We present the pressure-induced phase/chemical changes of lithium peroxide (Li2O2) to 63 GPa using diamond anvil cells, confocal micro-Raman spectroscopy, and synchrotron x-ray diffraction. The Raman data show the emergence of the major vibrational peaks associated with O2 above 30 GPa, indicating the subsequent pressure-induced reversible chemical decomposition (disassociation) in dense Li2O2. The x-ray diffraction data of Li2O2, on the other hand, show no dramatic structural change but remain well within a P63/mmc structure to 63 GPa. Nevertheless, the Rietveld refinement indicates a subtle change in the structural order parameter z of the oxygen position O (13, 23, z) at around 35 GPa, which can be considered as a second-order, isostructural phase transition. The nearest oxygen-oxygen distance collapses from 1.56 Å at ambient condition to 1.48 Å at 63 GPa, resulting in a more ionic character of this layered crystal lattice, 3Li++(LiO2)33. This structural change in turn advocates that Li2O2 decomposes to 2Li and O2, further augmented by the densification in specific molar volumes.

Lithium peroxide (Li2O2) is a prototypical alkali metal peroxide that crystallizes into an ionic solid at ambient conditions.1 The strong ionic character in alkali metal peroxides is in contrast to its isoelectronic counterpart H2O2 that crystallizes in a covalent lattice. Furthermore, unlike H2O2, strong hydrogen bonding is absent in alkali metal peroxides. Nevertheless, all peroxides including Li2O2 are strong oxidizers, because of the unusually weak peroxy (O–O) bonds and the comparatively unstable oxidation state (−1) of oxygen atoms;1,2 thus they can explode upon contact with organic materials or other reducing agents,3 similar to H2O2 that even detonates with a detonation velocity of ∼6.7 km/s.4 

Both alkali metal oxides and peroxides are oxygen-rich compounds that are used in applications as oxygen sources, such as in various electrochemical fuel cells.5,6 Of those, Li2O2 has the highest theoretical active oxygen content (34.8 wt. %). It is also a compound that has good thermal stability and is non-hygroscopic.1 Therefore, Li2O2 has received notable attention in rechargeable lithium battery applications, which are based on its decomposition mechanics.7 Li2O2, for example, is one of the main products of oxygen reduction reaction in Li/O2 electrochemical cells, together with Li2O and LiO2.6 Interestingly, Li2O2 crystals in these reaction mixtures are often found in various crystalline and amorphous structures, which deviate from the crystal structure of Li2O2 at ambient conditions.8–10 Neither the crystal structures and stabilities of these discharged solids, nor is the crystal chemistry associated with the decomposition process, are well understood. In this regard, high-pressure studies of Li2O2 can provide insights into the potential crystal structure in solid phases and the decomposition mechanism at various thermodynamic conditions and, thus, help in the development of Li/O2 (or air) electrochemical cells.

Despite numerous studies on alkali oxides and alkali peroxides at ambient conditions,11–14 little is known about high-pressure behaviors of Li2O2 and alkali metal peroxides in general. Therefore, the goal of this study is to investigate the crystal structure and stability of Li2O2 to 63 GPa using diamond anvil cells, confocal micro-Raman spectroscopy, and synchrotron x-ray diffraction. Our results show the pressure-induced reversible chemical decomposition (disassociation) of Li2O2 above 50 GPa, which may have implications to the applications in fuel cells and strong oxidizers.

Powder form of lithium peroxide (Li2O2) from Acros Organic (95% purity) was used without further purification. A small amount of Li2O2 samples is loaded into a small hole in a rhenium gasket (with dimensions of 0.1 mm in diameter and 0.03 mm in thickness) together with few small chips of ruby crystals. We used diamond anvils of 0.3 mm culet flats (∼0.3 carat of type Ia) in a membrane diamond anvil cell (m-DAC) to generate the pressures reached in this study. The internal pressures were determined from the R1 shift of ruby luminescence using quasi-hydrostatic ruby pressure scale.15 Spatially resolved Raman spectra were collected using a custom-built confocal micro-Raman system in backscattering geometry with 514.5 nm excitation line of CW Ar+ ion laser. The laser power was maintained at a minimum level (estimated to be ∼10–20 mW at the sample) to avoid any potential photo-induced transitions and/or decomposition of the sample.

Angle-dispersive x-ray diffraction (ADXD) data of the polycrystalline samples were obtained at the 16 ID-B beamline (λ = 0.4066 Å) at the HPCAT (Advanced Photon Source). The monochromatic x-rays were recorded on a high-resolution 2D MAR imaging plate. The Fit2D program16 was used to convert the diffraction peak positions (2θ) and the intensity information recorded on the image (MAR) plates to 1D profiles. Integrated diffraction patterns were indexed using Dicvol17 and XRDA.18 The proposed space group was selected based on the systematic absence of the observed diffraction peaks. The detailed structural information of Li2O2 was obtained by performing a full-scale Rietveld refinement.19 In this refinement, the scale factor, lattice parameters, zero shift, background, and peak width parameters, u, v, and w were all determined. The background parameters were not further refined after proper subtraction from the raw data.

