Effect of mechanical milling on both the structure and ...

2 downloads 0 Views 440KB Size Report
Jun 16, 2006 - curves were obtained using a modified Sievert's type apparatus ... chamber at different milling times inside a glove box and samples were ...
WHEC 16 / 13-16 June 2006 – Lyon France

Effect of mechanical milling on both the structure and the first hydriding-dehydriding properties of a MmNi5-Ni mixture Marcelo R. Esquivel a,b and Gabriel Meyer a,b a

Comisión Nacional de Energía Atómica-Centro Atómico Bariloch, Av. Bustillo km 9.5, Bariloche, Río Negro, Argentina, [email protected] b Consejo Nacional de Investigaciones Científicas y Técnicas, [email protected]

ABSTRACT: The effects of mechanical milling on both the structure and hydriding properties of a MmNi5-Ni mixture synthesized by mechanical alloying is presented. Mm and Ni were milled in Ar atmosphere to obtain the mixture. Alloy was heated at 600 ºC during 5 days to obtain a crystalline compound which served as reference. During the milling process, samples were withdrawn and studied by X-ray diffraction (XRD) and scanning electron microscopy (SEM). XRD measurements were used to determine the change of the crystallite size and strain on both MmNi5 and Ni peaks due to milling. SEM was used to observe the temporal evolution in morphology and size of the particles. From this study, the stages occurring during the grinding of the alloy were identified and characterized. Selected samples were analyzed and hydrogen absorption curves were obtained using a modified Sievert’s type apparatus designed in our laboratory. The data obtained from absorption curves was correlated with the defects and strain introduced into the material due to milling. KEYWORDS : MmNi5, Mechanical alloying, Mischmetal.

INTRODUCTION: Mechanical alloying has evolved gradually as a widely used method of synthesis [1,2]. Simplicity, low cost and easy scaling up are the main advantages attributed to this process over traditional ones such as full equilibrium methods or chemical synthesis. The findings obtained using this technique contributed to the development of alloys useful for diverse applications [3,4]. This current trend is also observed in the synthesis of MmNi5 based alloys [5,6]. Most of these works have been done using a high energy mill device [1]. Compounds were synthesized in short times of the order of minutes or hours and no further investigation was focused on the characteristic stages occurring during milling [5,6]. Features related to the evolution of the sample during this process can be analyzed if synthesis is achieved in a low energy mill [7]. One of these characteristics is the relationship between integrated milling time and the introduction of strain and defects into the material [1,2]. MmNi5 based alloys have been synthesized from individual compounds using this technique [5,6,7]. A work regarding the evolution of mechanical milling process of a MmNi5-Ni mixture obtained from Mm and Ni powders using a low energy mill was published previously [7]. In this work, the temporal evolution of the milling process of a MmNi5-Ni mixture is presented. Initial mixture is obtained by a combination of mechanical milling with low heating temperature [7]. The powder morphology is analyzed by scanning electron microscopy (SEM). From this analysis, the stages governing the process of milling are obtained. Results are compared to those obtained for the milling of a Mm-Ni mixture [7]. The crystallite size and strain of both Mm and Ni are studied using X-ray diffraction (XRD). Values obtained at intermediate milling times are compared to those obtained previously for Ni [7]. Hydrogen-alloy Interaction of samples extracted at different times is studied using a Sievert´s type device. From these results, the effect of milling time on both microstructure and hydrogen absorption process is analyzed. These results are also compared to those obtained from a sample synthesized previously [7]. The potential applications of these findings in both research and technological fields makes worth this study. This double objective aimed the elaboration of the present work. EXPERIMENTAL: 1/7

