- Email: [email protected]
Measurement and augmentation of effective thermal conductivity of La0.8Ce0.2Ni5 hydride bed
Accepted Manuscript Measurement and augmentation of effective thermal conductivity of La0.8Ce0.2Ni5 hydride bed Yogesh Madaria, E. Anil Kumar PII:
S0925-8388(16)32683-4
DOI:
10.1016/j.jallcom.2016.08.283
Reference:
JALCOM 38785
To appear in:
Journal of Alloys and Compounds
Received Date: 22 June 2016 Revised Date:
26 August 2016
Accepted Date: 27 August 2016
Please cite this article as: Y. Madaria, E. Anil Kumar, Measurement and augmentation of effective thermal conductivity of La0.8Ce0.2Ni5 hydride bed, Journal of Alloys and Compounds (2016), doi: 10.1016/j.jallcom.2016.08.283. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
Measurement and Augmentation of Effective Thermal Conductivity of La0.8Ce0.2Ni5 hydride bed Yogesh Madaria and E. Anil Kumar1
RI PT
Discipline of Mechanical Engineering, Indian Institute of Technology Indore, Indore, Madhya Pradesh, India 453552 ABSTRACT
An experimental set up for measuring Effective Thermal Conductivity (ETC) of Metal Hydride
SC
(MH) beds was fabricated using one-dimensional steady-state radial heat transfer absolute method. La0.8Ce0.2Ni5 alloy powder and its hydride were considered in the present study. Two types of compacts were fabricated for ETC augmentation of MH bed. First type was made from
M AN U
the mixture of 6 wt% graphite flakes with the La0.8Ce0.2Ni5 alloy powder. A three dimensional structure, having six radial fins, uniformly distributed around the outer periphery of a hollow cylinder, made with copper wire mesh, was integrated with the mixture of La0.8Ce0.2Ni5 powder and 6 wt% graphite flakes to make the second type of compacts. The ETC of the three types of beds namely loose metal powder bed (LMP), bed of compacts of metal alloy powder mixed with
TE D
graphite flakes (CMPGF) and bed of compacts of metal alloy powder mixed with graphite flakes and embedded with copper wire mesh structure (CMPGFCu) were measured and compared at different argon and hydrogen pressures, hydrogen concentrations and average bed temperatures. The maximum values of effective thermal conductivity for the bed configurations of
EP
LMP/CMPGF/CMPGFCu were 1.3/4.7/6.8 Wm-1K-1. These values of effective thermal conductivity were obtained at 70 bar hydrogen gas pressure and 60˚C average bed temperature. The decrease in mass fraction of hydrogen absorbed by the La0.8Ce0.2Ni5 hydride, as a result of
AC C
implementation of ETC augmentation techniques, was nominal (around 5.7% in case of CMPGF and 4.5% in case of CMPGFCu) when compared to the respective powder bed configurations. Keywords: Metal hydride bed, Effective thermal conductivity, Compacts, Steady state method.
1
Corresponding author: E-mail: [email protected]
1
ACCEPTED MANUSCRIPT
INTRODUCTION Certain metals, alloys or intermetallics can reversibly absorb and desorb hydrogen. By absorption of hydrogen these materials form a class of compounds called metal hydrides. The
RI PT
hydrogen absorption is an exothermic process and desorption is an endothermic process. By virtue of these reactions metal hydrides have been used in some important applications involving heat exchange such as sorption based cooling systems, hydrogen purifiers, hydrogen compressors, etc [1] other than hydrogen storage. For any of these applications hydrogen transfer (absorption/desorption) rate is very important, as it decides the cycle time. Also hydrogen
SC
absorption and desorption rates are heat and mass transfer dependent [2]. The prevailing conditions in metal hydride beds show that almost all the heat transfer through the metal hydride
important
role
in
heat
M AN U
bed is by conduction mode. Thus effective thermal conductivity (ETC) of the bed plays an transfer
from/to
the
metal
hydride
bed.
Cyclic
hydrogenation/dehydrogenation results in reduction in particle size of the metal hydride leading to final average size of particles in the range of 10-15 microns [3]. ETC varies with hydrogen concentration, particle size and bed porosity. Owing to these dependencies, it becomes difficult to measure the effective thermal conductivity of metal hydride beds. Several measurement
TE D
techniques have been developed and utilized for determination of ETC of MH beds. Sun and Deng [4], Kempf and Martin [5], Suda et al. [6] etc. applied steady state methods while Dedrick et al. [7], Ishido et al. [8], Gustafsson [9] etc. have opted for transient methods for measurement of ETC of metal hydride beds. Kapischke and Hapke [10] had proposed oscillation heating
EP
instead of continuous heating for ETC measurement. Also different simulation methods have been applied for estimation of ETC of MH beds. Sun and Deng [3] suggested a mathematical
AC C
model for estimation of effective thermal conductivity of MlNi4.5Mn0.5 (Ml: lanthanum rich mischmetal). The heat transfer was considered to be taking place through three mechanisms in parallel to each other, those were (i) conduction and radiation through voids, (ii) through a series path comprising of effective solid path and effective gas path and (iii) through diffusion between solid particles. The summation of the three fractional heat transfers gives total heat transfer which is then used to calculate the ETC of bed. Asakuma et al. [11] applied a homogenization method for calculation of ETC of LaNi4.7Al0.3Hx bed. The change in contact area due to hydrogenation/dehydrogenation was examined and the effect of particle size depreciation on the 2
ACCEPTED MANUSCRIPT
ETC of MH bed was accommodated. Authors’ group has proposed appropriate extensions to some renowned mathematical models, originally developed for non reactive packed beds, so as to make those models suitable for estimation of ETC of MH beds [12]. The extended models were validated with the experimental results from the work of E. Anil Kumar et al. [13].
RI PT
Effective thermal conductivity of metal hydride beds is generally in the range of 0.1-1.2 W m-1 K-1[4-10]. This leads to poor heat transfer characteristics of hydride beds. For better hydrogen sorption characteristics and more importantly for efficient working of metal hydride based heat exchanging systems the value of ETC must be improved to a value suitable for applications like
SC
metal hydride based hydrogen compressor, metal hydride based thermal energy storage system, etc. (target range for present work was 3-7 W m-1 K-1). Several augmentation techniques have
M AN U
been tried with different levels of successes. Extensive review on experimental methods, simulation techniques and augmentation of ETC of MH beds has been done by the authors [1]. Some of the common augmentation techniques are: inclusion of high thermal conductivity metallic foams/metallic structures/fins, mixing of graphite fibers, metallic coating of metal hydride particles and compaction etc. A rather unconventional technique was that of direct synthesis of single walled carbon nano tube (SWCNT) on MH particles which resulted in drastic
TE D
enhancement in ETC of MH bed. An ETC value of 16 W m-1 K-1 was claimed to be achievable with SWCNT filling ratio of 0.5 and a slightly reduced particle packing ratio although no experimental evidences are available for the success of this technique [14]. ETC augmentation techniques like insertion of metal foams, porous metal matrix compaction, microencapsulated
EP
compacts, etc. results in increase of parasitic thermal mass of the reactor bed therefore compromising with performance of MH beds [15-21]. As compared to other augmentation
AC C
techniques, compaction has grabbed more attention and has proven to be more advantageous [1]. Yan et al. [22] and Sanchez et al. [23] observed that the ETC does not depends only on expanded natural graphite to metal hydride (ENG/MH) mass ratio but also on the fill factor and porosity and therefore on the pressing conditions. The ETC can be enhanced by intensified pressing; however this might have a negative influence on the hydrogen flow through the compact. Yan et al. [22] suggested that ENG should be added as low as possible in meeting the heat transfer requirements for the endothermic desorption reaction. In their study with Mg90Ni10-ENG compacts Pohlmann et al. [24-25] found that with the increase in ENG content the Mg90Ni10 3
ACCEPTED MANUSCRIPT
flakes got more misaligned, probably due to possibility of more free movement within the lubricious ENG matrix. For better alignment of Mg90Ni10 flakes in radial direction, 5 wt% ENG content was found to be suitable. While Kim et al. [21] reported that the ETC for compacts of LaNi5 and recompressed expanded graphite (REG) increased from 0.1 W m-1 K-1 for 0 wt% REG
RI PT
to 4.5 W m-1 K-1 for 10 wt% REG and 6.3 W m-1K-1 for 100 wt% REG and Pohlmann et al. [26] found a linear increase in ETC with ENG content, Shim et al. [27] observed that the ETC value for compacts of MgH2 and ENG, increased slowly with ENG content up to 10 wt% but the rate
SC
of increase of ETC with ENG content was much sharper from 10 wt% to 20 wt%.
In the present work authors have attempted to exploit the advantages of compaction, mixing of graphite flakes and inclusion of high thermal conductivity metal structure for the ETC
M AN U
augmentation of La0.8Ce0.2Ni5 hydride bed. La0.8Ce0.2Ni5 hydride was chosen as La based intermetallic hydrides are suitable for developing metal hydride based hydrogen compressors, cooling systems etc. [2, 28-30]. Two types of compacts were formed with the view to serve different purposes: (i) To assess the level of augmentation achieved by compaction and mixing of graphite flakes, first type of compacts were made by pressing a mixture of La0.8Ce0.2Ni5 powder and 6 wt% graphite flakes. (ii) Keeping in mind the adverse effects of excessive amount
TE D
of graphite flakes and those of excessive pressing on hydrogen sorption rates to/from the metal hydride, in the second type of compact, inclusion of copper wire mesh structure was tried as an alternative to both increasing the quantity of graphite flakes and increasing compaction pressure.
