Phase evolution of Li2ND, LiD and LiND2 in hydriding/dehydriding of Li3N

Article abstract of DOI:10.1016/j.jallcom.2007.02.149

Neutron and synchrotron studies have been performed on in situ hydriding of Li₃N. Commercial Li₃N is composed of α phase (～70 wt.%) and β phase (～30 wt.%). We have performed experiments to convert β → α, and studied in situ deuteration using neutron diffraction experiments. We found concurrent phase evolution of Li₂ND, LiD, and LiND₂. Mass percentages of the phases evolved as a function of time and temperature have been quantified using General Structure Analysis System (GSAS) refinement of the neutron diffraction data. The problem of formation of the stable LiD is discussed in light of decreasing of the amount of LiD phase when the temperature is increased from 200 to 320 °C during dehydriding, and in addition the concentration of Li₂ND phase increased at this temperature. Lattice parameters, volume changes, phase evolutions in wt.% as a function of temperature and time are presented.

Full text of DOI:10.1016/j.jallcom.2007.02.149

Journal of Alloys and Compounds 446–447 (2007) 363–367
Phase evolution of Li2ND, LiD and LiND2
in hydriding/dehydriding of Li3N
Wen-Ming Chiena, Joshua Lamba, Dhanesh Chandraa,∗
,
Ashfia Huqb, James Richardson Jr.c, Evan Maxeyc
a Metallurgical Materials Engineering, MS 388, University of Nevada, Reno, NV 89557, United States
b Spallation Neutron Source, Oak Ridge National Laboratory, Oak Ridge, TN 37831, United States
c Intense Pulsed Neutron Source, Argonne National Laboratory, Argonne, IL 60439, United States
Received 17 February 2007; received in revised form 22 February 2007; accepted 22 February 2007
Available online 4 March 2007
Abstract
Neutron and synchrotron studies have been performed on in situ hydriding of Li3N. Commercial Li3N is composed of ␣ phase (∼70 wt.%) and
␤ phase (∼30 wt.%). We have performed experiments to convert ␤ → ␣, and studied in situ deuteration using neutron diffraction experiments. We
found concurrent phase evolution of Li2ND, LiD, and LiND2. Mass percentages of the phases evolved as a function of time and temperature have
been quantified using General Structure Analysis System (GSAS) refinement of the neutron diffraction data. The problem of formation of the stable
LiD is discussed in light of decreasing of the amount of LiD phase when the temperature is increased from 200 to 320 ◦C during dehydriding, and
in addition the concentration of Li2ND phase increased at this temperature. Lattice parameters, volume changes, phase evolutions in wt.% as a
function of temperature and time are presented.
© 2007 Elsevier B.V. All rights reserved.
Keywords: Li2NH; Li3N; LiNH2; Neutron diffraction; Phase analysis
1. Introduction
To increase the equilibrium pressure, Luo [2] showed that Mg
additions increase the plateau pressures, Eq. (3).
Several studies have indicated complex alkali metal based
hydrides, such as, lithium amide (LiNH2) and lithium imide
(Li2NH) are particularly attractive as potential candidates for
hydrogen storage due to their relatively high theoretical gravi-
metric hydrogen density (e.g., LiNH2 alone has 6.5 wt.%
hydrogen). Chen et al. [1] pioneered the progressive hydrogena-
tion of lithium nitride, Li3N, which leads to the absorption of
the theoretical 11.5 wt.% hydrogen by the reaction in Eq. (1):
2LiNH2 + MgH2 ⇔ Li2Mg(NH)2
(32 bar at 200 ◦C) (3)
Chen et al. [1] also suggested that full reversal to Li3N
may be possible at higher temperatures (>400 ◦C) as per
Eq. (4), which is also confounded with competing side-
reactions leading to ammonia formation as in Eqs. (5) and (6)
[3].
Li2NH + LiH ⇔ Li3N + H2
Li3N + 2H2 ⇔ Li2NH + LiH + H2 ⇔ LiNH2 + 2LiH
(total capacity : 11.5 wt.% H2)
(not easily reversible, ∼ 0.01 bar at 255 ◦C, 5 wt.% H2, ꢀH
(1)
= 165 kJ/mol)
(4)
(5)
(6)
If only imides are considered, then the following Eq. (2) is valid.