WHEC 16 / 13-16 June 2006 – Lyon France

Pure Ni (3.80 µm, 99.99%) (Sigma Aldrich) and drilled lumps of Mischmetal (99.7 %) (Alpha Aesar) of nominal composition 52.0 wt% Ce, 25.6 wt% La, 16.9 wt% Pr , 5.5 wt% Nd were mechanically milled under Argon atmosphere using a Uni-Ball-Mill II apparatus (Australian Scientific Instruments). Mischmetal (Mm) composition was verified by Neutron Activation Analysis (NAA) and Energy Dispersive Spectroscopy (EDS). Powder mixture in a proportion of 20% excess of Nickel over the stoichiometric MmNi5 composition and steel balls were set in a stainless steel chamber under Ar atmosphere in a glove box. Particles size and morphology were observed by Scanning Electron Microscopy (SEM). The balls to powder mass relation was 33.5/1. Representative amount of powder was withdrawn from chamber at different milling times inside a glove box and samples were analyzed by X-ray powder diffraction (XRD). During sample manipulation inside the glove box the oxygen level was monitored by a trace analyzer (Series 3000, Alpha Omega) and kept under 5 ppm to avoid material oxidation. Room temperature X-ray diffraction was achieved on a Philips PW 1710/01 Instrument with Cu Kα radiation (graphite monocromator). Diffraction patterns were analyzed by the Rietveld method [8] using DBWS software [9]. A Sievert´s type equipment was used to measure hydrogen absorption-desorption curves. The sample is placed in the reactor at constant temperature and a selected initial pressure. A PC-based data acquisition system monitors and controls the experiment variables. Experimental set-up device details can be found elsewhere [10]. RESULTS AND DISCUSSION: Milling evolution of a MmNi5-Ni mixture The diffractograms of samples extracted at different milling times are shown in Fig. 1. Initial sample (Fig. 1.a) is a MmNi5-Ni mixture obtained by mechanical milling and heated at 600 ºC during 5 days to obtain complete crystallization.

Figure 1. Diffractograms of a MmNi5-Ni mixture milled at different times. a) Initial. b) 10 h; c) 20 h ; d) 40 h; e) 100 h. Ni diffraction lines are shown in dotted lines.

2/7

WHEC 16 / 13-16 June 2006 – Lyon France MmNi5 and Ni crystalline parameters and mass percentage in initial mixture was determined using Rietveld Method [9]. Mass percentages, Rietveld method fitting parameters at different milling times for these and samples milled at other different times are summarized in Table 1. Table 1. MmNi5-Ni mixture parameters and mass percentage values. Milling Time 0h 4h 5h 10 h 20 h 40 h 60 h

MmNi5 Parameters (Å) Volume cell (Å)3 a c 4,9067 3,9774 82, 93 4,9171 3,9860 83,46 4,9114 3,9815 83,17 4,9170 3,9898 83,54 4,9072 3,9937 83,28 4,8930 4,0001 82,96 4,8882 4,0001 82,83

Ni Parameters (Å) a 3,5304 3,5396 3,5367 3,5400 3,5371 3,5330 3,5242

Percentage in Mixture (mass %) MmNi5 Ni 94 ± 1 6±1 94 ± 4 6±2 94 ± 2 6±2 94 ± 1 6±1 92 ± 2 8±2 92 ± 2 8±2 94 ± 1 6±2

Rwp value 20 18 15 18 18 18 20

MmNi5 structure was fitted assigning P6/mmm space group and Wyckoff positions corresponding to LaNi5 [11]. Lanthanides (Ce, La, Nd, Pr) in MmNi5 alloy are supposed to be distributed randomly in La positions in the structure. Ni parameters were fitted assigning space group and Wyckoff positions of reference pattern [12]. It is clearly observed that milling increments the peaks width (Figs. 1b to 1e) for both Ni and MmNi5. Ni diffraction lines are indicated in doted lines. Other diffraction lines correspond to MmNi5. No secondary phases are detected. The increase of peaks width can be assigned to a decrease in the crystallite size, to a decrease in the long range order or to an increase of the strain [13]. From Table 1, it is noticed that MmNi5 basal parameter increases with milling time from 0 h to 4h a relative length of 0.21% and reduces from 5 h to 60 h a relative length of 0.47%. This reduction reaches lower values than that of initial sample. MmNi5 c parameter increases with milling a relative length 0.57%. Volume cell value increases up to 10 h and decreases up to 60 h finally reaching approximately the same value before and after milling (variation of 0.12%). Despite the differences in mill devices, the behavior of both a and cell volume parameters is similar to that reported for the milling of LaNi5 at times shorter than 5 h [14]. A higher energy mill was used in reference [14]. c parameter behavior increases from 0 to 4h being different from c behavior observed in [14] where the values remain almost constant. Cell volume and a parameter changes are opposite to those observed in other reference [15]. The expansion observed in c parameter in this work should not be attributed to the formation of dumbbells [14,15]. Because an a and cell volume reduction should also be observed. A site interchange between lanthanide and Ni atoms can not explain the process either [15]. Because the mill used here is a device with lower energy than those of ref [14,15]. The increments in lattice parameters (a, c, cell volume) observed from 0 to 4h in this work could be due to the defects accumulated during the first hours of mechanical milling. The effects are isotropic because broadening of peaks is symmetrical and no preferential orientation is observed as shown in Figure 2. The figure shows a reduced range of the diffractogram of samples un milled (full line), milled 2 h (dotted line) and milled 4 h (dashed line). The displacement of the maxima of each MmNi5 peak is in agreement with data of Table 1 showing an increment in a, c and volume cell parameters. Ni (111) peaks also shows a maxima displacement indicating an increase of a parameter. The further decrease on lattice parameter a and further increase on lattice parameter c at times longer than 4 h can not be attributed to the creation of Ni dumbbells replacing a La atoms in a La site [15] since volume cell values remains almost constant. A preferential reordering of higher radius lanthanides along the C axis and lower atom radius lanthanides in the basal parameter could be the explanation to the cell parameters evolution. A detailed study should be done to clarify this point. Ni a parameter increases with milling from 0 h to 10 h. It decreases from 10 h to 60 h. Its maximum relative change reaches 0,27% at 10 h of milling. Effects of milling on Ni can not be appreciated at milling times longer than 10 h because of Ni diffraction lines disappearance. Crystallite size and strain induced by milling Table 2 shows the change in crystallite size and strain induced by milling at different milling times. As observed, crystallite size decreases with milling and strain increases.