EP
The ETC augmentations with the two types of compacts have been compared. Experimental Details
AC C
In this work one dimensional, steady state, radial heat transfer, absolute method was used for ETC measurement. Temperature profile prevailing in the MH bed is measured with ‘K’ type thermocouples with probe of 1 mm diameter. The steady state method is time consuming and some loss in accuracy is seen due to non-fulfillment of assumption of homogeneous beds caused by the presence of thermocouples within the bed [31]. In the present case, the thermocouple probes (each of 1 mm diameter and 2.04 cm2 surface area) have occupied only 0.2% of bed volume. Therefore the problem of loss in homogeneity was negligible. Taking this into 4
ACCEPTED MANUSCRIPT
consideration, steady state method becomes more accurate and reliable than transient methods of ETC measurement and hence was chosen for the present study. ETC cell
RI PT
Fig. 1 shows the schematic diagram of ETC cell. The ETC cell is made up of a SS 316 cylinder, which was designed and fabricated to attain the required conditions of one dimensional, steady state, radial heat transfer. For achieving unidirectional radial heat transfer the aspect ratio (Length/Diameter) of the reactor was kept 4.33:1. To ascertain steady-state heat transfer two
SC
constant temperatures (as heat source and heat sink) were maintained at the inner and outer radii of the hollow cylindrical MH bed. The MH bed in powder form (1) assumed a hollow cylindrical shape with a capsule heater (2) of 10 mm diameter, 130 mm length and 60 W rated capacity, at
M AN U
the inner radius and enveloped by the inner surface of high pressure cylinder (3) of 30 mm inner diameter, 38 mm outer diameter and 170 mm height, at the outer radius. Coolant jacket (4) has an annular volume between the outer surface of high pressure cylinder and inner wall of another cylinder of 45 mm inner diameter, 49 mm outer diameter and 170 mm length. Suitable fittings were provided to this cylinder for coolant inlet (5) and outlet (6). Two flanges, one at each end of the cylinders were welded to make this a single assembly. The top flange (7) was made as a
TE D
hollow disc of 30 mm inner diameter, 120 mm outer diameter and 15 mm thickness. The 30 mm bore was intended to facilitate the filling of material from this end of the cell. A lid (8) of 120 mm diameter and 7 mm thickness had a hole of quarter inch at its center. A quarter inch tube was fitted to it for gas passage (9) to / from the MH bed. PTFE gasket (10) was fitted into the grooves
EP
made in the top flange and lid to assure leak proof sealing of the ETC cell. Six holes of 10 mm diameter were drilled along an appropriate pitch circle diameter on the lid as well as on top
AC C
flange to fit the lid to the top flange using threaded joint. The lower flange (11) was equipped with a capsule heater (2) and arrangements for leak proof fitting of four probe type thermocouples (12) and connections for one thermocouple junction each brazed on heater surface and inner surface of high pressure cylinder (3). Thus temperatures at five points at different radii and same axial height and at two points at same radius and different axial heights were monitored for measurement of ETC. The temperatures at same radius and different axial heights were monitored to assure that temperature variation in lateral direction was negligible. Arrangement was also made for electrical connection to the capsule heater using an in-house 5
ACCEPTED MANUSCRIPT
developed high pressure feed-through. The mass of La0.8Ce0.2Ni5 alloy powder, filled in the ETC cell was 310 g. The capsule heater of 60 W rated capacity was used as the source for higher temperature. It was fixed coaxially to the MH bed using a PTFE fixture to minimize the heat transfer from heater to the body of the ETC cell. Also a syndanio (a composite made from treated
RI PT
raw asbestos fibers and special grade of cement) sheet was placed between the bottom surface of heater and the lower flange to assure that there was no metallic contact between heater and ETC body. To maintain a constant lower temperature at the outer radius of MH bed, a thermostatic fluid (Make: Julabo, Model: Thermal H10) was circulated through the coolant jacket at a flow
SC
rate of 10 litres per minute. For the circulation and extraction of heat from the thermostatic fluid a thermostatic bath was utilized. The thermostatic bath had the accuracy of maintaining the temperature of coolant within ± 0.1 ˚C at the set temperature. To prevent the heat loss to
M AN U
surroundings through either top lid or lower flange, syndanio (thermal conductivity 0.06 W m-1 K-1) sheets (13) were used above and below the La0.8Ce0.2Ni5 powder bed. Fig. 2 shows the photograph of components of ETC cell. XRD analysis of La0.8Ce0.2Ni5
TE D
XRD analysis of La0.8Ce0.2Ni5 was done by authors’ group [28]. The XRD pattern was obtained using a powder diffractometer and a Cu Kα radiation source with the wavelength 1.5406 Å.The result of this analysis is shown in Fig. 3. As shown in figure, presence of secondary phase of
Compacts details
EP
La2O3 was found. The proportion of La2O3 was 1.14%.
Two different types of compacts were developed. First type contained the La0.8Ce0.2Ni5 powder,
AC C
graphite flakes (Sigma-Aldrich) and tin powder (Source: Sigma Aldrich; Purity: 99.5%). Graphite flakes were used for enhancement of ETC and tin powder was used as binder. Average flake size of the graphite flakes used in this work was 150 µm. La0.8Ce0.2Ni5 was obtained in the form of ingot from LabTech Int. Co. Ltd. Bulgaria. For making compacts the ingots were initially crushed in the mortar and then ground using a cup mill in dry atmosphere. The graphite flakes (6 wt%) and tin powder (1.5 wt%) were then mixed with the alloy powder. The components were mixed thoroughly by dry milling using a cup mill. The rings were not used during milling to avoid deterioration of metal alloy powder and graphite flakes. The mixture of 6
ACCEPTED MANUSCRIPT
La0.8Ce0.2Ni5 alloy, graphite flakes and tin powder was then taken to a compact die in measured amount and was pressed on a hydraulic press with a pressure of 200 MPa to obtain the compacts. This pressure for compaction was decided on the basis of similar work reported in literature [17, 20-27].
The die was designed to provide an axial hole of 10.1 mm in the compacts to
RI PT
accommodate the capsule heater. Fig. 4 depicts the steps involved in the making of compacts. While the compaction of metal alloy powder considerably decreases the permeability, the addition of ENG flakes results in overcoming this effect along with significant increase in ETC of the bed [21, 24-27]. The radial dimensions of the compacts were designed to fit the ETC cell
SC
described in the previous section. Hence the compacts were of 29.8 mm outer diameter and 10.1 mm inner diameter. Holes were drilled in the compacts for accommodating the thermocouple
M AN U
probes. Thermal paste having a composition of aluminium oxide suspended in silicon compound, was applied on the thermocouple probes to fill the micro gaps between compacts and different components of ETC cell. In case of second type of compacts copper wire mesh structure was embedded in the graphite flakes and metal powder mixture. The intentions behind including copper wire mesh structure were; to exploit the benefit of high thermal conductivity of copper while keeping the volume fraction occupied by augmentation source to suitably lower values and Table 1 provides detailed
TE D
to provide a set of fins for radial heat transfer enhancement.
information about the copper wire mesh used in this work. Dimension wise, both types of compacts were similar. Fig. 5 shows both the types of compacts. For measurement of ETC, the compacts were piled one over the other in the ETC cell, with the capsule heater passing through
EP
the axial hole of the compacts.
Table 2 depicts the packing densities and bed porosities for the three types of beds. To measure
AC C
the bed porosity following steps were performed: (i) measuring the volume of the empty reactor (in the present case it was 9.18 x 10-5 m3) (ii) filling the reactor with the metal alloy (either in powder or compact form) (iii) supply of known mass of argon gas to the filled reactor and recording its pressure and temperature (iv) calculation of void volume in the reactor using the mass of argon, pressure and temperature in the reactor. This void volume and the reactor volume is then used for calculation of bed porosity. The bed porosity decreases from 58 % to 34% due to compaction. The packing density increases from 3.79 to 5.47 g.cm-3.
7
ACCEPTED MANUSCRIPT
Experimental Set up Fig. 6 shows the schematic diagram of experimental set up used in the present work. In our experiments, we intended not only to evaluate the ETC augmentation achieved through the
RI PT
augmentation technique but also to evaluate the relative effect of the augmentation technique on the amount of hydrogen absorbed by the metal hydride. To fulfill our intentions the experimental set up was designed for activation of metal hydride, quantification of hydrogen absorption / desorption and ETC measurement. Quarter inch seamless SS 316 tubes were used for building
SC
the gas flow circuit. These tubes could sustain pressure up to 500 bar. The gas flow in this circuit was controlled through high pressure bellow sealed valves (working pressure up to 172 bar, in the temperature range of -28˚C to 343˚C). Calibrated SS 316 cylinders (2250 ml-1No., 1000 ml-
M AN U
2 No.s) were utilized for supply of hydrogen gas to the ETC cell in metered quantities. One of the two 1000 ml cylinders was used as reference volume while the other two cylinders formed the major part of supply volume. For monitoring hydrogen pressure, piezoresistive type pressure transducers were installed at different locations in the set up. P1 and P2 were installed for monitoring the supply pressure and transducer P3 was used for monitoring the pressure in the ETC cell. The working pressure range for all these pressure transducers was 0 to 100 bar with
TE D
0.1% accuracy on full scale. With the aim to minimize the error in the measurement of absorption quantity of hydrogen, a differential pressure transducer DP was used between the reference volume and selected volumes. The operating pressure range of the DP was 0 to 5 bar. ‘K’ type thermocouples with an accuracy of ±0.5˚C and time constant of 0.2 s were used at
EP
different locations in the hydrogen supply circuit and in the ETC cell to monitor the hydrogen gas temperature as shown in Fig. 6. A rotary type vacuum pump was used for evacuating the
AC C
system as per the requirement. Pirani gauge attached to the vacuum line was utilized for monitoring the vacuum level in the circuit. Computerized data logger system was used for monitoring and recording all the temperatures and pressure data during the experiments. High purity argon (99.999% purity) was used for volume measurement of different parts of the circuit and for leakage tests.
To calibrate the fabricated ETC measurement cell, pilot tests were performed using different test specimens of known thermal conductivity (stainless steel, brass and PTFE) with dimensions 30 8
ACCEPTED MANUSCRIPT
mm outer diameter, 10 mm inner diameter and 130 mm height to resemble the MH bed. Tests were performed for different source and sink temperature ranges. The maximum error in measurement of thermal conductivity of standard sample is less than ± 5%.
RI PT
Experimental Procedure After filling the ETC cell with material (either in powder form or in compacts form), the cell was heated to a temperature of 200ºC and then evacuated for few hours while maintaining the temperature. This was done to remove moisture and foreign gases present in the cell. The ETC
SC
was then measured under vacuum and different argon pressures for this configuration of bed when the metal alloy was not converted to hydride. To measure the ETC of metal hydrides, first
M AN U
the metal alloy was required to be activated. The activation procedure for La0.8Ce0.2Ni5 has already been discussed [2].
The ETC estimation experiment comprised following steps (Refer Fig. 6): 1. Evacuation of entire setup to 10-3 mbar pressure by opening all valves except v1, v2 and
TE D
v3.
2. The temperature of coolant was set to a fixed value (the coolant temperature was determined by trial and error method to achieve a particular average bed temperature) and coolant circulation is started.
EP
3. Power was supplied to the capsule heater through a DC power supply unit.
AC C
4. All the thermocouples readings were monitored in the ETC cell in intervals of 15 minutes till the steady state is reached.
5. Temperatures T4 to T11 were recorded after attaining steady state. 6. The ETC value for the MH bed at vacuum was estimated using equation (1) as [6]:
ETC =
Q 2π L ( dT d ln r )
(1)
9
ACCEPTED MANUSCRIPT
7. Valves v4 and v11 were closed and hydrogen (99.999% pure) was supplied to volumes V2, V3, V4, V5, V6 and V7 to a predetermined incremental pressure by opening valve v1. After reaching steady state, valve v1 was closed and initial pressures in the volumes V2, V3, V4,
RI PT
V5, V6 and V7 were recorded. 8. Valve v11 was opened to allow hydrogen to the ETC cell. The final pressure values after attaining steady state were recorded. The amount of hydrogen absorbed by the La0.8Ce0.2Ni5 hydride powder was calculated by mass balance on the basis of drop in
SC
hydrogen pressure in the supply volumes.
9. Steps 2 to 8 are repeated to estimate the value of ETC corresponding to different
Results and Discussion
M AN U
hydrogen concentration, pressure and average bed temperatures.
Compaction has proven to be one of the best methods for augmentation of ETC of MH beds. Mixing of graphite flakes has also been implemented by some researchers with different metal hydrides [21-27, 31-36]. It is an established fact that mixing of graphite flakes will result in increase in ETC but the quantity of ENG added is an important parameter as ENG
TE D
flakes need more space as compared to that required for same quantity of metal hydride. The increase in ETC and heat transfer characteristics (and consequently better hydrogen sorption kinetics) may be at a cost of decreased amount of hydrogen absorbed for that volume of MH
EP
bed. Inclusion of metallic structures also results in enhancement of ETC of MH beds again at the cost of decrease in amount of hydrogen absorbed. A common phenomenon that can be observed in the graphs in Figs. 7, 8, 10-12 is that, for the
AC C
same set of operating parameters, the ETC values obtained with CMPGFCu is around 1.5 times of CMPGF and 6 times of LMP bed thus clearly showing the level of augmentation of ETC achieved by different types of compacts. Fig. 7 shows the variation of ETC with pressure for La0.8Ce0.2Ni5 bed with LMP, CMPGF and CMPGFCu. As the solid thermal conductivity shows no variation with pressure, the variation in ETC with pressure is mainly the consequence of variation in gas thermal conductivity, if the filling gas is an inert gas. Thermal conductivity of gas is dependent on the kinetic energy of its molecules and their collisions. In a fixed volume powder bed with a 10
ACCEPTED MANUSCRIPT
certain filling gas, the variation in pressure results in variation in the mean free path of the gas molecules consequently affecting the rate of collisions between the molecules. Hence, for such beds, the variation shown by ETC with change in pressure is similar to the variation shown by the gas thermal conductivity.