Li2NH + H2 ⇔ LiNH2 + LiH
2LiNH2 ⇔ Li2NH + NH3
(3.7 wt.% NH3, ΔH = +84 kJ/mol NH3)
(reversible at ∼ 1 bar at 285 ◦C, 6.5 wt.% H2, ΔH
= 45 kJ/mol)
(2)
NH3 + LiH → 2LiNH2 + H2
(5.8 wt.% H2, ΔH = −39 kJ/mol NH3)
∗
Corresponding author. Tel.: +1 775 784 4960; fax: +1 775 784 4316.
E-mail address: dchandra@unr.edu (D. Chandra).
0925-8388/$ – see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.jallcom.2007.02.149

364
W.-M. Chien et al. / Journal of Alloys and Compounds 446–447 (2007) 363–367
Fig. 1. Schematic of the in situ deuteriding/de-deuteriding experimental setup.
However, these hydrogenation/dehydrogenation reactions have
not been shown to be reversible or to have adequate hydrogen
release kinetics at moderate temperatures.
Issues with ammonia evolution were reported by Ichikawa et
al. [4] who indicated that the NH3 release is controllable; the
NH3 evolved during a first-order reaction, then is consumed in
a secondary reaction shown in Eq. (6); the kinetics are very fast
for this reaction. The lithium-based complex hydrides have great
potential as materials for fuel cell, on-board vehicular, and other
applications [5–7].
2. Experimental
A lithium nitride sample (Li3N, 80 mesh) was obtained from Sigma–Aldrich
in the form of a powder. Time-of-flight (TOF) NPD data was collected for
the Li3N sample using the General Purpose Powder Diffractometer (GPPD)
at the Intense Pulsed Neutron Source (IPNS), Argonne National Laboratory.
Diffraction data were collected on all detector banks. The sample was loaded
in the Inconel alloy pressure sample holder in a helium-filled re-circulating
glove-box. The in situ deuteriding/de-deuteriding experimental setup is shown
in Fig. 1. Deuterium gas was supplied at 2 bar during the deuteriding experiment.
Rietveld refinements were performed using the General Structure Analysis Sys-
tem (GSAS) computer software [20,21]. To de-deuteride, for the results shown
in Fig. 3, we used active evacuation via a Varian V-70 turbo pump system, backed
up with a dry mechanical pump. The NPD data were acquired on two banks of
detectors at IPNS (Argonne National Laboratories), every 15 min. The phases
were identified first by standard search methods, and then the compositional
changes were tracked using the GSAS program.
Commercial Li3N powder is a two-phase mixture of ␣-Li3N
and ␤-Li3N. The structure of the ␣ phase is hexagonal with
˚
P6/mmm space group with lattice constants a = 3.648(1) A and
˚
c = 3.875(1) A (Z = 1) [8,9]. The ␤ phase is a high-pressure
˚
phase, and is a hexagonal structure with a = 3.552(1) A and
˚
c = 6.311(3) A (P63/mmc, Z = 2) [10]. The low-temperature
3. Results and discussions
structural behavior studies of the ␣-Li3N and ␤-Li3N phases
have been performed by Huq et al. [11] and Chien et al. [12]
In this study, we performed both diffraction studies on the
pre-treated ␣ phase Li3N sample as well as ␣ + ␤ Li3N com-
mercial powders. At room temperature, the NPD data show
that the commercial Li3N sample contains two different phases:
65 wt.% ␣-Li3N phase and 35 wt.% ␤-Li3N phase. Isothermal
in situ deuteriding NPD data was taken at 200 ◦C at 2 bar D2
pressure. Isothermal evolution (wt.%) of different phases as a
function of time is shown in Fig. 2. Initially, as the tempera-
ture was equilibrated to 200 ◦C, the only change observed in the
diffraction pattern was ␤-Li3N transforming to ␣-Li3N phase
(region 1); the ␣-Li3N phase increases from 65 to 79 wt.% with
a corresponding decrease in the mass of the ␤ phase. In region
2, the hydride phases Li2ND and LiD begin to appear. Con-
currently, there is a decrease in the mass of the ␣-Li3N phase,
suggesting that the ␣-Li3N is converting to Li2ND and LiD. The
␤-Li3N phase concentration continues to decrease in region 2.
The evolution of the amide LiND2 phase was observed after
∼145 min, in region 3. It can be noted that in region 3, con-
centrations of LiND2 and LiD are greater than that of Li2ND,
using neutron powder diffraction (NPD). The structure of Li2NH
´
˚
was reported as cubic (Fm-3m) with a = 5.0742(2) A by Nori-
take et al. [13] using synchrotron X-ray diffraction (XRD), and
by Ohoyama et al. [14] using NPD. Recently, Balogh et al.
