AM; 80 Mg-3Ni-5Al12C), the Mg17Al12 decayed and the residuals entered (Mg). After A (82aMgH2-1(Mg)-5Ni2Al3-12AM; 77 Mg-2Ni-3Al-12C6H), Mg was mostly bound in MgH2, Sporadic Mg particles were also observed. Al was scattered uniformly; no localized phase particles composed of Ni, Al and Mg were observed. After D (83(Mg)-4Ni2Al3-1NiAl-12AM; 80 Mg-3Ni-5Al-12C), new-formed phases NimAln replaced completely originally introduced phase Mg17Al12 in hydrogen-free sample.
Sample LiF After BM (75(Mg)-12LiF-13AM; 75 Mg-3Li-9F-13C), Small particles LiF (~3 mm) were observed on the surface of great Mg grains. C was scattered uniformly. After A (83bMgH2-2LiF2(Mg)-13AM; 68 Mg-4Li-10F-12C-6H), the identity of LiF phase persisted even after the hydrogen absorption. High pressure phase bMgH2 was formed by BM; (Mg) phase was a consequence of incomplete absorption. After D (75(Mg)-11LiF1MgH2-13AM; 73 Mg-4Li-10F-13C), the original state observed after BM was almost exactly restored.
Sample NaF2 After BM (80(Mg)-7NaMgF1.17H1.83e13AM; 75 Mg-6Na-6F-13C), NaF decomposed during BM and the decay products entered the (Mg) phase. A small fraction of ternary hydride was formed. All chemical components were scattered uniformly.
8318
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 4 4 ( 2 0 1 9 ) 8 3 1 5 e8 3 2 4
After A (84bMgH2-1NaMgH3-2(Mg)-13AM; 74 Mg-5Na-3F-12C6H), high pressure hydride bMgH2 was formed, ternary hydride changed its composition and F remained solved probably in AM and/or in 9Mg0 phases. After D (80(Mg)7NaMgF1.17H1.83e13AM; 79 Mg-4Na-4F-13C), the original state was approached, observed after BM.
Sample LiCl After BM (73(Mg)-11LiCl-4bMgH2-12AM; 76 Mg-2Li-10Cl-12C), both phase and chemical composition agreed approximately with nominal values; high pressure bMgH2 was formed. After A (72aMgH2-15Li-1(Mg)-12AM; 67 Mg-2Li-12Cl-12C-6H), LiCl was found completely decayed. Li precipitated in elemental form. After D (2(Mg)-98AM; 73 Mg-2Li-12Cl-12C), AM phase grew significantly. It was most likely due to interaction of Cl with Mg [43] that produced finally extremely fine (almost amorphous) phase. Positions of chlorides lines [43] are plotted in Fig. S2c. In case of fine structure they may broaden and superimpose over LiCl pattern. In results, they may be involved in overall AM phase.
Sample NaCl After BM (77(Mg)-8NaCl-3aMgH2-12AM; 75 Mg-5Na-8Cl-12C), NaCl phase was found partly decomposed; the decay products entered (Mg) and/or AM phase. After A (74aMgH2-9NaCl-5Mg12AM; 70 Mg-5Na-7Cl-12C-6H), NaCl phase remained also after hydrogen absorption. After D (10(Mg)-90AM; 74 Mg-5Na9Cl-12C), NaCl phase decayed and the transformation products entered the phase that behaves as AM. AM contained also prevailing fraction of Mg. Most likely, this was a result of fragmentation of MgH2 into very fine particles during interaction of Mg with Cl e reasoning is similar as in the case of sample LiCl.
Observed structure transformations e summary Chemical composition of all samples in states BM and D were reasonably equal to elements nominal concentrations in milling batches. On the basis of the present observation, the response of alloys structure during A/D cycle can be ranked into three distinct classes: Samples Ge, LiF and NaF form the first one. The phase composition of these samples in D is about the same as in A. Samples Si and NiSi can be also involved in this class, since the formation of Mg2Si compound is completed after the first A and then remains without changes. On the other hand, samples Al, NiAl and Ga in the second class went through a significant phase reformation leading to different phase composition in BM and D. Samples NaCl and LiCl in the last class showed also significant phase changes, but they tended to form amorphous phase already during BM. Great fraction of AM phase was found also in D.