RI PT
In case of metal hydride beds, when the bed is exposed to hydrogen, the reaction between solid particles and hydrogen causes significant variation in bed porosity and solid particle geometry consequently affecting the ETC of the bed. Therefore, to evaluate the effect of variation of gas thermal conductivity (with pressure) on the ETC of the bed, initially the ETC
SC
was measured for the non- hydrogenated bed of metal alloy. As shown in Fig. 7a, the variation of ETC with pressure forms a tilted ‘S’ shaped curve. Similar trend of ETC
M AN U
variation with pressure for metal hydride beds have been reported in literature [15-19, 22, 29]. The curve can be divided into three regimes. For low pressure region (from 0.001 to 1.5 bar) the gas pressure does not show any effect on the thermal conductivity of gas as the density of gas present in the voids of the bed, is low and the mean free path of the gas molecule is large thus reducing the number of collisions. This leads to unaffected gas thermal conductivity and consequently the ETC of bed remains almost constant in this region. With
TE D
the increase in pressure from 1.5 bar to 40 bar, the density of gas increases and mean free path of molecule decreases. As a result the variation in pressure in this regime, shows magnified effect on thermal conductivity of gas and hence there is a drastic increase in ETC. This trend continues till the pressure value reaches break away pressure [1]. The effect of
EP
augmentation techniques is clearly visible in the figure. As the bed porosity decreases with compaction thus the apparent solid thermal conductivity increases, causing increase in ETC
AC C
of bed. The variation in thermal conductivity of gas does not seem to be affected by compaction as the variation in ETC with respect to gas pressure remains almost similar for all the three types of beds. As shown in Fig. 5b, the ETC augmentation structure made with copper wire mesh has radial fins. These fins enhance the radial heat conduction. In case of compacts embedded with CWM structure the ETC augmentation achieved due to compaction and addition of graphite flakes is further increased due to the CWM structure. Hence a considerable difference in the increase in ETC values in the cases of CMHGF and CMHGFCu can be observed in Fig. 7b. After break away pressure, further increase in 11
ACCEPTED MANUSCRIPT
pressure will not cause any change in thermal conductivity of gas and hence ETC of bed. The effect of variation in pressure is supplemented with other effects when the gas in the bed is hydrogen instead of non-reactive gas, because with exposure to hydrogen, the metal alloy particles react with it to make metal hydride. There is a swelling effect due to hydrogenation
RI PT
of these particles. With this swelling the contact profile between solid particles changes causing decrease in contact resistance. Thus, although change in pressure has no direct effect on thermal conductivity of solid but due to metal alloy – hydrogen interaction, the ETC amplifies for the same pressure as compared to that with argon (an inert gas). Fig. 7b shows
SC
the variation of ETC with pressure when the filling gas is hydrogen. The trends of ETC variation in the intermediate regime were similar (but larger) to that of argon as filling gas
M AN U
verifying the explanation given for the case of argon. The effect of augmentation was even better when combined with the effect of hydrogenation. The hydrogenation results in decrease in bed porosity and increase in the thermal conductivity. This effect was already gained by compaction which was further amplified by hydrogenation of beds. Table 3 depicts the effects of pressure variation as well as that of ETC augmentation on ETC of all the three types of beds.
TE D
The effect of hydrogen concentration on ETC can be seen as the net effect of the following four phenomena: 1.The reaction between metal alloy particles and hydrogen causes swelling of solid particles and thus increasing contact area between particles which ultimately results in increase in ETC of bed. 2. The increase in hydrogen concentration involves increase in
EP
hydrogen pressure thus the effect of increase in pressure gets involved in the effect of hydrogen concentration on ETC. 3. The swelling of solid particles causes decrease in bed
AC C
porosity as the bed volume is constant. This decrease in porosity is a positive factor for ETC of the MH bed. 4. As the hydrogen atoms get embedded to solid particles, the electronic configurations of solid particles change. The major impact is of variation in bed porosity and hydrogen gas pressure. The effect of variation of hydrogen concentration on ETC has been shown in Fig. 8. Extracts of the data from Fig. 8 has been tabulated in Table 4. The porosities in both the cases of pellets beds are lower as compared to that of powder bed at the beginning of hydrogenation process. With the increase in hydrogen concentration the bed porosity further decreases thus resulting in an increase in ETC. In case of pellets, probably this 12
ACCEPTED MANUSCRIPT
decrease in porosity is more significant than in case of powder bed thus resulting in significant increase in ETC in the pellets beds. Moreover, as can be seen in Fig. 7 of the revised manuscript, the increase in ETC with increase in hydrogen pressure is much lower in case of powder bed. These two combined effects lead to insignificant variation in thermal
RI PT
conductivity in case of powder bed. Also it is observed that the ETC for all the three types of beds increases more sharply from hydrogen concentration of 0 to about 0.1 wt%. From PCI diagram (Fig. 9) at 60˚C, it can be observed that change in hydrogen concentration from 0 to 0.1 wt%, the corresponding equilibrium pressure of the bed is increasing from 1 bar to 50
SC
bar. The maximum increase in ETC is occurring in this pressure range as can be seen from Fig. 7b. Above 0.1 wt% hydrogen concentration, the variation in pressure with hydrogen
M AN U
concentration is not significant. Hence concentration variation has less effect on ETC value in this region. For the same temperature and pressure conditions, hydrogen concentration (in wt%) : ETC (in W m-1 K-1 ) for LMP/CMPGF/CMPGFCu were observed to be 0.8 : 1.22 / 0.76 : 4.39 / 0.78 : 5.67. The values of hydrogen concentration clearly suggest that although there were significant enhancements in the ETC values with augmentation techniques, the decreases in quantities of hydrogen absorption were below 5.2%. The amplitude of effect of
TE D
ETC augmentation on quantity of hydrogen absorption is in conformation with the reported values in literature for similar ENG content and compaction pressure [17-18, 21, 23-27]. Fig. 10 shows the variation of ETC with pressure during hydrogen absorption and desorption processes for powder beds. As shown in the figure, the ETC during desorption is marginally
EP
more than that during absorption at a particular pressure. This can be explained with the help of PCI curve as shown in Fig. 9. The PCI characteristics of La0.8Ce0.2Ni5 hydride was
AC C
performed by authors’ group [28]. The detailed information regarding the experimental set up and methodology adopted etc. are provided elsewhere [28]. At a given pressure the concentration of hydrogen in metal hydride is slightly higher in case of desorption curve compared to absorption curve due to hysteresis. Due to the higher concentration ETC value during desorption is slightly higher than that of absorption. During desorption the reactor (ETC cell) is connected to a sample cylinder which is filled with hydrogen at atmospheric temperature and a particular pressure (which is below the reactor pressure). The metal hydride in the reactor desorbs hydrogen corresponding to the desorption temperature. The 13
ACCEPTED MANUSCRIPT
thermal conductivity is measured when the system (reactor and sample cylinder) achieves an equilibrium pressure. For the next step, the valve v11 is closed, thus disconnecting the ETC cell from the sample cylinder. The sample cylinder is then brought to a lower pressure and again the ETC cell is connected to sample cylinder by opening the valve v11. This process is
RI PT
repeated to obtain lower equilibrium pressures during desorption process.
Temperature variation is reported to have different types of effects on ETC in different studies [5]. In the present work, the average temperature variation seemed ineffective in
SC
bringing about any change in the ETC of MH beds although the effect of augmentation was clearly visible. For an average bed temperature of 30 ºC and 50 bar pressure, ETC values were 1.37 / 5.1 / 7 W m-1 K-1 for LMP/CMPGF/CMPGFCu. As shown in Figs. 11 (a) and 11
M AN U
(b) although the ETC is changing with pressure but the graph between ETC and average temperature remained almost a horizontal line in all the cases except in case of CMPGFCu bed for 10 bar pressure where a slight decrease in ETC is observed when average bed temperature was increased from 50˚C to 70˚C. The reason behind this may be that the increasing effect on ETC caused due to the increasing thermal conductivity of hydrogen with temperature might be balanced by the decreasing effect caused due to decrease in thermal
TE D
conductivity of the solid particles.
The effect of ETC augmentation is clear from Fig. 12, which shows the radial (spatial) temperature profile for the MH beds at steady state for different configurations. As can be
EP
seen that the slope of the radial temperature profile reduces for CMPGF and it further reduces for CMPGFCu beds for both the MH’s when compared to that of LMP beds. The
AC C
slopes of radial temperature profiles for LMP/CMPGF/CMPGFCu with argon as filling gas were -5255.17 / -3955.17 / -3312.07 K/m whereas with hydrogen these were -3670.69 / 1036.21 / -731.03 K/m respectively. This clearly shows the improvement in heat transfer characteristics with the augmentation of ETC. Since filling of argon gas will result into comparatively lower values of ETC for these beds as compared to those with hydrogen as filling gas, the slopes for the cases of hydrogen are lower than those for argon for the respective beds.
14
ACCEPTED MANUSCRIPT
The ETC values obtained in the present work with the two types of compacts are close to those reported in literature for the similar ENG content, compaction pressure and operating parameters [21, 23-27]. The uncertainties / accuracies of different measuring instruments used for this work are given in Table 5. Using these values the maximum uncertainty in the
RI PT
estimation of ETC values was found to be ± 5.1%. Conclusion
The ETC augmentations for metal hydride beds of La0.8Ce0.2Ni5 are attempted using different
SC
types of compacts. The effect of compaction and mixing of graphite flakes resulted in the increase of ETC to about four times that of powder bed while the increase in ETC was found
M AN U
to be around six times that of powder bed for the CMPGFCu configuration. The decrease in quantity of hydrogen absorbed by La0.8Ce0.2Ni5 hydride due to ETC augmentation was very less (< 6%) as the mass fraction and hence the volume fraction occupied by graphite flakes was intentionally kept low. The inclusion of copper wire mesh resulted in continuous heat conduction path in the radial direction as the mesh consists of seven radial porous fins. Thus inclusion of copper wire mesh lead to enhancement in the heat transfer in radial direction and
TE D
consequently in the ETC. Using high pressure for compaction with higher wt% of graphite flakes for making metal hydride compacts will also increase the conduction path in the radial direction, however it reduces the rate of hydrogen diffusion in the metal hydride beds. Hence, it can be concluded that the compacts made with mixture of metal alloy powder and graphite
EP
flakes and embedded with copper wire mesh structure is a better method of ETC augmentation for MH beds with minimal loss in hydrogen absorption quantity.
AC C
Nomenclature CWM
Copper wire mesh structure
ENG
Expanded natural graphite
ETC
Effective thermal conductivity, W m-1 K-1
L
Length of metal hydride bed, m 15
ACCEPTED MANUSCRIPT
Loose metal alloy / composite powder
CMPGF
Compacts of metal alloy / composite powder and graphite flakes
CMPGFCu
Compacts of metal alloy / composite powder, graphite flakes and embedded with copper wire mesh structure Heat transfer rate, W
R
Radius of metal hydride bed, m
T
Temperature, K
M AN U
SC
Q
RI PT
LMP
Acknowledgements
One of the authors, Dr. E. Anil Kumar, wishes to sincerely acknowledge the Science and Engineering Research Board, Department of Science and Technology, Government of India
References
TE D
for partially funding the work vide Project No. SERB/F/0810/2013-14.