[15] reported the crystal structure and phase transformation of
deuterated lithium imide (Li2ND) using both NPD and XRD.
The crystal structure of LiNH2 was determined by Juza and
Opp [16], and re-determined by Jacobs and Juza (single-crystal
XRD) [17], Miwa et al. [18], and Yang et al. by NPD [19] as
˚
tetragonal structure with a = 5.03442 and c = 10.25558 A (space
group I-4).
Li-based hydrides show a high capacity for reversible hydro-
gen storage. However, understanding phase formation behaviors
during temperature change is also important for the hydrogen
storage application. In this study, we report results of NPD
studies, such as phase evolution of the amide, imide, LiD, dur-
ing hydrogenation of Li3N. Also the effect of temperature on
deuteriding and de-deuteriding of Li3N.

W.-M. Chien et al. / Journal of Alloys and Compounds 446–447 (2007) 363–367
365
for the deuteration and de-deuteration in Fig. 3. Deuteriding
was performed at room temperature, 150, 200, and 250 ◦C. As
expected, the ␣-Li3N phase is stable up to 150 ◦C. As temper-
ature was increased from 150 to 200 ◦C, (from ∼19 to 30 h)
the hydride phases, Li2ND, LiD, and LiND2 form simultane-
ously under these conditions (2 bar of deuterium pressure). Also
note that as the temperature is raised there is time lag to equi-
librate. The amount of Li2ND phase increases to ∼19 wt.%,
the LiD phase to ∼21 wt.%, and the LiND2 phase ∼14 wt.%
after isothermally holding temperature at 200 ◦C for 9 h. From
Fig. 3, it can seen that the ␣-Li3N phase concentration decreases
rapidly in the 200 and 250 ◦C regions at the deuteriding region,
and only ∼10 wt.% is residual after 12 h of isothermal holding at
200 ◦C. In the 250 ◦C deuteriding region, initially, the amount of
the Li2ND phase formed increases rapidly, and then decreases
due to the formation of the LiND2 and LiD phases. It can also
be noted that the formation of LiD phase continues to increase
during deuteration between 200 and 250 ◦C, while the Li2ND
decreases and LiND2 increases with a possible saturation at 40 h;
with 25 wt.% of Li2ND, 37 wt.% LiD, and 28 wt.% LiND2. In
the 200 ◦C region, there is a tendency to form Li2ND and LiD
phases in preference to LiND2, similar to what we observed in
Fig. 2. The total amount of hydrogen storage is calculated to be
7.26 wt.% at the end of 44 h.
Fig. 2. Isothermal in situ deuteriding NPD results (at 200 ◦C and 2 bar D2 pres-
sure of the commercial Li3N sample (␣-Li3N and ␤-Li3N) show the different
amounts of phase formation as a function of time.
indicating either direct formation of LiND2 from Li3N or rapid
transformation of Li2ND → LiND2 during these multiple con-
current transformations. It does appear that, as the amount of
␣-Li3N phase decreases, there is a corresponding increase in
the amount of LiND2 formed. The reactions in regions 1–3 at
200 ◦C are summarized in Table 1.
Another in situ deuteriding/de-deuteriding experiment was
performed by using the pre-treated Li3N sample which
contained only the ␣-Li3N phase; the ␤-Li3N to ␣-Li3N transfor-
mation occurs readily above 250 ◦C. The results are presented
The neutron data indicate a total storage of 7.26 wt.% of
hydrogen after 44 h at 255 ◦C. Our recent work at 255 ◦C showed
that at 2 bar, the material has an effective equilibrium capacity
of 8.7 wt.%. Four grams of material was used in the NPD stud-
ies, indicating that 0.29 g of hydrogen was stored, or 83% of the
material’s potential capacity. The deuteration wt.% capacity of
the materials was converted to hydrogen capacity.
Table 1
The overall hydrogen absorption reaction of Li3N can be
obtained from Eq. (1) as:
Summary of isothermal transformations in each region at 200 ◦C
Region 1
Region 2
Region 3
Up to ∼100 min
100–145 min
145–650 min
Li3N + 2H2 ↔ LiNH2 + 2LiH
␤-Li3N → ␣-Li3N
␤-Li3N → ␣-Li3N
␤-Li3N → ␣-Li3N
␣-Li3N → Li2ND + LiD
Li2ND → LiND2 (rapid)
␣-Li3N → Li2ND + LiD
The Gibbs free energy of the ongoing reaction in our case at a
hydrogen pressure of 2 bar, can be obtained using the following
Fig. 3. The phase evolutions of the in situ deuteriding and de-deuteriding of the pure ␣-Li3N sample as a function of temperature and time.