Sorption kinetics Reproducibility A and D kinetics were studied in a repeated A/D cycling regime. The concentration-time curves, cH(t), were recorded after the stabilization of this process. Usually one or two A/D cycles were sufficient to reach the stable state, where the hydrogen storage capacity, cHmax, varied within about ±0.2% and reproducibility of time to 0.9 of cHmax, t09, was better
than 10 s for A and about 0.75 min for D e see in supplementary material - Fig. S3.
Optimum composition Measurements of cH(t) for varied values of cX and cCB (cX is concentration of additive X, concentration of the carbon black phase, cCB, is equal to concentration of carbon, cC) revealed that if the hydrogen sorption capacity, cHmax should be kept at about 6 wt % H2, and time t09 as short as possible, cX and cCB must be about 10e14 wt %. This is illustrated in Figs. S4 and S5 for alloys NaFx and Gax. It was observed, most likely due to alloying effects, that the sorption capacity might even exceed slightly the capacity of pure Mg (e.g., for sample Ga1 e Fig. S5). This led, however, to lowered sorption rate, which is general conclusion found for all studied alloys: the effect of X and CB upon the cHmax and t09 by desorption was significantly inverse. Increase in cHmax led to worse desorption kinetics.
Influence of X upon sorption kinetics Comparison of relative influence of additive X upon the sorption kinetics was carried out at temperature 623 K and with approximately equal content of CB and X. Measured sorption curves cH(t) are plotted in Fig. 1 for times up to 0.1 h compared with results for pure Mg. Hydrogen concentration cH(t) approached values about 6 wt % H2 in all cases after sufficiently long absorption time t. Very good sorption kinetics was found for sodium halides X ¼ NaF, NaCl and for alloys that were modified by Daltontype intermetallics X ¼ Mg2Si, Mg2Ge that proved to be stable during A/D cycle e see paragraph 3.2. Alloys Ga and alloy Na modified by pure sodium without halide element (introduced in the form of NaH hydride) showed relative slowest desorption kinetics. It can be seen that all additives shown in Fig. 1 improved the hydrogen sorption kinetics compared to hydrogen sorption in pure Mg. Worse desorption kinetics was observed with alloys modified by combination Ni þ Mg17Al12 and Ni þ Mg2Si e see Fig. 2. This was due most likely to unwanted interaction of Ni with other alloys components. Parasite phases of the type NimAln in alloy NiAl were even identified by XRD (see in paragraph Structure after BM and A/D cycle). Unwanted interaction between Mg, Ni and Si that must not be manifested directly in an explicit formation of unwished phases can be also expected [44].
Measurements of pressure-composition isotherms (PCT) Measurement of PCT curves was carried out with alloys Si, Al, Ga2, NaCl and LiF at temperatures between 573 K and 658 K. Other alloys, and also Mg e as a reference material e were studied at a single temperature 623 K. Typical results are illustrated in Fig. 3. Van't Hoff diagram is shown in Fig. 4, where the present results are compared with literature data for Mg. Our value of peq ¼ 0.54 ± 0.03 MPa obtained for Mg at temperature T ¼ 623 K agree quite well with other literature data for Mg plotted in Fig. 3. Entropies DS and enthalpies DH of hydride decomposition are listed in Tables 2e4. Values of DHcomp for alloys, where hydrogen equilibrium pressure during D was measured only at 623 K, were calculated using known (almost invariant [45] p. 86) value DS ¼ 0.135 ± 0.002 J/mol/K from Eq. (1).
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 4 4 ( 2 0 1 9 ) 8 3 1 5 e8 3 2 4
8319
Fig. 1 e Sorption curves cH(t) measured at 623 K for cCB ¼ 12e14 wt% CB. (a) e A, (b) e D. 1: Al, 2: Si, 3: Ga2, 4: LiF, 5: Na, 6: NaF2, 7: Ge, 8: NaCl, 9: LiCl. Sorption curves for pure Mg measured in the present work are also plotted for comparison.
DHcomp
DS lnpeq : ¼ R,T R
(1)
In Eq. (1), R stands for gas constant, T is temperature, DS is entropy of hydride decomposition and peq is equilibrium hydrogen pressure during D at temperature T. Calculated
Fig. 2 e Influence of Ni co-alloying together with Mg17Al12 and Mg2Si (alloys NiAl and NiSi) upon sorption curves cH(t): comparison with alloys without Ni (alloys Si and Al); T ¼ 623 K.
values of DHcomp are listed in Table 2 also in cases where DH was directly measured in PCT experiments. It follows from comparison of DH and DHcomp that Eq. (1) gives quite reasonable results.