1. Madaria Y, Anil Kumar E, Murthy SS. Effective thermal conductivity of reactive packed
EP
beds of hydriding materials. App Therm Eng 2016;98:976-90. 2. Madaria Y, Anil Kumar E. Effect of heat transfer enhancement on the performance of
AC C
metal hydride based hydrogen compressor. Int J Hydrogen Energy 2016;41:3961-73. 3. Sun DW, Song JD. Theoretical descriptions and experimental measurements on the effective thermal conductivity in metal hydride powder beds. J Less-Common Met 1990;160:387-95. 4. Sun DW, Deng SJ. Theoretical descriptions and experimental measurements on the effective thermal conductivity of metal hydride powder beds. J Less-Common Met 1990;160:387-95. 5. Kempf A, Martin W. Measurement of the effective thermal properties of TiFe0.85Mn0.15 and its hydrides, Int J Hydrogen Energy 1986;11:107-16. 6. Suda S, Kobayashi N, Yoshida K, Ishido Y, Ono S. Experimental measurements of thermal conductivity. J Less-Common Met 1980;74:127-36. 16
ACCEPTED MANUSCRIPT
7. Dedrick DE, Kanouff MP, Replogle BC, Gross KJ. Thermal properties characterization of sodium alanates. J Alloy Compd 2005;389:299-305. 8. Ishido Y, Kawamura M, Ono S. Thermal conductivity of magnesium-nickel hydride powder beds in a hydrogen atmosphere. Int J Hydrogen Energy 1982;7:173-82.
RI PT
9. Gustafsson SE. Transient plane source techniques for thermal conductivity and thermal diffusivity measurements of solid materials. Rev Sci Instrum 1991;62:797-804. 10. Kapischke J, Hapke J, Measurement of effective thermal conductivity of a Mg-MgH2 packed bed with oscillating heating. Exp Therm Fluid Sci 1998;17:347-55.
SC
11. Y. Asakuma, S. Miyauchi, T. Yamamoto, H. Aoki, T. Muira, Homogenization method for effective thermal conductivity of metal hydride bed, Int J Hydrogen Energy 2004;29:209–16. 12. Madaria Y, Anil Kumar E, Maiya MP, Murthy SS. Simulation of effective thermal conductivity of metal hydride packed beds. Heat Transf Eng 2016;37:616-24.
M AN U
13. Anil Kumar E, Maiya MP, Murthy SS. Measurement and analysis of effective thermal conductivity of MmNi4.5Al0.5 hydride bed. Ind Eng Chem Res 2011;50:12990-9. 14. Inoue S, Yushi I, Yukihiko M. Drastic enhancement of effective thermal conductivity of a metal hydride packed bed by direct synthesis of single-walled carbon nanotubes. Int J Hydrogen Energy 2012;37:1836-41.
TE D
15. Suda S, Komazaki Y, Kobayashi N. Effective thermal conductivity of metal hydride beds. J LessCommon Met 1983;89:317-24. 16. Nagel M, Komazaki Y. Effective thermal conductivity of a metal hydride bed augmented with a copper wire matrix. J Less-Common Met 1986;120:35-43.
EP
17. Bershadsky E, Josephy Y, Ron M. Permeability and thermal conductivity of porous metallic matrix hydride compacts. J Less-Common Met 1989;153:65-78. 18. Ron M, Bershadsky M, Josephy Y. The thermal conductivity of porous metal matrix hydride compacts. J Less-Common Met 1991;172-174:1138-46.
AC C
19. Suda S, Komazaki Y. The effective thermal conductivity of a metal hydride bed packed in a multiple-waved sheet metal structure. J Less-Common Met 1991;172-174:1130-7. 20. Ron M, Bershadsky E, Josephy Y. Thermal conductivity of PMH compacts, measurements and evaluation. Int J Hydrogen Energy 1992;17:623-30. 21. Kim KJ, Montoya B, Razani A, Lee K-H. Metal hydride compacts of improved thermal conductivity. Int J Hydrogen Energy 2001;26:609-13. 22. Yan M-Y, Sun F, Liu X-P, Ye J-H. Effects of compaction pressure and graphite content on hydrogen storage properties of Mg(NH2)2-2LiH hydride. Int J Hydrogen Energy 2014;39:1965661. 17
ACCEPTED MANUSCRIPT
23. Sanchez RA, Klein H-P, Groll M. Expanded graphite as heat transfer matrix in metal hydride beds. Int J Hydrogen Energy 2003;28:5156-27. 24. Pohlmann C, Rontzsch L, Kalinichenka S, Hutsch T, Weibgarber T, Kieback B.
RI PT
Hydrogen storage properties of compacts of melt-spun Mg90Ni10 flakes and expanded natural graphite. J Alloy Compd 2011;509S:S625-8. 25. Pohlmann C, Rontzsch L, Kalinichenka S, Hutsch T, Kieback B. Magnesium alloy-
graphite composites with tailored heat conduction properties for hydrogen storage applications. Int J Hydrogen Energy 2010;35:12829-36. 26. Pohlmann C, Rontzsch L, Heubner F, Weibgarber T, Kieback B. Solid-state hydrogen
SC
storage in Hydralloy-graphite composites. J Power Sources 2013;231:97-105.
M AN U
27. Shim J-H, Park M, Lee YH, Kim S, Im YH, Suh J-Y, et al. Effective thermal conductivity of MgH2 compacts containing expanded natural graphite under a hydrogen atmosphere. Int J Hydrogen Energy 2014;39:349-55. 28. Sharma VK, Anil Kumar E. Studies on La based intermetallic hydrides to determine their suitability in metal hydride based cooling systems. Intermetallics 2015;57:60-7. 29. Muthukumr P, Maiya MP, Srinivasa Murthy S. Experiments on a metal hydride based hydrogen compressor. Int J Hydrogen Energy 2005;30:879-92. 30. Kim JK, Park IS, Kim KJ, Gawlik K. A hydrogen-compression system using porous metal hydride pellets of LaNi5-xAlx. Int J Hydrogen Energy 2008;33:870-7.
TE D
31. Chaise A, Rango P.de, Marty Ph, Fruchart D, Miraglia S, Olives R, Garrier S. Enhancement of hydrogen sorption in magnesium hydride using expanded natural graphite. Int J Hydrogen Energy 2009;34:8589-96.
EP
32. Y. Ishido, M. Kawamura, S. Ono. Thermal conductivity of magnesium-nickel hydride powder beds in a hydrogen atmosphere. Int J Hydrogen Energy 1982;7:173-82.
AC C
33. Yasuda N, Tsuchiya T, Okinaka N, Akyama T. Thermal conductivity and cycle characteristics of metal hydride sheet formed using aramid pulp and carbon fiber. Int J Hydrogen Energy 2013;38:1657-61. 34. Wang H, Prasad AK, Advani SG. Hydrogen storage system based on hydride materials with enhanced thermal conductivity. Int J Hydrogen Energy 2012;37:290-8. 35. Zamengo M, Ryu J, Kato Y. Thermochemical performance of magnesium hydroxide – expanded graphite pellets for chemical pump. App Therm Eng 2014;64:339-47. 36. Dieterich M, Pohlmann C, Burger I, Linder M, Rontzsch L. Long – term cycle stability of metal hydride – graphite composite. Int J Hydrogen Energy 2015;40:16375-82.
18
ACCEPTED MANUSCRIPT
Value
Surface area per unit volume
8333 m2 m-3
Density
5.6 g cm-3
Thermal conductivity
392 W m-1 K-1
Wire diameter
0.12 mm
Purity
99.9%
AC C
EP
TE D
M AN U
SC
Parameter
RI PT
Table 1 Physical properties of copper wire mesh used for making compacts
ACCEPTED MANUSCRIPT
Bed porosity (%)
LMP
3.79
58
CMPGF
4.94
38
CMPGFCu
5.47
34
AC C
EP
TE D
M AN U
SC
Packing density (g.cm-3)
RI PT
Table 2 Packing density and bed porosity of different types of beds
ACCEPTED MANUSCRIPT
Table 3 Variation of ETC with pressure for different filling gases Pressure
ETC with Ar as filling gas -1
(bar)
ETC with H2 as filling gas
-1
(W m-1 K-1)
(W m K ) CMPGF
CMPGFCu
LMP
CMPGF
CMPGFCu
0.01
0.58
1.9
3.01
0.58
1.9
3.0
1.5
0.62
2.1
3.3
0.59
2.0
3.3
10
0.68
2.6
3.6
0.69
2.9
4.28
40
0.93
3.1
4.05
1.29
4.63
6.2
60
0.95
3.2
4.22
1.3
4.7
6.8
70
0.95
3.2
4.22
1.3
4.7
6.8
AC C
EP
TE D
M AN U
SC
RI PT
LMP
ACCEPTED MANUSCRIPT
Table 4 Variation of ETC with hydrogen concentration for different types of beds
RI PT
ETC (W m-1 K-1) CMPGF CMPGFCu 1.0 1.33 2.64 3.42 4.12 4.94 4.57 6.31 4.73 6.85
TE D
M AN U
SC
LMP 0.59 0.71 1.19 1.27 1.31
EP
0.5 bar 5 bar 10 bar 20 bar 50 bar
Hydrogen concentration (wt%) LMP CMPGF PMPGFCu 0.02 0.01 0.01 0.08 0.069 0.07 0.29 0.25 0.23 1.38 1.29 1.25 1.45 1.43 1.41
AC C
Pressure
ACCEPTED MANUSCRIPT
Table 5 Uncertainties of various measured parameters for estimating ETC S. No.
Measurement
Device
Accuracy
Minimum
Maximum
quantity
uncertainty
measured Bed temperature
K type thermocouple
± 0.5 K
2.
Hydrogen gas
Piezoresistive type
± 0.000625 bar
1 bar
±0.0625%
pressure
transducer
Temperature of
K type thermocouple
± 0.5 K
283 K
±0.177%
ETC cell
± 0.5 W
30 W
±1.66%
3.
4.
Heat transfer rate estimation
SC
thermostatic fluid
293 K
±0.171%
RI PT
1.
Mass of alloy
Digital weighing balance
±0.1g
25 g
±0.4%
6.
Length
Digital vernier caliper
±0.1mm
5 mm
±2%
AC C
EP
TE D
M AN U
5.
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 1 Schematic diagram of ETC cell
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
Fig. 2 Components of ETC cell
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
Fig. 3 XRD Pattern of La0.8Ce0.2Ni5 alloy [28]
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
Fig. 4 Steps in compact making
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 5 (a) Compacts of La0.8Ce0.2Ni5 powder mixed with graphite flakes
AC C
EP
TE D
M AN U
(b) Compacts with copper wire mesh
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 6 Experimental set up for ETC measurement
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
(a)
(b) Fig. 7 Variation of effective thermal conductivity with pressure (a) Filling gas: Argon (b) Filling gas: Hydrogen
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
Fig. 8 Variation of effective thermal conductivity with hydrogen concentration
ACCEPTED MANUSCRIPT
100
Pressure (bar)
RI PT
10
1
SC
10°C 20°C 40°C 60°C 80°C
M AN U
Absorption Desorption 0.1 0
0.2
0.4
0.6 0.8 1 Storage Capacity (wt%)
1.2
1.4
1.6
AC C
EP
TE D
Fig. 9 Pressure concentration Isotherms for La0.8Ce0.2Ni5 hydride [28]
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
Fig. 10 Variation of effective thermal conductivity with pressure during hydrogen absorption and desorption
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
(a)
(b) Fig. 11 Variation of effective thermal conductivity with average bed temperature (a) Hydrogen pressure = 10 bar (b) Hydrogen pressure = 50 bar
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
Fig. 12 Radial bed temperature profile of MH bed
ACCEPTED MANUSCRIPT
Highlights of the research article
AC C
EP
TE D
M AN U
SC
RI PT
• ETC of La0.8Ce0.2Ni5 hydride bed was measured. • Two different types of pellets were developed for augmentation of ETC. • Pellets with embedded copper wire mesh were found to be better option for ETC augmentation.