3/7

WHEC 16 / 13-16 June 2006 – Lyon France MmNi5 diffraction lines (011), (110) and (111) were selected to analyze both changes in strain and size of crystallites. (111) Ni diffraction line was selected. In each case selection obeyed to its higher relative intensity at all milling times. As observed in Table 2, MmNi5 decreases with milling time. Anisotropy is observed in milling because relative changes within the initial and final milling time are different for (011) and (110) and (111). These structural changes were introduced by milling. Strain changes are also observed in Table 2. It is deduced that milling introduces changes in the structure. It is also noted the anisotropy of the phenomena.

Figure 2. MmNi5 peaks displacement and symmetry at different milling times. Table 2. Changes in crystallite size and strain in MmNi5 and Ni structures Milling time (h) 0h 4h 5h 10 h 20 h 40 h 60 h 100 h

Crystallite size (Å) MmNi5 (0,1,1) (1,1,0) (1,1,1) 421 727 508 439 466 491 391 448 491 298 355 329 180 236 134 159 214 128 113 200 94 127 184 79

Ni (1,1,1) 270 --------------------------------------------------------

(0,1,1) 0.048 0.067 0.076 0.098 0.160 0.185 0.200 0.232

Strain (%) MmNi5 (1,1,0) 0.026 0.054 0.055 0.068 0.100 0.115 0.123 0.134

(1,1,1) --------0.044 0.044 0.064 0.150 0.165 0.223 0.223

Stages present during milling Figure 3 shows a mosaic image of samples milled at different times. The particles morphology as obtained by SEM is useful to determine the stages present during milling. Unlike the stages present during the mechanical alloying of Mm-Ni powders to form MmNi5 [7], the mechanical grinding of the MmNi5-Ni mixture leads to a global increment on the size of the particles. Since particles of an average size of 5 µm of un milled powders (Fig 3.a) transforms to agglomerates of sizes bigger than 20 µm (Fig 3 h). Since initial milled powders are previously milled and heated powders, initial size is low enough not to allow fracture 4/7

WHEC 16 / 13-16 June 2006 – Lyon France process to progress. On the contrary cold welding slowly dominates and global particle size is increased (see Figures f -h). Competition between both processes is found at intermediate milling times (see figures de).

Figure 3. SEM images of samples extracted at different milling times. a) Un milled powders. b) 2 h. c) 3 h. d) 10 h. e) 30 h. f) 60 h g) 100 h. h) 200 h. Absorption curves Figure 4 shows the first absorption curves for different as-milled non activated MmNi5-Ni mixtures. Y axis represents the weight percent ( mH / (mMmNi5 + mH)) of absorbed H2. The initial pressure is 6000 kPa

Figure 4. a) Absorption curves at 25 ºC. b) Absorption curves at 90 ºC. Initial Pressure is 6000 kPa 5/7