366
W.-M. Chien et al. / Journal of Alloys and Compounds 446–447 (2007) 363–367
equation:
ꢀG = ꢀG◦ + RT ln KP=2 bar
= −RT ln KP at x= 7.26 wt.% + RT ln KP=2 bar
(7)
As Li3N, LiNH2, and LiH are all solid species, their activities
are assumed to be equal to 1. The equilibrium constant for this
reaction is estimated to be:
aLiNH a2
1
pH2
2
LiH
KP =
=
H2
,
aLi Np2
3
2
and if the imposed pressure is 2 bar then
ꢀ
ꢁ
ꢀ
ꢁ
1
pH2
1
pH2
ꢀG = −RT ln
+ RT ln
(8)
2
2
p
eq
p=2 bar
Our neutron data showed ∼7.26 wt.% hydrogen (Fig. 3) at
250 ◦C, whereas our equilibrium absorption isotherms, using
Li3N as starting material taken at 255 ◦C, showed that at 2 bar
hydrogen pressure there should be ∼8.7 wt.% hydrogen stored
(Fig. 4). Thus the absorption NPD and isotherm show a discrep-
ancy of 1.44 wt.% hydrogen; although there is a small difference
in temperature at which these two data sets were taken. From
our isotherm (Fig. 4) we can obtain the absorption equilibrium
pressure (Peq) of ∼0.8 bar at 7.26 wt.% hydrogen. This yields
ꢀGEq. (8) = −8.0 kJ/mol for reaction (1), indicating that the reac-
Fig. 4. Pressure-composition hydrogen absorption isotherm of Li3N at 255 ◦C.
tion would continue to absorb the additional 1.4 wt.% hydrogen
if given more time. However, we started to evacuate the sys-
tem prematurely before that remaining 1.4 wt.% hydrogen was
absorbed in the sample.
De-deuteriding was started after 44 h at 250 ◦C. In this 250 ◦C
region, the amount of the Li2ND phase increases rapidly to
Fig. 5. The unit-cell volumes of (a) ␣-Li3N, (b) LiD and (c) LiND2 phases vs. temperature.

W.-M. Chien et al. / Journal of Alloys and Compounds 446–447 (2007) 363–367
367
50 wt.% with a corresponding decrease of LiND2 phase to
10 wt.% in 5 h. There is a much less pronounced decrease in the
amount of LiD phase at this temperature; only 5 wt.% decrease
observed. After 48 h, the temperature was decreased to 200 ◦C
during de-deuteration. At this temperature the amount of Li2ND
(54%), LiD (31%), LiND2 (10%) remained constant up to 65 h.
Chen et al. [1] reported a rather rapid increase in hydride phase
formation at 200 ◦C during de-hydriding, but we did not observe
this behavior. We acknowledge that if we had raised the temper-
ature by ∼10 K; we probably would have observed the changes
that Chen et al. [1] had reported. It can be noted that the amount
of hydrogen desorbed increased rapidly for about 2 h or so at
320 ◦C which is perhaps what Chen et al. observed. Increas-
ing the temperature to 320 ◦C rapidly increased the amount of
Li2ND (from 54% to 76%) and decreased the amount of LiD
(from 31% to 20%). No LiND2 phase was measurable after 68 h;
perhaps the intensity of its Bragg peaks were too low for detec-
tion above the background level. The amount of the LiD phase
decreased when the de-deuteriding temperature was increased
to 320 and 350 ◦C, and the amount of the Li2ND phase increased
as the temperature was increased during de-deuteriding. There
is no LiND2 phase observed upon de-deuteriding above 320 ◦C.
The changes in unit-cell volumes were measured at different
temperatures during deuteration in equilibrated samples at dif-
ferent temperatures. In this experiment, the deuterated sample
was heated to 250 ◦C, the sample holder was then sealed, and
the supply of deuterium to the sample was shut off. The sample
was then cooled down from 250 to 100 ◦C, and then reheated
from 100 to 300 ◦C to measure the unit-cell volumes. The vol-
ume expansions as a function of temperatures of the ␣-Li3N,
LiD, and LiND2 phases are shown in Fig. 5, and tabulated in the
insert table. All these phases show an increase in unit-cell vol-
umes, asexpected. TheBraggpeaksoftheLi2NDphasewerenot
observed at these temperatures. The volume of the Li3N phase
to decrease slightly when the temperature is increased from 250
to 320 ◦C during de-deuteriding.