S0925-8388(16)32683-4
DOI:
10.1016/j.jallcom.2016.08.283
Reference:
JALCOM 38785
To appear in:
Journal of Alloys and Compounds
Received Date: 22 June 2016 Revised Date:
26 August 2016
Accepted Date: 27 August 2016
Please cite this article as: Y. Madaria, E. Anil Kumar, Measurement and augmentation of effective thermal conductivity of La0.8Ce0.2Ni5 hydride bed, Journal of Alloys and Compounds (2016), doi: 10.1016/j.jallcom.2016.08.283. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
Measurement and Augmentation of Effective Thermal Conductivity of La0.8Ce0.2Ni5 hydride bed Yogesh Madaria and E. Anil Kumar1
RI PT
Discipline of Mechanical Engineering, Indian Institute of Technology Indore, Indore, Madhya Pradesh, India 453552 ABSTRACT
An experimental set up for measuring Effective Thermal Conductivity (ETC) of Metal Hydride
SC
(MH) beds was fabricated using one-dimensional steady-state radial heat transfer absolute method. La0.8Ce0.2Ni5 alloy powder and its hydride were considered in the present study. Two types of compacts were fabricated for ETC augmentation of MH bed. First type was made from
M AN U
the mixture of 6 wt% graphite flakes with the La0.8Ce0.2Ni5 alloy powder. A three dimensional structure, having six radial fins, uniformly distributed around the outer periphery of a hollow cylinder, made with copper wire mesh, was integrated with the mixture of La0.8Ce0.2Ni5 powder and 6 wt% graphite flakes to make the second type of compacts. The ETC of the three types of beds namely loose metal powder bed (LMP), bed of compacts of metal alloy powder mixed with
TE D
graphite flakes (CMPGF) and bed of compacts of metal alloy powder mixed with graphite flakes and embedded with copper wire mesh structure (CMPGFCu) were measured and compared at different argon and hydrogen pressures, hydrogen concentrations and average bed temperatures. The maximum values of effective thermal conductivity for the bed configurations of
EP
LMP/CMPGF/CMPGFCu were 1.3/4.7/6.8 Wm-1K-1. These values of effective thermal conductivity were obtained at 70 bar hydrogen gas pressure and 60˚C average bed temperature. The decrease in mass fraction of hydrogen absorbed by the La0.8Ce0.2Ni5 hydride, as a result of
AC C
implementation of ETC augmentation techniques, was nominal (around 5.7% in case of CMPGF and 4.5% in case of CMPGFCu) when compared to the respective powder bed configurations. Keywords: Metal hydride bed, Effective thermal conductivity, Compacts, Steady state method.
1
Corresponding author: E-mail: [email protected]
1
ACCEPTED MANUSCRIPT
INTRODUCTION Certain metals, alloys or intermetallics can reversibly absorb and desorb hydrogen. By absorption of hydrogen these materials form a class of compounds called metal hydrides. The
RI PT
hydrogen absorption is an exothermic process and desorption is an endothermic process. By virtue of these reactions metal hydrides have been used in some important applications involving heat exchange such as sorption based cooling systems, hydrogen purifiers, hydrogen compressors, etc [1] other than hydrogen storage. For any of these applications hydrogen transfer (absorption/desorption) rate is very important, as it decides the cycle time. Also hydrogen
SC
absorption and desorption rates are heat and mass transfer dependent [2]. The prevailing conditions in metal hydride beds show that almost all the heat transfer through the metal hydride
important
role
in
heat
M AN U
bed is by conduction mode. Thus effective thermal conductivity (ETC) of the bed plays an transfer
from/to
the
metal
hydride
bed.
Cyclic
hydrogenation/dehydrogenation results in reduction in particle size of the metal hydride leading to final average size of particles in the range of 10-15 microns [3]. ETC varies with hydrogen concentration, particle size and bed porosity. Owing to these dependencies, it becomes difficult to measure the effective thermal conductivity of metal hydride beds. Several measurement
TE D
techniques have been developed and utilized for determination of ETC of MH beds. Sun and Deng [4], Kempf and Martin [5], Suda et al. [6] etc. applied steady state methods while Dedrick et al. [7], Ishido et al. [8], Gustafsson [9] etc. have opted for transient methods for measurement of ETC of metal hydride beds. Kapischke and Hapke [10] had proposed oscillation heating
EP
instead of continuous heating for ETC measurement. Also different simulation methods have been applied for estimation of ETC of MH beds. Sun and Deng [3] suggested a mathematical
AC C
model for estimation of effective thermal conductivity of MlNi4.5Mn0.5 (Ml: lanthanum rich mischmetal). The heat transfer was considered to be taking place through three mechanisms in parallel to each other, those were (i) conduction and radiation through voids, (ii) through a series path comprising of effective solid path and effective gas path and (iii) through diffusion between solid particles. The summation of the three fractional heat transfers gives total heat transfer which is then used to calculate the ETC of bed. Asakuma et al. [11] applied a homogenization method for calculation of ETC of LaNi4.7Al0.3Hx bed. The change in contact area due to hydrogenation/dehydrogenation was examined and the effect of particle size depreciation on the 2
ACCEPTED MANUSCRIPT
ETC of MH bed was accommodated. Authors’ group has proposed appropriate extensions to some renowned mathematical models, originally developed for non reactive packed beds, so as to make those models suitable for estimation of ETC of MH beds [12]. The extended models were validated with the experimental results from the work of E. Anil Kumar et al. [13].
RI PT
Effective thermal conductivity of metal hydride beds is generally in the range of 0.1-1.2 W m-1 K-1[4-10]. This leads to poor heat transfer characteristics of hydride beds. For better hydrogen sorption characteristics and more importantly for efficient working of metal hydride based heat exchanging systems the value of ETC must be improved to a value suitable for applications like
SC
metal hydride based hydrogen compressor, metal hydride based thermal energy storage system, etc. (target range for present work was 3-7 W m-1 K-1). Several augmentation techniques have
M AN U
been tried with different levels of successes. Extensive review on experimental methods, simulation techniques and augmentation of ETC of MH beds has been done by the authors [1]. Some of the common augmentation techniques are: inclusion of high thermal conductivity metallic foams/metallic structures/fins, mixing of graphite fibers, metallic coating of metal hydride particles and compaction etc. A rather unconventional technique was that of direct synthesis of single walled carbon nano tube (SWCNT) on MH particles which resulted in drastic
TE D
enhancement in ETC of MH bed. An ETC value of 16 W m-1 K-1 was claimed to be achievable with SWCNT filling ratio of 0.5 and a slightly reduced particle packing ratio although no experimental evidences are available for the success of this technique [14]. ETC augmentation techniques like insertion of metal foams, porous metal matrix compaction, microencapsulated
EP
compacts, etc. results in increase of parasitic thermal mass of the reactor bed therefore compromising with performance of MH beds [15-21]. As compared to other augmentation
AC C
techniques, compaction has grabbed more attention and has proven to be more advantageous [1]. Yan et al. [22] and Sanchez et al. [23] observed that the ETC does not depends only on expanded natural graphite to metal hydride (ENG/MH) mass ratio but also on the fill factor and porosity and therefore on the pressing conditions. The ETC can be enhanced by intensified pressing; however this might have a negative influence on the hydrogen flow through the compact. Yan et al. [22] suggested that ENG should be added as low as possible in meeting the heat transfer requirements for the endothermic desorption reaction. In their study with Mg90Ni10-ENG compacts Pohlmann et al. [24-25] found that with the increase in ENG content the Mg90Ni10 3
ACCEPTED MANUSCRIPT
flakes got more misaligned, probably due to possibility of more free movement within the lubricious ENG matrix. For better alignment of Mg90Ni10 flakes in radial direction, 5 wt% ENG content was found to be suitable. While Kim et al. [21] reported that the ETC for compacts of LaNi5 and recompressed expanded graphite (REG) increased from 0.1 W m-1 K-1 for 0 wt% REG
RI PT
to 4.5 W m-1 K-1 for 10 wt% REG and 6.3 W m-1K-1 for 100 wt% REG and Pohlmann et al. [26] found a linear increase in ETC with ENG content, Shim et al. [27] observed that the ETC value for compacts of MgH2 and ENG, increased slowly with ENG content up to 10 wt% but the rate
SC
of increase of ETC with ENG content was much sharper from 10 wt% to 20 wt%.
In the present work authors have attempted to exploit the advantages of compaction, mixing of graphite flakes and inclusion of high thermal conductivity metal structure for the ETC
M AN U
augmentation of La0.8Ce0.2Ni5 hydride bed. La0.8Ce0.2Ni5 hydride was chosen as La based intermetallic hydrides are suitable for developing metal hydride based hydrogen compressors, cooling systems etc. [2, 28-30]. Two types of compacts were formed with the view to serve different purposes: (i) To assess the level of augmentation achieved by compaction and mixing of graphite flakes, first type of compacts were made by pressing a mixture of La0.8Ce0.2Ni5 powder and 6 wt% graphite flakes. (ii) Keeping in mind the adverse effects of excessive amount
TE D
of graphite flakes and those of excessive pressing on hydrogen sorption rates to/from the metal hydride, in the second type of compact, inclusion of copper wire mesh structure was tried as an alternative to both increasing the quantity of graphite flakes and increasing compaction pressure.