Fig. 3 e Example of measured PCT curves. Alloy NaCl.
8320
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 4 4 ( 2 0 1 9 ) 8 3 1 5 e8 3 2 4
Table 3 e Enthalpy of hydride decay in alloys NaFx calculated from Eq. (1) using measured values of peq at T ¼ 623 K and average value of DS ¼ 0.135 ± 0.002 J/mol H2/K. sample
NaF1 NaF2 NaF3 NaF4
peq
DHcomp
MPa
kJ/mol H2
0.600 0.573 0.567 0.604
± 0.005 ± 0.003 ± 0.005 ± 0.003
74.8 75.1 75.1 74.8
± 3.0 ± 3.0 ± 3.0 ± 3.0
Table 4 e Enthalpy of hydride decay in alloys Gax calculated from Eq. (1) using measured values of peq at T ¼ 623 K and average value of DS ¼ 0.135 ± 0.002 J/mol H2/K. sample
Ga1 Ga2 Ga3
Fig. 4 e Van't Hoff diagram. Comparison of present results for desorption with literature data known for Mg (1 e [44], 2 e [46], 3 e [47], 4 e [48]).
The calculated values of DHcomp are equal in frame of errors with measured values DH and with known literature values for pure MgH2 hydride. This means that the modification of Mg by X did not lead to any significant destabilization of thermodynamic stability of its hydride MgH2. The slight decrease in DHcomp observed in some cases may represent decrease in temperature of the start of hydrogen desorption by several degrees only. To quantify this temperature shift more precisely, temperature programmed desorption was studied.
Temperature programmed desorption (TPD) Fully hydrogen charged samples (2.5 MPa of H2 at temperature 653 K) were desorbed in temperature-increase regime using
peq
DHcomp
MPa
kJ/mol H2
0.55 ± 0.04 0.56 ± 0.04 0.55 ± 0.04
75.3 ± 2.0 75.5 ± 3.0 75.3 ± 2.0
the rate 10 K/min. Relative hydrogen concentration a (a ¼ cH/ cHmax) was registered in dependence on time of desorption, t. The numerically obtained time derivative da/dt is plotted in Fig. 5 as a function of increasing temperature T(t). The temperature of desorption peak Tpeak identifies the temperature of the most intensive hydrogen desorption and therefore, it can be defined to be the temperature of desorption in modified Mg. Contrary to the fact that measured enthalpies DH were almost un-affected by X, it can be clearly seen that the desorption temperature Tpeak depends on X quite convincingly. The lowest desorption temperature showed Mg modified by Mg2Si, the highest Tpeak was obtained for NiAl a NiSi. These extreme values correspond quite reasonably with observed kinetic behavior of alloys Si on one hand and NiAl and NiSi on the other (see Fig. 2). It is also obvious in Fig. 5 that the peak temperature, Tpeak, measured for alloys NiAl and NiSi was even higher than Tpeak in Mg. All other alloys showed decrease in hydrogen desorption temperatures.
Table 2 e Enthalpy and entropy of hydride decay. Mg and alloys with approximately equal concentration of additive and carbon (cX ≅ cCB ≅ 12e14 wt %) are listed. material
Mg Mg Mg Mg Si Al Ga2 NaCl LiF Ge NaF2 LiCl
DH
DS
DHcomp*)
kJ/mol H2 74.1 76.6 74.2 74.5 77.2 77.6 76.9 77.6 75.3
kJ/mol H2/K
kJ/mol H2
± 2.9
0.133 ± 0.005 0.137 0.134
75.6 ± 3.0 75.0 ± 3.0 75.0 ± 3.0
± 5.8 ± 1.5 ± 3.2 ± 3.0
0.138 0.140 0.138 0.140 0.135
± 0.009 ± 0.002 ± 0.005 ± 0.004
75.2 74.4 75.5 74.3 75.0 74.8 75.1 74.2
± 3.0 ± 3.0 ± 3.0 ± 3.0 ± 3.0 ± 3.0 ± 3.0 ± 3.0
remark
[44], 100e1000 mm Hg [46], 587 Ke662 K [47] [48] This work This work This work This work This work This work This work This work
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 4 4 ( 2 0 1 9 ) 8 3 1 5 e8 3 2 4
Fig. 5 e TPD results.