Acknowledgements
We thank the U.S. Department of Energy for program sup-
port through the US DOE Metal Hydride Center of Excellence
(MHCoE) at Sandia National Laboratories, Livermore, CA.
Argonne National Laboratory’s work was supported by the U.S.
Department of Energy, Office of Science, Basic Energy Sci-
ences, under contract DE-AC02-06CH11357.
References
[1] P. Chen, Z. Xiong, J. Luo, J. Lin, K. Tan, Nature 21 (2002) 302–304.
[2] W. Luo, J. Alloys Compd. 381 (2004) 284–287.
[3] F.E. Pinkerton, J. Alloys Compd. 400 (2005) 76–82.
[4] T. Ichikawa, N. Hanada, S. Isobe, H.Y. Leng, H. Fujii, J. Phys. Chem. B
108 (2004) 7887–7892.
[5] L. Schlapbach, A. Zu¨ttel, Nature 414 (2001) 353.
[6] D. Chandra, J. Petrovic, R.G. Bautista, A. Imam, Overview of the Advanced
Materials for Energy Conversion – III, Symposium in Honor of Drs. Gary
Sandrock, Louis Schlapbach, and Seijirau Suda For Lifetime Achievements
in Metal Hydride Research and Development; D. Chandra, J. Petrovic, R.
Bautista, A. Imam, Advanced Materials for Energy Conversion III, Pro-
ceeding of the TMS Annual Symposium in San Antonio, TX, TMS Press,
vol. III, 383 pages, March 2006, pp. 3–7, ISBN: 978-0-87339-610-3.
[7] D. Chandra, J.J. Reilly, R. Chellappa, J. Met. 58 (2006) 26–32.
[8] A. Rabenau, H. Schulz, J. Less-Common Met. 50 (1976) 155–159.
[9] D.H. Gregory, P.M. O’Meara, A.G. Gordon, J.P. Hodges, S. Short, J.D.
Jorgensen, Chem. Mater. 14 (2002) 2063–2070.
[10] H.J. Beister, S. Haag, R. Kniep, K. Strossner, K. Syassen, Angew. Chem.
Int. Ed. Engl. 27 (8) (1988) 1101–1103.
[11] A. Huq, J.W. Richardson, E.R. Maxey, D. Chandra, W. Chien, J. Alloys
Compd. 436 (2007) 256–260.
[12] W. Chien, D. Chandra, A. Huq, J.W. Richardson, E.R. Maxey, S. Fakra, M.
Kunz, MS&T 2006: Fund. Charact. 1 (2006) 501–507.
[13] T. Noritake, H. Nozaki, M. Aoki, S. Towata, G. Kitahara, Y. Nakamori, S.
Orimo, J. Alloys Compd. 393 (2005) 264–268.
[14] K. Ohoyama, Y. Nakamori, S. Orimo, K. Yamada, J. Phys. Soc. Jpn. 74 (1)
(2005) 483–487.
[15] M.P. Balogh, C.Y. Jones, J.F. Herbst, L.G. Hector Jr., M. Kundrat, J. Alloys
Compd. 420 (2006) 326–336.
3
˚
increased from 44.59 to 45.2 A , the LiD phase from 67.76 to
3
3
˚
˚
69.58 A , and the LiND2 phase from 260.79 to 266.85 A , from
100 to 300 ◦C, respectively.
4. Summary and conclusions
[16] R. Juza, K. Opp, Z. Anorg. Allgem. Chem. 266 (1951) 313–324.
[17] H. Jacobs, R. Juza, Z. Anorg. Allgem. Chem. 391 (3) (1972) 271–279.
[18] K. Miwa, N. Ohba, S. Towata, Phys. Rev. B71 (2005) 195109-1.
[19] J.B. Yang, X.D. Zhou, Q. Cai, W.J. James, W.B. Yelon, Appl. Phys. Lett.
88 (2006) 041914-1.
[20] A.C. Larson, R.B. Von Dreele, General Structure Analysis System (GSAS),
Los Alamos National Laboratory Report LAUR 86-748 (2000).
[21] B.H. Toby, J. Appl. Cryst. 34 (2001) 210–213.
Neutron scattering studies of deuteration of ␣-Li3N showed
that the different amounts of LiD, and LiND2, and Li2ND phases
form during isothermal heating at different temperatures. The
total amount of hydrogen storage is calculated to be 7.26 wt.%
during deuteriding at 250 ◦C. The stable LiD phase was found