EP
The ETC augmentations with the two types of compacts have been compared. Experimental Details
AC C
In this work one dimensional, steady state, radial heat transfer, absolute method was used for ETC measurement. Temperature profile prevailing in the MH bed is measured with ‘K’ type thermocouples with probe of 1 mm diameter. The steady state method is time consuming and some loss in accuracy is seen due to non-fulfillment of assumption of homogeneous beds caused by the presence of thermocouples within the bed [31]. In the present case, the thermocouple probes (each of 1 mm diameter and 2.04 cm2 surface area) have occupied only 0.2% of bed volume. Therefore the problem of loss in homogeneity was negligible. Taking this into 4
ACCEPTED MANUSCRIPT
consideration, steady state method becomes more accurate and reliable than transient methods of ETC measurement and hence was chosen for the present study. ETC cell
RI PT
Fig. 1 shows the schematic diagram of ETC cell. The ETC cell is made up of a SS 316 cylinder, which was designed and fabricated to attain the required conditions of one dimensional, steady state, radial heat transfer. For achieving unidirectional radial heat transfer the aspect ratio (Length/Diameter) of the reactor was kept 4.33:1. To ascertain steady-state heat transfer two
SC
constant temperatures (as heat source and heat sink) were maintained at the inner and outer radii of the hollow cylindrical MH bed. The MH bed in powder form (1) assumed a hollow cylindrical shape with a capsule heater (2) of 10 mm diameter, 130 mm length and 60 W rated capacity, at
M AN U
the inner radius and enveloped by the inner surface of high pressure cylinder (3) of 30 mm inner diameter, 38 mm outer diameter and 170 mm height, at the outer radius. Coolant jacket (4) has an annular volume between the outer surface of high pressure cylinder and inner wall of another cylinder of 45 mm inner diameter, 49 mm outer diameter and 170 mm length. Suitable fittings were provided to this cylinder for coolant inlet (5) and outlet (6). Two flanges, one at each end of the cylinders were welded to make this a single assembly. The top flange (7) was made as a
TE D
hollow disc of 30 mm inner diameter, 120 mm outer diameter and 15 mm thickness. The 30 mm bore was intended to facilitate the filling of material from this end of the cell. A lid (8) of 120 mm diameter and 7 mm thickness had a hole of quarter inch at its center. A quarter inch tube was fitted to it for gas passage (9) to / from the MH bed. PTFE gasket (10) was fitted into the grooves
EP
made in the top flange and lid to assure leak proof sealing of the ETC cell. Six holes of 10 mm diameter were drilled along an appropriate pitch circle diameter on the lid as well as on top
AC C
flange to fit the lid to the top flange using threaded joint. The lower flange (11) was equipped with a capsule heater (2) and arrangements for leak proof fitting of four probe type thermocouples (12) and connections for one thermocouple junction each brazed on heater surface and inner surface of high pressure cylinder (3). Thus temperatures at five points at different radii and same axial height and at two points at same radius and different axial heights were monitored for measurement of ETC. The temperatures at same radius and different axial heights were monitored to assure that temperature variation in lateral direction was negligible. Arrangement was also made for electrical connection to the capsule heater using an in-house 5
ACCEPTED MANUSCRIPT
developed high pressure feed-through. The mass of La0.8Ce0.2Ni5 alloy powder, filled in the ETC cell was 310 g. The capsule heater of 60 W rated capacity was used as the source for higher temperature. It was fixed coaxially to the MH bed using a PTFE fixture to minimize the heat transfer from heater to the body of the ETC cell. Also a syndanio (a composite made from treated
RI PT
raw asbestos fibers and special grade of cement) sheet was placed between the bottom surface of heater and the lower flange to assure that there was no metallic contact between heater and ETC body. To maintain a constant lower temperature at the outer radius of MH bed, a thermostatic fluid (Make: Julabo, Model: Thermal H10) was circulated through the coolant jacket at a flow
SC
rate of 10 litres per minute. For the circulation and extraction of heat from the thermostatic fluid a thermostatic bath was utilized. The thermostatic bath had the accuracy of maintaining the temperature of coolant within ± 0.1 ˚C at the set temperature. To prevent the heat loss to
M AN U
surroundings through either top lid or lower flange, syndanio (thermal conductivity 0.06 W m-1 K-1) sheets (13) were used above and below the La0.8Ce0.2Ni5 powder bed. Fig. 2 shows the photograph of components of ETC cell. XRD analysis of La0.8Ce0.2Ni5
TE D
XRD analysis of La0.8Ce0.2Ni5 was done by authors’ group [28]. The XRD pattern was obtained using a powder diffractometer and a Cu Kα radiation source with the wavelength 1.5406 Å.The result of this analysis is shown in Fig. 3. As shown in figure, presence of secondary phase of
Compacts details
EP
La2O3 was found. The proportion of La2O3 was 1.14%.
Two different types of compacts were developed. First type contained the La0.8Ce0.2Ni5 powder,
AC C
graphite flakes (Sigma-Aldrich) and tin powder (Source: Sigma Aldrich; Purity: 99.5%). Graphite flakes were used for enhancement of ETC and tin powder was used as binder. Average flake size of the graphite flakes used in this work was 150 µm. La0.8Ce0.2Ni5 was obtained in the form of ingot from LabTech Int. Co. Ltd. Bulgaria. For making compacts the ingots were initially crushed in the mortar and then ground using a cup mill in dry atmosphere. The graphite flakes (6 wt%) and tin powder (1.5 wt%) were then mixed with the alloy powder. The components were mixed thoroughly by dry milling using a cup mill. The rings were not used during milling to avoid deterioration of metal alloy powder and graphite flakes. The mixture of 6
ACCEPTED MANUSCRIPT
La0.8Ce0.2Ni5 alloy, graphite flakes and tin powder was then taken to a compact die in measured amount and was pressed on a hydraulic press with a pressure of 200 MPa to obtain the compacts. This pressure for compaction was decided on the basis of similar work reported in literature [17, 20-27].
The die was designed to provide an axial hole of 10.1 mm in the compacts to
RI PT
accommodate the capsule heater. Fig. 4 depicts the steps involved in the making of compacts. While the compaction of metal alloy powder considerably decreases the permeability, the addition of ENG flakes results in overcoming this effect along with significant increase in ETC of the bed [21, 24-27]. The radial dimensions of the compacts were designed to fit the ETC cell
SC
described in the previous section. Hence the compacts were of 29.8 mm outer diameter and 10.1 mm inner diameter. Holes were drilled in the compacts for accommodating the thermocouple
M AN U
probes. Thermal paste having a composition of aluminium oxide suspended in silicon compound, was applied on the thermocouple probes to fill the micro gaps between compacts and different components of ETC cell. In case of second type of compacts copper wire mesh structure was embedded in the graphite flakes and metal powder mixture. The intentions behind including copper wire mesh structure were; to exploit the benefit of high thermal conductivity of copper while keeping the volume fraction occupied by augmentation source to suitably lower values and Table 1 provides detailed
TE D
to provide a set of fins for radial heat transfer enhancement.
information about the copper wire mesh used in this work. Dimension wise, both types of compacts were similar. Fig. 5 shows both the types of compacts. For measurement of ETC, the compacts were piled one over the other in the ETC cell, with the capsule heater passing through
EP
the axial hole of the compacts.
Table 2 depicts the packing densities and bed porosities for the three types of beds. To measure
AC C
the bed porosity following steps were performed: (i) measuring the volume of the empty reactor (in the present case it was 9.18 x 10-5 m3) (ii) filling the reactor with the metal alloy (either in powder or compact form) (iii) supply of known mass of argon gas to the filled reactor and recording its pressure and temperature (iv) calculation of void volume in the reactor using the mass of argon, pressure and temperature in the reactor. This void volume and the reactor volume is then used for calculation of bed porosity. The bed porosity decreases from 58 % to 34% due to compaction. The packing density increases from 3.79 to 5.47 g.cm-3.
7
ACCEPTED MANUSCRIPT
Experimental Set up Fig. 6 shows the schematic diagram of experimental set up used in the present work. In our experiments, we intended not only to evaluate the ETC augmentation achieved through the
RI PT
augmentation technique but also to evaluate the relative effect of the augmentation technique on the amount of hydrogen absorbed by the metal hydride. To fulfill our intentions the experimental set up was designed for activation of metal hydride, quantification of hydrogen absorption / desorption and ETC measurement. Quarter inch seamless SS 316 tubes were used for building
SC
the gas flow circuit. These tubes could sustain pressure up to 500 bar. The gas flow in this circuit was controlled through high pressure bellow sealed valves (working pressure up to 172 bar, in the temperature range of -28˚C to 343˚C). Calibrated SS 316 cylinders (2250 ml-1No., 1000 ml-
M AN U
2 No.s) were utilized for supply of hydrogen gas to the ETC cell in metered quantities. One of the two 1000 ml cylinders was used as reference volume while the other two cylinders formed the major part of supply volume. For monitoring hydrogen pressure, piezoresistive type pressure transducers were installed at different locations in the set up. P1 and P2 were installed for monitoring the supply pressure and transducer P3 was used for monitoring the pressure in the ETC cell. The working pressure range for all these pressure transducers was 0 to 100 bar with
TE D
0.1% accuracy on full scale. With the aim to minimize the error in the measurement of absorption quantity of hydrogen, a differential pressure transducer DP was used between the reference volume and selected volumes. The operating pressure range of the DP was 0 to 5 bar. ‘K’ type thermocouples with an accuracy of ±0.5˚C and time constant of 0.2 s were used at
EP
different locations in the hydrogen supply circuit and in the ETC cell to monitor the hydrogen gas temperature as shown in Fig. 6. A rotary type vacuum pump was used for evacuating the
AC C
system as per the requirement. Pirani gauge attached to the vacuum line was utilized for monitoring the vacuum level in the circuit. Computerized data logger system was used for monitoring and recording all the temperatures and pressure data during the experiments. High purity argon (99.999% purity) was used for volume measurement of different parts of the circuit and for leakage tests.
To calibrate the fabricated ETC measurement cell, pilot tests were performed using different test specimens of known thermal conductivity (stainless steel, brass and PTFE) with dimensions 30 8
ACCEPTED MANUSCRIPT
mm outer diameter, 10 mm inner diameter and 130 mm height to resemble the MH bed. Tests were performed for different source and sink temperature ranges. The maximum error in measurement of thermal conductivity of standard sample is less than ± 5%.
RI PT
Experimental Procedure After filling the ETC cell with material (either in powder form or in compacts form), the cell was heated to a temperature of 200ºC and then evacuated for few hours while maintaining the temperature. This was done to remove moisture and foreign gases present in the cell. The ETC
SC
was then measured under vacuum and different argon pressures for this configuration of bed when the metal alloy was not converted to hydride. To measure the ETC of metal hydrides, first
M AN U
the metal alloy was required to be activated. The activation procedure for La0.8Ce0.2Ni5 has already been discussed [2].
The ETC estimation experiment comprised following steps (Refer Fig. 6): 1. Evacuation of entire setup to 10-3 mbar pressure by opening all valves except v1, v2 and
TE D
v3.
2. The temperature of coolant was set to a fixed value (the coolant temperature was determined by trial and error method to achieve a particular average bed temperature) and coolant circulation is started.
EP
3. Power was supplied to the capsule heater through a DC power supply unit.
AC C
4. All the thermocouples readings were monitored in the ETC cell in intervals of 15 minutes till the steady state is reached.
5. Temperatures T4 to T11 were recorded after attaining steady state. 6. The ETC value for the MH bed at vacuum was estimated using equation (1) as [6]:
ETC =
Q 2π L ( dT d ln r )
(1)
9
ACCEPTED MANUSCRIPT
7. Valves v4 and v11 were closed and hydrogen (99.999% pure) was supplied to volumes V2, V3, V4, V5, V6 and V7 to a predetermined incremental pressure by opening valve v1. After reaching steady state, valve v1 was closed and initial pressures in the volumes V2, V3, V4,
RI PT
V5, V6 and V7 were recorded. 8. Valve v11 was opened to allow hydrogen to the ETC cell. The final pressure values after attaining steady state were recorded. The amount of hydrogen absorbed by the La0.8Ce0.2Ni5 hydride powder was calculated by mass balance on the basis of drop in
SC
hydrogen pressure in the supply volumes.
9. Steps 2 to 8 are repeated to estimate the value of ETC corresponding to different
Results and Discussion
M AN U
hydrogen concentration, pressure and average bed temperatures.
Compaction has proven to be one of the best methods for augmentation of ETC of MH beds. Mixing of graphite flakes has also been implemented by some researchers with different metal hydrides [21-27, 31-36]. It is an established fact that mixing of graphite flakes will result in increase in ETC but the quantity of ENG added is an important parameter as ENG
TE D
flakes need more space as compared to that required for same quantity of metal hydride. The increase in ETC and heat transfer characteristics (and consequently better hydrogen sorption kinetics) may be at a cost of decreased amount of hydrogen absorbed for that volume of MH
EP
bed. Inclusion of metallic structures also results in enhancement of ETC of MH beds again at the cost of decrease in amount of hydrogen absorbed. A common phenomenon that can be observed in the graphs in Figs. 7, 8, 10-12 is that, for the
AC C
same set of operating parameters, the ETC values obtained with CMPGFCu is around 1.5 times of CMPGF and 6 times of LMP bed thus clearly showing the level of augmentation of ETC achieved by different types of compacts. Fig. 7 shows the variation of ETC with pressure for La0.8Ce0.2Ni5 bed with LMP, CMPGF and CMPGFCu. As the solid thermal conductivity shows no variation with pressure, the variation in ETC with pressure is mainly the consequence of variation in gas thermal conductivity, if the filling gas is an inert gas. Thermal conductivity of gas is dependent on the kinetic energy of its molecules and their collisions. In a fixed volume powder bed with a 10
ACCEPTED MANUSCRIPT
certain filling gas, the variation in pressure results in variation in the mean free path of the gas molecules consequently affecting the rate of collisions between the molecules. Hence, for such beds, the variation shown by ETC with change in pressure is similar to the variation shown by the gas thermal conductivity.