8321
Fig. 6 e Equilibrium hydrogen pressure, peq, measured during desorption at 623 K, in dependence on desorption temperature Tpeak, measured by TPD using the rate 10 K/ min. Shaded area e values of peq calculated from literature data on enthalpy and entropy of hydride decay (Table 2).
Effect of studied additives upon HS parameters To assess the influence of studied X's upon the parameters that determine the HS behavior of studied alloys, it is effectual to combine results of kinetic (t09), PCT (peq) at 623 K and TPD (Tpeak) measurements. In Fig. 6, peq is plotted in dependence on Tpeak. The shaded area covers known literature values of peq at 623 K [44,46e48]. Unfortunately, no details on Tpeak are known to literature data on peq for Mg from Refs. [44,46e48]. However, our point peq ¼ 0.54 ± 0.05 MPa measured at 623 K in pure Mg at Tpeak ¼ 631 K (plotted in Fig. 6) may serve as a reference point to compare results measured with experimental alloys with pure Mg. It can be seen that our value of peq falls well into the interval of literature data. The alloys lying above the shaded region, show decrease in the stability of MgH2. It is obvious that alloys Si and NaCl show the most beneficial hydrogen storage performance: The lowest stability of hydride phase (i.e., the highest peq), which is lower than that by pure Mg, and, at the same time, the lowest hydrogen desorption temperature Tpeak. Results of kinetic study and TPD measurements are compared in Fig. 7. It is obvious that the alloys that slightly destabilize MgH2, led also to acceleration of hydrogen sorption. The worst additives with the longest time t09 on the other hand, were X ¼ MgmGan, and combinations Ni þ Mg2Si and Ni þ Mg17Al12. To assess the results showed in Figs. 6 and 7 in perspective of structure changes during A/D (see par. 3.2), it can be concluded that the best HS performance (high peq and, at the same time, low Tpeak and t09) show alloys, which phase composition seems to be invariant during hydrogen sorption (Ge, LiF, NaF, Si), and/or alloys that tend to form easily significant fraction of AM (NaCl, LiCl). Alloy Al is seemingly a special case in these scheme. However, if it is considered that X ¼ Mg17Al12 is,
in fact, mashed in ball mill into Mg and Al (see par. 3.2.), it is obvious that the sorption experiment was not conducted with alloy Al but with alloy Al’ with phase composition (Mg)-Al-CB that can be considered invariant during subsequent A/D cycling. This idea can explain that Al’ show acceptable HS performance (Figs. 6 and 7).
Fig. 7 e Hydrogen desorption rate at 623 K scaled by time t09 in dependence on peak temperature Tpeak measured by TPD.
8322
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 4 4 ( 2 0 1 9 ) 8 3 1 5 e8 3 2 4
Fig. 8 e Comparison of hydrogen sorption kinetic curves with literature (broken lines): Mg-Al-C-Nb2O5 [50]. Mg-Sb-C [16], Mg-M-Li [30], MgH2-VO2 [49]. (a) e A, (b) e D. Thick solid lines in (a, b) e NaF2, thick dotted lines in (b) e Si, Ge, NaCl).
Comparison with literature data Contrary to the fact that a great number of studies were devoted to the hydrogen sorption kinetics in Mg-based alloys, it is not easy to compare adequately the data sets. This is because of diverse test temperatures, various sample preparation and different experimental details. This may make difficult the comparison of desorption data. Comparison of desorption kinetics is strongly conditioned by hydrogen pressure p. The high value of p may considerably slow-down the hydrogen desorption in the vicinity of peq. Therefore, it is advisable to make comparison of desorption data sets at higher temperature, with p well below peq. In Fig. 8, the present sorption data are compared with literature [16,30,49,50] for the temperature 623 K, where the hydrogen pressure during desorption p 0.03 MPa was well bellow equilibrium pressure p y 0.7 MPa. All the shown literature data were obtained under similar conditions as the present results. For the sake of simplicity, the literature data are plotted in the form of broken lines going from origin and through knots in 0.9cHmax and in cHmax. Individual literature kinetic curves cH(t) in Fig. 8 refer to different content of C, Nb2O5 and different grain size [50], to different content of Sb [16], and to different concentration of added elements (Ag, Al, Zn, Y) [50], or VO2 [49] (for details on alloys labeling e see the original papers). It is obvious that the present data represent improvement in hydrogen sorption kinetics compared to pure Mg (see in Fig. 8a for A and in Fig. 8b for D [16]) and to other available literature results [16,49,50] irrespective of composition and other experimental parameters. Just to note: the admixture of transition metal oxides (VO2, Nb2O5) and compounds containing Li are considered to be the most efficient modifier of HS Mg-based alloys [3,5]. It is easy to find that the hydrogen
desorption kinetics observed in the present work is better than kinetics found in Mg(In)-MgF2 and in Mg85In5Al5Ti5 [18,19] and also much better than that reported in Ref. [10] for Mg alloys modified by yttrium.