Figure 1 shows the pressure-induced changes in the Raman spectra of Li2O2 to 63 GPa. The observed Raman spectrum at 1 GPa consists of the most characteristic stretching mode of O22(vsLi2O2) at 798 cm−1 and lattice modes at 143 and 267 cm−1. Note that the lattice modes reported at 78 and 103 cm−1 at ambient conditions20 are not observed because of the spectral cutoff of the Raman notch filter used in the present Raman setup. The lattice mode at 143 cm−1 (vL1Li2O2) is relatively weaker than that at 267 cm−1 (vL2Li2O2). In comparison, the internal vibrational mode (vsLi2O2) is very strong. All these Raman modes shift to higher wavenumber, upon compression to 63 GPa. In addition to these modes, several new modes emerge at around 430 and 1590 cm−1, above 30 GPa, which have the characteristic frequencies of a lattice mode (νL2) and the vibron (νS) of ε-O2.21 Note that the pressure dependence of the low frequency lattice mode observed for Li2O2(vL1Li2O2) is similar to that of the low frequency lattice mode (νL1) observed for ε-O2 above 30 GPa. Even though both the modes corresponding to the ε-O2 lattice modes (νL1, νL2) appear at around 30 GPa, the ε-O2 oxygen vibron (νS) only appears well above 40–45 GPa. The intensity of these peaks continues to increase upon further compression, particularly of the 430 cm−1 peak, which becomes the strongest of all peaks observed, above 50 GPa. Although all the Li2O2 modes still remain strong to the maximum pressure (63 GPa) studied, their intensities notably decrease above 50 GPa. Figure 2 summarizes the pressure dependent spectral shifts of Li2O2 in comparison to those of pure oxygen.21 

FIG. 1.

The pressure-induced Raman spectra of pure Li2O2 at ambient temperature, showing the emergence of peaks corresponding to O2 at around 430 and 1590 cm−1 above 30 GPa. The intensity of these new peaks increases with pressure. Note that all vibrational peaks of Li2O2 still remain strong to the highest pressure studied. The Raman spectrum in red collected at 1.3 GPa upon unloading shows the reversibility of the transition. Microscopic images of Li2O2 samples on the right show the progressive color changes upon compression.

FIG. 1.

The pressure-induced Raman spectra of pure Li2O2 at ambient temperature, showing the emergence of peaks corresponding to O2 at around 430 and 1590 cm−1 above 30 GPa. The intensity of these new peaks increases with pressure. Note that all vibrational peaks of Li2O2 still remain strong to the highest pressure studied. The Raman spectrum in red collected at 1.3 GPa upon unloading shows the reversibility of the transition. Microscopic images of Li2O2 samples on the right show the progressive color changes upon compression.

Close modal
FIG. 2.

The pressure-dependent Raman peak shifts of Li2O2 at ambient temperature, together with those of oxygen (ε-O2)21 for comparison. The blue vertical line signifies the onset of the phase transition above ∼30 GPa. The solid symbols indicate the shifts on uploading (compression), and the open symbols indicate that on unloading (decompression), confirming the reversibility of the transition.

FIG. 2.

The pressure-dependent Raman peak shifts of Li2O2 at ambient temperature, together with those of oxygen (ε-O2)21 for comparison. The blue vertical line signifies the onset of the phase transition above ∼30 GPa. The solid symbols indicate the shifts on uploading (compression), and the open symbols indicate that on unloading (decompression), confirming the reversibility of the transition.

Close modal

Interestingly, the spectral change occurs reversibly upon pressure unloading even when the sample was compressed to 63 GPa. This spectral change accompanies a dramatic color change of the sample from transparent to orange above 30 GPa and to red above 40 GPa. The red color further darkens as the pressure increases, as in ε-O2. However, it is important to note that the observed spectral and color changes suggest that Li2O2 does not immediately lead to chemical decomposition—not until at least 60 GPa. Furthermore, it decomposes only partially, based on (i) the persistence of both the lattice and vibron modes associated with Li2O2 up to 63 GPa, (ii) the absence of Raman evidence for lithia—Li2O16 up to 63 GPa, and (iii) the reversibility of the pressure-induced spectral changes with a little hysteresis. In turn, these observations indicate a subtle change in peroxide O–O packing in Li2O2 crystal similar to that of ε-O2 above 30 GPa, prior to the partial decomposition above 60 GPa. Thus it is interesting to note that the high-pressure behavior of Li2O2 is similar to that in H2O2. H2O2 undergoes a structural phase transition at 13 GPa followed by chemical decomposition to H2 + O2 above 18 GPa.22 

To gain insight into the observed spectral change, we have obtained the ADXD of Li2O2 to 63 GPa as shown in Fig. 3. The diffraction patterns of Li2O2 remain qualitatively unchanged to the maximum pressure, but all diffraction peaks shift toward higher angles as pressure increases. There is no apparent change in the diffraction patterns that can be attributed to the presence of a structural phase transition or a chemical reaction. Small features, especially, at around 10° and 11° at 4 GPa are from the (100) and (101) reflections from the Re gasket. No other diffraction features that could be attributed for Li, O2, or Li2O were observed in crystalline phases.