RI PT
In case of metal hydride beds, when the bed is exposed to hydrogen, the reaction between solid particles and hydrogen causes significant variation in bed porosity and solid particle geometry consequently affecting the ETC of the bed. Therefore, to evaluate the effect of variation of gas thermal conductivity (with pressure) on the ETC of the bed, initially the ETC
SC
was measured for the non- hydrogenated bed of metal alloy. As shown in Fig. 7a, the variation of ETC with pressure forms a tilted ‘S’ shaped curve. Similar trend of ETC
M AN U
variation with pressure for metal hydride beds have been reported in literature [15-19, 22, 29]. The curve can be divided into three regimes. For low pressure region (from 0.001 to 1.5 bar) the gas pressure does not show any effect on the thermal conductivity of gas as the density of gas present in the voids of the bed, is low and the mean free path of the gas molecule is large thus reducing the number of collisions. This leads to unaffected gas thermal conductivity and consequently the ETC of bed remains almost constant in this region. With
TE D
the increase in pressure from 1.5 bar to 40 bar, the density of gas increases and mean free path of molecule decreases. As a result the variation in pressure in this regime, shows magnified effect on thermal conductivity of gas and hence there is a drastic increase in ETC. This trend continues till the pressure value reaches break away pressure [1]. The effect of
EP
augmentation techniques is clearly visible in the figure. As the bed porosity decreases with compaction thus the apparent solid thermal conductivity increases, causing increase in ETC
AC C
of bed. The variation in thermal conductivity of gas does not seem to be affected by compaction as the variation in ETC with respect to gas pressure remains almost similar for all the three types of beds. As shown in Fig. 5b, the ETC augmentation structure made with copper wire mesh has radial fins. These fins enhance the radial heat conduction. In case of compacts embedded with CWM structure the ETC augmentation achieved due to compaction and addition of graphite flakes is further increased due to the CWM structure. Hence a considerable difference in the increase in ETC values in the cases of CMHGF and CMHGFCu can be observed in Fig. 7b. After break away pressure, further increase in 11
ACCEPTED MANUSCRIPT
pressure will not cause any change in thermal conductivity of gas and hence ETC of bed. The effect of variation in pressure is supplemented with other effects when the gas in the bed is hydrogen instead of non-reactive gas, because with exposure to hydrogen, the metal alloy particles react with it to make metal hydride. There is a swelling effect due to hydrogenation
RI PT
of these particles. With this swelling the contact profile between solid particles changes causing decrease in contact resistance. Thus, although change in pressure has no direct effect on thermal conductivity of solid but due to metal alloy – hydrogen interaction, the ETC amplifies for the same pressure as compared to that with argon (an inert gas). Fig. 7b shows
SC
the variation of ETC with pressure when the filling gas is hydrogen. The trends of ETC variation in the intermediate regime were similar (but larger) to that of argon as filling gas
M AN U
verifying the explanation given for the case of argon. The effect of augmentation was even better when combined with the effect of hydrogenation. The hydrogenation results in decrease in bed porosity and increase in the thermal conductivity. This effect was already gained by compaction which was further amplified by hydrogenation of beds. Table 3 depicts the effects of pressure variation as well as that of ETC augmentation on ETC of all the three types of beds.
TE D
The effect of hydrogen concentration on ETC can be seen as the net effect of the following four phenomena: 1.The reaction between metal alloy particles and hydrogen causes swelling of solid particles and thus increasing contact area between particles which ultimately results in increase in ETC of bed. 2. The increase in hydrogen concentration involves increase in
EP
hydrogen pressure thus the effect of increase in pressure gets involved in the effect of hydrogen concentration on ETC. 3. The swelling of solid particles causes decrease in bed
AC C
porosity as the bed volume is constant. This decrease in porosity is a positive factor for ETC of the MH bed. 4. As the hydrogen atoms get embedded to solid particles, the electronic configurations of solid particles change. The major impact is of variation in bed porosity and hydrogen gas pressure. The effect of variation of hydrogen concentration on ETC has been shown in Fig. 8. Extracts of the data from Fig. 8 has been tabulated in Table 4. The porosities in both the cases of pellets beds are lower as compared to that of powder bed at the beginning of hydrogenation process. With the increase in hydrogen concentration the bed porosity further decreases thus resulting in an increase in ETC. In case of pellets, probably this 12
ACCEPTED MANUSCRIPT
decrease in porosity is more significant than in case of powder bed thus resulting in significant increase in ETC in the pellets beds. Moreover, as can be seen in Fig. 7 of the revised manuscript, the increase in ETC with increase in hydrogen pressure is much lower in case of powder bed. These two combined effects lead to insignificant variation in thermal
RI PT
conductivity in case of powder bed. Also it is observed that the ETC for all the three types of beds increases more sharply from hydrogen concentration of 0 to about 0.1 wt%. From PCI diagram (Fig. 9) at 60˚C, it can be observed that change in hydrogen concentration from 0 to 0.1 wt%, the corresponding equilibrium pressure of the bed is increasing from 1 bar to 50
SC
bar. The maximum increase in ETC is occurring in this pressure range as can be seen from Fig. 7b. Above 0.1 wt% hydrogen concentration, the variation in pressure with hydrogen
M AN U
concentration is not significant. Hence concentration variation has less effect on ETC value in this region. For the same temperature and pressure conditions, hydrogen concentration (in wt%) : ETC (in W m-1 K-1 ) for LMP/CMPGF/CMPGFCu were observed to be 0.8 : 1.22 / 0.76 : 4.39 / 0.78 : 5.67. The values of hydrogen concentration clearly suggest that although there were significant enhancements in the ETC values with augmentation techniques, the decreases in quantities of hydrogen absorption were below 5.2%. The amplitude of effect of
TE D
ETC augmentation on quantity of hydrogen absorption is in conformation with the reported values in literature for similar ENG content and compaction pressure [17-18, 21, 23-27]. Fig. 10 shows the variation of ETC with pressure during hydrogen absorption and desorption processes for powder beds. As shown in the figure, the ETC during desorption is marginally
EP
more than that during absorption at a particular pressure. This can be explained with the help of PCI curve as shown in Fig. 9. The PCI characteristics of La0.8Ce0.2Ni5 hydride was
AC C
performed by authors’ group [28]. The detailed information regarding the experimental set up and methodology adopted etc. are provided elsewhere [28]. At a given pressure the concentration of hydrogen in metal hydride is slightly higher in case of desorption curve compared to absorption curve due to hysteresis. Due to the higher concentration ETC value during desorption is slightly higher than that of absorption. During desorption the reactor (ETC cell) is connected to a sample cylinder which is filled with hydrogen at atmospheric temperature and a particular pressure (which is below the reactor pressure). The metal hydride in the reactor desorbs hydrogen corresponding to the desorption temperature. The 13
ACCEPTED MANUSCRIPT
thermal conductivity is measured when the system (reactor and sample cylinder) achieves an equilibrium pressure. For the next step, the valve v11 is closed, thus disconnecting the ETC cell from the sample cylinder. The sample cylinder is then brought to a lower pressure and again the ETC cell is connected to sample cylinder by opening the valve v11. This process is
RI PT
repeated to obtain lower equilibrium pressures during desorption process.
Temperature variation is reported to have different types of effects on ETC in different studies [5]. In the present work, the average temperature variation seemed ineffective in
SC
bringing about any change in the ETC of MH beds although the effect of augmentation was clearly visible. For an average bed temperature of 30 ºC and 50 bar pressure, ETC values were 1.37 / 5.1 / 7 W m-1 K-1 for LMP/CMPGF/CMPGFCu. As shown in Figs. 11 (a) and 11
M AN U
(b) although the ETC is changing with pressure but the graph between ETC and average temperature remained almost a horizontal line in all the cases except in case of CMPGFCu bed for 10 bar pressure where a slight decrease in ETC is observed when average bed temperature was increased from 50˚C to 70˚C. The reason behind this may be that the increasing effect on ETC caused due to the increasing thermal conductivity of hydrogen with temperature might be balanced by the decreasing effect caused due to decrease in thermal
TE D
conductivity of the solid particles.
The effect of ETC augmentation is clear from Fig. 12, which shows the radial (spatial) temperature profile for the MH beds at steady state for different configurations. As can be
EP
seen that the slope of the radial temperature profile reduces for CMPGF and it further reduces for CMPGFCu beds for both the MH’s when compared to that of LMP beds. The
AC C
slopes of radial temperature profiles for LMP/CMPGF/CMPGFCu with argon as filling gas were -5255.17 / -3955.17 / -3312.07 K/m whereas with hydrogen these were -3670.69 / 1036.21 / -731.03 K/m respectively. This clearly shows the improvement in heat transfer characteristics with the augmentation of ETC. Since filling of argon gas will result into comparatively lower values of ETC for these beds as compared to those with hydrogen as filling gas, the slopes for the cases of hydrogen are lower than those for argon for the respective beds.
14
ACCEPTED MANUSCRIPT
The ETC values obtained in the present work with the two types of compacts are close to those reported in literature for the similar ENG content, compaction pressure and operating parameters [21, 23-27]. The uncertainties / accuracies of different measuring instruments used for this work are given in Table 5. Using these values the maximum uncertainty in the
RI PT
estimation of ETC values was found to be ± 5.1%. Conclusion
The ETC augmentations for metal hydride beds of La0.8Ce0.2Ni5 are attempted using different
SC
types of compacts. The effect of compaction and mixing of graphite flakes resulted in the increase of ETC to about four times that of powder bed while the increase in ETC was found
M AN U
to be around six times that of powder bed for the CMPGFCu configuration. The decrease in quantity of hydrogen absorbed by La0.8Ce0.2Ni5 hydride due to ETC augmentation was very less (< 6%) as the mass fraction and hence the volume fraction occupied by graphite flakes was intentionally kept low. The inclusion of copper wire mesh resulted in continuous heat conduction path in the radial direction as the mesh consists of seven radial porous fins. Thus inclusion of copper wire mesh lead to enhancement in the heat transfer in radial direction and
TE D
consequently in the ETC. Using high pressure for compaction with higher wt% of graphite flakes for making metal hydride compacts will also increase the conduction path in the radial direction, however it reduces the rate of hydrogen diffusion in the metal hydride beds. Hence, it can be concluded that the compacts made with mixture of metal alloy powder and graphite
EP
flakes and embedded with copper wire mesh structure is a better method of ETC augmentation for MH beds with minimal loss in hydrogen absorption quantity.
AC C
Nomenclature CWM
Copper wire mesh structure
ENG
Expanded natural graphite
ETC
Effective thermal conductivity, W m-1 K-1
L
Length of metal hydride bed, m 15
ACCEPTED MANUSCRIPT
Loose metal alloy / composite powder
CMPGF
Compacts of metal alloy / composite powder and graphite flakes
CMPGFCu
Compacts of metal alloy / composite powder, graphite flakes and embedded with copper wire mesh structure Heat transfer rate, W
R
Radius of metal hydride bed, m
T
Temperature, K
M AN U
SC
Q
RI PT
LMP
Acknowledgements
One of the authors, Dr. E. Anil Kumar, wishes to sincerely acknowledge the Science and Engineering Research Board, Department of Science and Technology, Government of India
References
TE D
for partially funding the work vide Project No. SERB/F/0810/2013-14.