Conclusion Several chosen additives X were prejudged as potentially good catalysts for hydrogen sorption kinetics in mixture Mg-X-C. The aim of the present work was to find a composition that would improve the HS kinetics of Mg. It was found that admixture of NaCl, Mg2Si, and LiF most effectively increase the hydrogen equilibrium pressure at 623 K (slightly destabilize MgH2) and decrease t09 and Tpeak. NaCl and LiF shift the peak temperature to the lowest temperature Tpeak ¼ 597 K. On the other hand, Mg modified by MgmGan and by combinations of Ni þ Mg17Al12 and Ni þ Mg2Si showed bad HS performance at temperature 623 K.
Acknowledgements This work was supported by project of Czech Science Foundation No. 17-21683S and by Central European Institute of Technology, CEITEC-Institute of Physics of Materials, AS CR Brno, Czech Republic.
Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.ijhydene.2019.02.084.
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 4 4 ( 2 0 1 9 ) 8 3 1 5 e8 3 2 4
references
[1] Huot J. Metal hydrides. In: Hirscher M, editor. Handbook of hydrogen storage. Weinheim: WILEY-VCH Verlag GmbH &Co. KGaA; 2010. p. 81e110. [2] Crivelo JC, Denis RV, Dornheim M, Felderhoff M, Grant DM, Huot J, Jensen TR, deJongh P, Latroche M, Walker GS, Webb CJ, Yartys VA. Mg-based compounds for hydrogen storage. Appl Phys 2016;85:122e39. [3] Khafidz NZK, Yaakob Z, Lim KL, Timmiati SN. The kinetics of lightweight solid-state hydrogen storage materials: a review. Int J Hydrog Energy 2016;41:13131e51. [4] Barthelemy H, Weber M, Barbier F. Hydrogen storage: recent improvements and industrial perspectives. Int J Hydrog Energy 2017;42:7254e62. [5] Ren J, Musyoka NM, Langmi HW, Mathe M, Liao S. Current research trends and perspectives on materials-based hydrogen storage solutions: a critical review. Int J Hydrog Energy 2017;42:289e311. [6] Zhang J, Yan S, Qu H. Recent progress in magnesium hydride modified through catalysis and nanoconfinement. Int J Hydrog Energy 2018;43:1545e65. [7] Wang H, Lin HJ, Cai WT, Ouyang LY, Zhu M. Tuning kinetics and thermodynamics of hydrogen storage in light metal element based systems e a review of recent progress. J Alloy Compd 2016;658:280e300. [8] Zhang X, Leng Z, Gao M, Hu J, Du F, Yao J, Pan H, Liu Y. Enhanced hydrogen storage properties of MgH2 catalyzed with carbon-supported nanocrystalline TiO2. J Power Sources 2018;398. 183-92. [9] Korablov D, Besenbacher F, Jensen TR. Kinetics and thermodynamics of hydrogenation-dehydrogenation for Mg25%TM (TM ¼ Ti, Nb or V) composites synthesized by reactive ball milling in hydrogen. Int J Hydrog Energy 2018;43:16804e14. [10] Yang T, Li Q, Liu N, Liang C, Yin F, Zhang Y. Improved hydrogen absorption and desorption kinetics of magnesiumbased alloy via addition of yttrium. J Power Sources 2018;378:636e45. [11] Yahya MS, Ismail M. Synergistic catalytic effect of SrTiO3 and Ni on the hydrogen storage properties of MgH2. Int J Hydrog Energy 2018;43:6244e55. [12] Yahya MS, Sulaiman NN, Mustafa NS, Yap FAH, Ismail M. Improvement of hydrogen storage properties in MgH2 catalysed by K2NbF7. Int J Hydrog Energy 2018;43:14532e40. [13] Xu E, Li H, You X, Bu C, Zhang L, Wang Q, Zhao Z. Influence of micro-amount O2 or N2 on the hydrogenation/ dehydrogenation kinetics of hydrogen-storage material MgH2. Int J Hydrog Energy 2017;42:8057e62. [14] Floriano R, Deledda S, Hauback BC, Leiva DR, Botta WJ. Iron and niobium based additives in magnesium hydride: microstructure and hydrogen storage properties. Int J Hydrog Energy 2017;42:6810e9. [15] Xie L, Li J, Zhang T, Song L, Kou H. Microstructure and hydrogen storage properties of Mg-Ni-Ce alloys with a longperiod stacking ordered phase. J Power Sources 2017;338:91e102. [16] Cermak J, Kral L, Roupcova P. Improvement of hydrogen storage kinetics in ball-milled magnesium doped with antimony. Int J Hydrog Energy 2017;42:6144e51. [17] Korablov D, Besenbacher F, Jensen TR. Ternary compounds in the magnesium-titanium hydrogen storage system. Int J Hydrog Energy 2014;39:9700e8. [18] Ouyang LZ, Cao ZJ, Wang H, Liu JW, Sun DL, Zhang QA, Zhu M. Enhanced dehydriding thermodynamics and kinetics in Mg(In)eMgF2 composite directly synthesized by plasma milling. J Alloy Compd 2014;586:113e7.