WHEC 16 / 13-16 June 2006 – Lyon France

First hydriding curves of non activated samples were used instead of hydrogen cycled samples. This procedure was preferred to analyze the effects of mechanical milling on microstructure than the hydrogen interaction in activated samples after many hydrogen absorption-desorption cycles. As observed, the wt% at the same absorption time is higher as milling time increases. This is a straight forward effect of the effect of milling time. The introduction of defects and deformation on microstructure and surface produces a higher reactivity. The introduction of bulk defects also contributes to a quicker absorption. As temperature increases the effect of milling is decreased. This can be attributed to an enhanced superficial reactivity and diffusion of H2 due to the combined effects of temperature and strain introduced by mechanical milling. CONCLUSIONS: In this work, the effect of mechanical milling on microstructure and further hydrogen absorption of a MmNi5-Ni mixture was analyzed. MmNi5-Ni mass ratio was determined by Rietveld analysis. Changes in MmNi5 cell parameters was analyzed and compared to bibliography. The evolution of cell parameters a and c was attributed to preferential distribution of lanthanides in La position. The effect of milling on crystallite size and structure strain was analyzed by XRD. Strain anisotropy was observed in MmNi5. Particles size during milling evolution was analyzed at short and long times by SEM observations. Three main stages were observed. Fracture predominates in the first stage at times shorter than 10 h. Equilibrium between fracture and cold welding is observed between 10 h and 30 h. At higher times, cold welding predominates and a marked increase in particles size is observed. Hydrogen absorption curves of samples milled at different integrated milling times were analyzed at 25 ºC and 90 ºC. Higher milling times were found to correlate with a higher and quicker hydrogen absorption. A further analysis of samples treated at intermediate temperatures and intermediate integrated milling times is the subject of an incoming work.

REFERENCES: [1] C. Suryanarayana, Mechanical alloying and milling, Progress in Materials Science, 46, 9-29, Pergamon, 2001. [2] Lü, M. O. Lai, Mechanical alloying, 11-21, Kluwer Academic Publishers, Boston, 1998 [3] Z. Dehouche, N. Grimard, F. Laurencelle, J. Goyette, T.K. Bose, Hydride alloys properties investigations for hydrogen sorption compressor, J. Alloys Compd., 399, 224-236, 2005. [4] P. Muthukumar, M. Prakash Maiya, S. Srinivasa Murthy, Experiments on a metal hydride-based hydride storage device, Int. J. Hydrogen Energy, 30, 1569-1581, 2005. [5] M. Jurczyk, W. Rajewski, W. Majchrzycki, G. Wójcik; Mechanically alloyed MmNi5-type materials for metal hydride electrodes, J. Alloys Compd., 290, 262-266, 1999. [6] M. Jurcyk, Hydrogen storage properties of amorphous and nanocrystalline MmNi4.2Al0.8 alloys, J. Alloys Compd., 307, 279-282, 2000. [7] M.R. Esquivel, J.J. Andrade Gamboa, F.C. Gennari and G. Meyer, Synthesis of MmNi5 by combined mechanical milling-low temperature heating process, Proceedings of the Sohn International Symposium on Advanced Processing of Metals and Materials: Principles, Technologies and Industrial Practice, In press. [8] R.A. Young, ed., The Rietveld Method, (International Union of Crystallography, Oxford University Press, 1993). [9] “DBWS, 9411 an upgrade of the DBWS programs for Rietveld refinement with PC and mainframe computers”, J. Appl. Cryst,. 28, (1995) , 366 -367. [10] G. Meyer, D. Rodriguez, F. Castro, G. Fernández, Automatic device for precise characterization of hydride-forming materials, Proceedings of the 11 th World Hydrogen Energy Conference, Stuttgart, Hydrogen Energy Progress, Vol XI, 1996, p 1293-1298. [11] P. Fischer, A. Furrer, G. Bush, L. Schalapbach: Neutron scattering investigations of the LaNi5 hydrogen storage system, Helvetica Phisica Acta, 1977, vol. 50, pp 421-430. [12] JCPDF, Powder Diffraction File, International Center for Diffraction Data, Swarthmore, PA, 1996, Card No 040850. [13] S. Enzo, E. Bonetti, I. Soletta, G. Cocco, Structural changes induced by the mechanical alloying of crystalline metal powders, J. Phys D: Appl. Phys., 1991, vol. 24, pp 209-216.

6/7

WHEC 16 / 13-16 June 2006 – Lyon France [14] S. Corré, M Bououdina, N. Kuriyama, D. Fruchart, G. Adachi, Effects of mechanical grinding on the hydrogen storage and electrochemical properties of LaNi5, J. Alloys Compd., 1999, vol. 292, pp 166-173. [15] G. Liang, J. Huot, R. Schultz, Hydrogen storage properties of the mechanically alloyed LaNi5-based materials, J. Alloys Comp., 2001, 320, pp 133-139.

ACKNOWLEDGEMENTS: The authors wish to thank Comisión Nacional de Energía Atómica of Argentina (Project CNEA P5-PID-95-2) , Secretaría de Ciecia, Técnica y Posgrado Universidad Nacional de Cuyo of Argentina, Consejo Nacional de Investigaciones Científicas y Técnicas of Argentina (Project PIP-6448) and Agencia Nacional de Promoción Científica y Tecnológica of Argentina (Project PICT 12-15065) for partial finantial support.

7/7