1. Madaria Y, Anil Kumar E, Murthy SS. Effective thermal conductivity of reactive packed
EP
beds of hydriding materials. App Therm Eng 2016;98:976-90. 2. Madaria Y, Anil Kumar E. Effect of heat transfer enhancement on the performance of
AC C
metal hydride based hydrogen compressor. Int J Hydrogen Energy 2016;41:3961-73. 3. Sun DW, Song JD. Theoretical descriptions and experimental measurements on the effective thermal conductivity in metal hydride powder beds. J Less-Common Met 1990;160:387-95. 4. Sun DW, Deng SJ. Theoretical descriptions and experimental measurements on the effective thermal conductivity of metal hydride powder beds. J Less-Common Met 1990;160:387-95. 5. Kempf A, Martin W. Measurement of the effective thermal properties of TiFe0.85Mn0.15 and its hydrides, Int J Hydrogen Energy 1986;11:107-16. 6. Suda S, Kobayashi N, Yoshida K, Ishido Y, Ono S. Experimental measurements of thermal conductivity. J Less-Common Met 1980;74:127-36. 16
ACCEPTED MANUSCRIPT
7. Dedrick DE, Kanouff MP, Replogle BC, Gross KJ. Thermal properties characterization of sodium alanates. J Alloy Compd 2005;389:299-305. 8. Ishido Y, Kawamura M, Ono S. Thermal conductivity of magnesium-nickel hydride powder beds in a hydrogen atmosphere. Int J Hydrogen Energy 1982;7:173-82.
RI PT
9. Gustafsson SE. Transient plane source techniques for thermal conductivity and thermal diffusivity measurements of solid materials. Rev Sci Instrum 1991;62:797-804. 10. Kapischke J, Hapke J, Measurement of effective thermal conductivity of a Mg-MgH2 packed bed with oscillating heating. Exp Therm Fluid Sci 1998;17:347-55.
SC
11. Y. Asakuma, S. Miyauchi, T. Yamamoto, H. Aoki, T. Muira, Homogenization method for effective thermal conductivity of metal hydride bed, Int J Hydrogen Energy 2004;29:209–16. 12. Madaria Y, Anil Kumar E, Maiya MP, Murthy SS. Simulation of effective thermal conductivity of metal hydride packed beds. Heat Transf Eng 2016;37:616-24.
M AN U
13. Anil Kumar E, Maiya MP, Murthy SS. Measurement and analysis of effective thermal conductivity of MmNi4.5Al0.5 hydride bed. Ind Eng Chem Res 2011;50:12990-9. 14. Inoue S, Yushi I, Yukihiko M. Drastic enhancement of effective thermal conductivity of a metal hydride packed bed by direct synthesis of single-walled carbon nanotubes. Int J Hydrogen Energy 2012;37:1836-41.
TE D
15. Suda S, Komazaki Y, Kobayashi N. Effective thermal conductivity of metal hydride beds. J LessCommon Met 1983;89:317-24. 16. Nagel M, Komazaki Y. Effective thermal conductivity of a metal hydride bed augmented with a copper wire matrix. J Less-Common Met 1986;120:35-43.
EP
17. Bershadsky E, Josephy Y, Ron M. Permeability and thermal conductivity of porous metallic matrix hydride compacts. J Less-Common Met 1989;153:65-78. 18. Ron M, Bershadsky M, Josephy Y. The thermal conductivity of porous metal matrix hydride compacts. J Less-Common Met 1991;172-174:1138-46.
AC C
19. Suda S, Komazaki Y. The effective thermal conductivity of a metal hydride bed packed in a multiple-waved sheet metal structure. J Less-Common Met 1991;172-174:1130-7. 20. Ron M, Bershadsky E, Josephy Y. Thermal conductivity of PMH compacts, measurements and evaluation. Int J Hydrogen Energy 1992;17:623-30. 21. Kim KJ, Montoya B, Razani A, Lee K-H. Metal hydride compacts of improved thermal conductivity. Int J Hydrogen Energy 2001;26:609-13. 22. Yan M-Y, Sun F, Liu X-P, Ye J-H. Effects of compaction pressure and graphite content on hydrogen storage properties of Mg(NH2)2-2LiH hydride. Int J Hydrogen Energy 2014;39:1965661. 17
ACCEPTED MANUSCRIPT
23. Sanchez RA, Klein H-P, Groll M. Expanded graphite as heat transfer matrix in metal hydride beds. Int J Hydrogen Energy 2003;28:5156-27. 24. Pohlmann C, Rontzsch L, Kalinichenka S, Hutsch T, Weibgarber T, Kieback B.
RI PT
Hydrogen storage properties of compacts of melt-spun Mg90Ni10 flakes and expanded natural graphite. J Alloy Compd 2011;509S:S625-8. 25. Pohlmann C, Rontzsch L, Kalinichenka S, Hutsch T, Kieback B. Magnesium alloy-
graphite composites with tailored heat conduction properties for hydrogen storage applications. Int J Hydrogen Energy 2010;35:12829-36. 26. Pohlmann C, Rontzsch L, Heubner F, Weibgarber T, Kieback B. Solid-state hydrogen
SC
storage in Hydralloy-graphite composites. J Power Sources 2013;231:97-105.
M AN U
27. Shim J-H, Park M, Lee YH, Kim S, Im YH, Suh J-Y, et al. Effective thermal conductivity of MgH2 compacts containing expanded natural graphite under a hydrogen atmosphere. Int J Hydrogen Energy 2014;39:349-55. 28. Sharma VK, Anil Kumar E. Studies on La based intermetallic hydrides to determine their suitability in metal hydride based cooling systems. Intermetallics 2015;57:60-7. 29. Muthukumr P, Maiya MP, Srinivasa Murthy S. Experiments on a metal hydride based hydrogen compressor. Int J Hydrogen Energy 2005;30:879-92. 30. Kim JK, Park IS, Kim KJ, Gawlik K. A hydrogen-compression system using porous metal hydride pellets of LaNi5-xAlx. Int J Hydrogen Energy 2008;33:870-7.
TE D
31. Chaise A, Rango P.de, Marty Ph, Fruchart D, Miraglia S, Olives R, Garrier S. Enhancement of hydrogen sorption in magnesium hydride using expanded natural graphite. Int J Hydrogen Energy 2009;34:8589-96.
EP
32. Y. Ishido, M. Kawamura, S. Ono. Thermal conductivity of magnesium-nickel hydride powder beds in a hydrogen atmosphere. Int J Hydrogen Energy 1982;7:173-82.
AC C
33. Yasuda N, Tsuchiya T, Okinaka N, Akyama T. Thermal conductivity and cycle characteristics of metal hydride sheet formed using aramid pulp and carbon fiber. Int J Hydrogen Energy 2013;38:1657-61. 34. Wang H, Prasad AK, Advani SG. Hydrogen storage system based on hydride materials with enhanced thermal conductivity. Int J Hydrogen Energy 2012;37:290-8. 35. Zamengo M, Ryu J, Kato Y. Thermochemical performance of magnesium hydroxide – expanded graphite pellets for chemical pump. App Therm Eng 2014;64:339-47. 36. Dieterich M, Pohlmann C, Burger I, Linder M, Rontzsch L. Long – term cycle stability of metal hydride – graphite composite. Int J Hydrogen Energy 2015;40:16375-82.
18
ACCEPTED MANUSCRIPT
Value
Surface area per unit volume
8333 m2 m-3
Density
5.6 g cm-3
Thermal conductivity
392 W m-1 K-1
Wire diameter
0.12 mm
Purity
99.9%
AC C
EP
TE D
M AN U
SC
Parameter
RI PT
Table 1 Physical properties of copper wire mesh used for making compacts
ACCEPTED MANUSCRIPT
Bed porosity (%)
LMP
3.79
58
CMPGF
4.94
38
CMPGFCu
5.47
34
AC C
EP
TE D
M AN U
SC
Packing density (g.cm-3)
RI PT
Table 2 Packing density and bed porosity of different types of beds
ACCEPTED MANUSCRIPT
Table 3 Variation of ETC with pressure for different filling gases Pressure
ETC with Ar as filling gas -1
(bar)
ETC with H2 as filling gas
-1
(W m-1 K-1)
(W m K ) CMPGF
CMPGFCu
LMP
CMPGF
CMPGFCu
0.01
0.58
1.9
3.01
0.58
1.9
3.0
1.5
0.62
2.1
3.3
0.59
2.0
3.3
10
0.68
2.6
3.6
0.69
2.9
4.28
40
0.93
3.1
4.05
1.29
4.63
6.2
60
0.95
3.2
4.22
1.3
4.7
6.8
70
0.95
3.2
4.22
1.3
4.7
6.8
AC C
EP
TE D
M AN U
SC
RI PT
LMP
ACCEPTED MANUSCRIPT
Table 4 Variation of ETC with hydrogen concentration for different types of beds
RI PT
ETC (W m-1 K-1) CMPGF CMPGFCu 1.0 1.33 2.64 3.42 4.12 4.94 4.57 6.31 4.73 6.85
TE D
M AN U
SC
LMP 0.59 0.71 1.19 1.27 1.31
EP
0.5 bar 5 bar 10 bar 20 bar 50 bar
Hydrogen concentration (wt%) LMP CMPGF PMPGFCu 0.02 0.01 0.01 0.08 0.069 0.07 0.29 0.25 0.23 1.38 1.29 1.25 1.45 1.43 1.41
AC C
Pressure
ACCEPTED MANUSCRIPT
Table 5 Uncertainties of various measured parameters for estimating ETC S. No.
Measurement
Device
Accuracy
Minimum
Maximum
quantity
uncertainty
measured Bed temperature
K type thermocouple
± 0.5 K
2.
Hydrogen gas
Piezoresistive type
± 0.000625 bar
1 bar
±0.0625%
pressure
transducer
Temperature of
K type thermocouple
± 0.5 K
283 K
±0.177%
ETC cell
± 0.5 W
30 W
±1.66%
3.
4.
Heat transfer rate estimation
SC
thermostatic fluid
293 K
±0.171%
RI PT
1.
Mass of alloy
Digital weighing balance
±0.1g
25 g
±0.4%
6.
Length
Digital vernier caliper
±0.1mm
5 mm
±2%
AC C
EP
TE D
M AN U
5.
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 1 Schematic diagram of ETC cell
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
Fig. 2 Components of ETC cell
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
Fig. 3 XRD Pattern of La0.8Ce0.2Ni5 alloy [28]
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
Fig. 4 Steps in compact making
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 5 (a) Compacts of La0.8Ce0.2Ni5 powder mixed with graphite flakes
AC C
EP
TE D
M AN U
(b) Compacts with copper wire mesh
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 6 Experimental set up for ETC measurement
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
(a)
(b) Fig. 7 Variation of effective thermal conductivity with pressure (a) Filling gas: Argon (b) Filling gas: Hydrogen
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
Fig. 8 Variation of effective thermal conductivity with hydrogen concentration
ACCEPTED MANUSCRIPT
100
Pressure (bar)
RI PT
10
1
SC
10°C 20°C 40°C 60°C 80°C
M AN U
Absorption Desorption 0.1 0
0.2
0.4
0.6 0.8 1 Storage Capacity (wt%)
1.2
1.4
1.6
AC C
EP
TE D
Fig. 9 Pressure concentration Isotherms for La0.8Ce0.2Ni5 hydride [28]
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
Fig. 10 Variation of effective thermal conductivity with pressure during hydrogen absorption and desorption
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
(a)
(b) Fig. 11 Variation of effective thermal conductivity with average bed temperature (a) Hydrogen pressure = 10 bar (b) Hydrogen pressure = 50 bar
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
Fig. 12 Radial bed temperature profile of MH bed
ACCEPTED MANUSCRIPT
Highlights of the research article
AC C
EP
TE D
M AN U
SC
RI PT
• ETC of La0.8Ce0.2Ni5 hydride bed was measured. • Two different types of pellets were developed for augmentation of ETC. • Pellets with embedded copper wire mesh were found to be better option for ETC augmentation.