8323
[19] Cao Z, Ouyang L, Wu Y, Wang H, Liu J, Fang F, Sun D, Zhang Q, Zhu M. Dual-tuning effects of In, Al, and Ti on the thermodynamics and kinetics of Mg85In5Al5Ti5 alloy synthesized by plasma milling. J Alloy Compd 2015;623:354e8. [20] Valentoni A, Mulas G, Enzo S, Garroni S. Remarkable hydrogen storage properties of MgH2 doped with VNbO5. Phys Chem Chem Phys 2018;20:4100e8. n M, Al-Ajmi F. Discovering a new [21] El-Eskandarany MS, Banya MgH2 metastable phase. RSC Adv 2018;8:15667e74. [22] Pukazhselvan D, Nasani N, Correia P, Carbo-Argibay E, OteroIrurueta G, Stroppa DG, Fagg DP. Evolution of reduced Ti containing phase(s) in MgH2/TiO2 system and its effect on the hydrogen storage behavior of MgH2. J Power Sources 2017;362:174e83. [23] Yahya MS, Ismail M. Catalytic effect of SrTiO3 on the hydrogen storage behaviour of MgH2. J Energy Chemistry 2019;28:46e53. https://doi.org/10.1016/j.jechem.2017.10.020. [24] Ismail M, Mustafa NS, Juahir N, Yap FAH. Catalytic effect of CeCl3 on the hydrogen storage properties of MgH2. Mater Chem Phys 2016;170:77e82. [25] Sulaiman NN, Juahir N, Mustafa NS, Yap FAH, Ismail M. Improved hydrogen storage properties of MgH2 catalyzed with K2NiF6. J Energy Chemistry 2016;25:832e9. [26] Yahya MS, Ismail M. Improvement of hydrogen storage properties of MgH2 catalyzed by K2NbF7 and multiwall carbon nanotube. J Phys Chem C 2018;122:11222e33. [27] Li Q, Lin Q, Chou K-C, Jiang L. A study on the hydridingdehydriding kinetics of Mg1.9Al0.1Ni. J Mater Sci 2001;39:61e5. [28] Yhu Y, Liu Z, Yang Y, Gu H, Li L, Cai M. Hydrogen storage properties of MgeNieC system hydrogen storage materials prepared by hydriding combustion synthesis and mechanical milling. Int J Hydrog Energy 2010;35:6350e5. [29] Cermak J, Kral L. Alloying of Mg/Mg2Ni eutectic by chosen non-hydride forming elements: relation between segregation of the third element and hydride storage capacity. J Power Sources 2012;197:116e20. [30] Suarez-Alcantara K, Palacios-Lazcano AF, Funatsu T, Cabanas-Moreno JG. Mg-M-LiH alloys prepared by mechanical milling and their hydrogen storage characteristics. Int J Hydrog energy 2015;40:17344e53. [31] Yap FAH, Yahya MS, Ismail M. Enhancement of hydrogen storage properties in 4MgH2-Na3AlH6 composite catalyzed by TiF3. Int J Hydrog Energy 2017;42:21096e104. [32] Zhang H, Fu L, Xuan W, Qin H, Ji Z. Ni-Al films induced surface modification of La2Mg17 alloy leading to improved dehydrogenation properties. J Power Sources 2018;385:27e31. [33] Cermak J, Kral L. Hydrogen diffusion in MgeH and MgeNieH alloys. Acta Mater 2008;56:2677e86. [34] Ding X, Chen R, Jin Y, Chen X, Guo J, Su Y, Ding H, Fu H. Activation mechanism and dehydrogenation behavior in bulk hypo/hypereutectic Mg-Ni alloy. J Power Sources 2018;374:158e65. [35] Vajo JJ, Mertens F, Ahn CC, Bowman jr RC, Fultz B. Altering hydrogen storage properties by hydride destabilization through alloy formation: LiH and MgH2 destabilized with Si. J Phys Chem B 2004;108:13977e83. [36] Liu H, Wu C, Zhou H, Chen T, Liu Y, Wang X, Dong Z, Ge H, Li S, Yan M. Synergistically thermodynamic and kinetic tailoring of the hydrogen desorption properties of MgH2 by co-addition of AlH3 and CeF3. RSC Adv 2015;5:22091e6. [37] Lyci G, Surendra KS. Structural stability of metal hydrides, alanates and borohydrates of alkali and alkali-earth elements: a review. Int J Hydrog Energy 2010;35:5454e70. [38] Zhang B, Wu Y. Microstructural evolution and improved hydrogen storage properties for the Li3N-MgH2 system by addition of LiNH2 during the hydrogenation/ dehydrogenation. Int J Hydrog Energy 2015;40:9298e305.
8324
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 4 4 ( 2 0 1 9 ) 8 3 1 5 e8 3 2 4
[39] Oumellal Y, Courty M, Rougier A, Nazri GA, Aymard L. Electrochemical reactivity of magnesium hydride toward lithium: new synthesis route of nano-particles suitable for hydrogen storage. Int J Hydrog Energy 2014;39:5852e7. [40] Sulaiman NN, Ismail M. Effects of TiF3 addition on the hydrogen storage properties of 4MgH2 þ Cd composite. Int J Hydrog Energy 2018:1e9. accepted, https://doi.org/10.1016/j. ijhydene.2018.03.095. [41] Cermak J, Kral L. Hydrogen absorption in carbon allotropes and talc. Kovove Mater Metallic Mater 2018;56:75e80. [42] Zhou S, Zhang X, Li T, Wang N, Chen H, Zhang T, Yu H, Niu H, Liu D. Nano-confined magnesium for hydrogen storage from reactive milling with anthracite carbon as milling aid. Int J Hydrog Energy 2014;39:13628e33. [43] Vejeeston P, Ravindran P, Hauback BC, Fjellvag H, Kjekshus A, Furuseth S, Hanfland M. Phys Rev B 2006;73:224102. [44] Kennelley JA, Varwig JW, Myers HW. Magnesium-hydrogen relationships. J Phys Chem 1960;64:703e4. [45] Dornheim M. Tailoring reaction enthalpies of hydrides. In: Hirscher M, editor. Handbook of hydrogen storage.
[46] [47] [48]
[49]
[50]
Weinheim: WILEY-VCH Verlag GmbH &Co. KGaA; 2010. p. 187e214. Stampfer JF, Holley CE, Suttle FJ. The magnesium-hydrogen system. J Am Chem Soc 1960;82:3504e8. Schwarz RB. Hydrogen storage in Magnesium-based alloys. MRS Bull 1999;11:40e4. Gerard N, Ono S. Hydride formation and decomposition kinetics. In: Schlapbach L, editor. Hydrogen in intermetallic compounds II. Berlin, Heidelberg, New York, London, Paris, Tokyo, Hong Kong, Barcelona, Budapest: Springer-Verlag; 1992. p. 165e96. Milosevic S, Kurko S, Pasquini L, Matovic L, Vujasin R, Novakovic N, Grbovic Novakovic J. Fast hydrogen sorption from MgH2-VO2(B) composite materials. J Power Sources 2016;307:481e8. Hardian R, Pistida C, Chaudhary A-L, Capurso G, Gizer G, Cao H, Milanese C, Girella A, Santoru A, Yigit D, Dieringa H, Kainer KU, Klassen T, Dornheim M. Waste Mg-Al based alloys for hydrogen storage. Int J Hydrog Energy 2018;43:16738e48.