FIG. 3.

Angle dispersive x-ray diffraction patterns of Li2O2 upon compression at ambient temperature, showing no dramatic change to the maximum pressure (63 GPa) studied. The features marked by asterisks are the (100) and (101) reflections from Re gasket. The weak Re (101) peak is shown on the left of Li2O2 (102) peak (at ∼11.5°) above 56 GPa.

FIG. 3.

Angle dispersive x-ray diffraction patterns of Li2O2 upon compression at ambient temperature, showing no dramatic change to the maximum pressure (63 GPa) studied. The features marked by asterisks are the (100) and (101) reflections from Re gasket. The weak Re (101) peak is shown on the left of Li2O2 (102) peak (at ∼11.5°) above 56 GPa.

Close modal

All observed diffraction patterns can be explained well in terms of two hexagonal structure models proposed previously: (1) P-6 space group—the original model suggested by Föppl.9 and (2) P63/mmc space group, later modified from the original Föppl’s model by Cota et al.10 to include an inversion symmetry and shift atomic positions by +14. Figure 4 shows the measured and calculated diffraction patterns at 4 GPa for P63/mmc and at 63 GPa for both P63/mmc and P-6 for comparison. We used the Rietveld method to refine the structural parameters, peak profiles, zero-shift, oxygen position, and isotropic temperature factors. In the refinement with the P-6 model, the Li position was refined first due to the larger scattering cross section of oxygen. No preferred orientation was included in the refinement. Both the models produce almost identical refinement results with similar χ2-values; therefore, it is not possible to select a better model based on the fitted results. The refined structure gives the unit cell parameters at 4.0 GPa with P63/mmc space group to be, a = 3.108(0) Å and c = 7.548(0) Å and density ρ = 2.413 g/cm3, with reduced χ2 = 1.89. At 63.0 GPa, both the models give the same lattice parameters of a = 2.801(0) Å and c = 6.721(0) Å and density ρ = 3.337 g/cm3, with reduced χ2 = 2.41 for P63/mmc and χ2 = 2.09 for P-6. The detailed structure information from the refinement is listed in Table I.

FIG. 4.

Rietveld refinement of the ADXD patterns of Li2O2 at 4 GPa and 63 GPa based on two previously proposed structural models, P63/mmc and P-6. The measured, refined, and difference spectra are marked in black cross symbols, black line, and red line, respectively. The calculated (hkl) peak positions of Li2O2 are marked in short blue and green bars (for P63/mmc and P-6 models, respectively) and that of Re at 4 GPa in yellow bars. Note that the two structures equally well describe the measured diffraction patterns at all pressures. The detailed structural parameters are summarized in Table I.

FIG. 4.

Rietveld refinement of the ADXD patterns of Li2O2 at 4 GPa and 63 GPa based on two previously proposed structural models, P63/mmc and P-6. The measured, refined, and difference spectra are marked in black cross symbols, black line, and red line, respectively. The calculated (hkl) peak positions of Li2O2 are marked in short blue and green bars (for P63/mmc and P-6 models, respectively) and that of Re at 4 GPa in yellow bars. Note that the two structures equally well describe the measured diffraction patterns at all pressures. The detailed structural parameters are summarized in Table I.

Close modal
TABLE I.

Refined structural parameters of Li2O2 at selected pressures using Rietveld method.

Model-I10 Model-II9 
Pressure4 GPa63 GPa63 GPa
Space group P63/mmc P63/mmc P-6 
a = b (Å) 3.108 (0) 2.801 (0) 2.802 (0) 
c (Å) 7.548 (3) 6.721 (3) 6.721 (3) 
V (Å363.145 (22) 45.667 (18) 82.70 
ρ (g/cm32.413 3.337 3.335 
χ2 1.89 2.41 2.09 
Li1 (0.0,0.0,0.0) (0.0,0.0,0.0) (0.0,0.0,0.0) 
Li2 (1/3,2/3,1/4) (1/3,2/3,1/4) (1/3,2/3,1/2) 
Li3 … … (2/3,1/3,0.243) 
O1 (1/3,2/3,0.853) (1/3,2/3,0.860) (0.0,0,0.610) 
O2 … … (1/3,2/3,0.114) 
dO−O (Å) 1.557 (5) 1.477 (1) 1.478 (2) 
dO−Li1 (Å) 2.101 (2) 1.871 (0) 1.791 (1) 
dO−Li2 (Å) 1.962 (1) 1.778 (2) 1.779 (4) 
dO−Li3 (Å) … … 1.780 (3) 
dLi1−Li2 (Å) 2.604 (1) 2.332 (1) … 
dLi1−Li3 (Å) … … 2.490 (5) 
dLi2−Li3 (Å) … … 2.190 (5) 
Model-I10 Model-II9 
Pressure4 GPa63 GPa63 GPa
Space group P63/mmc P63/mmc P-6 
a = b (Å) 3.108 (0) 2.801 (0) 2.802 (0) 
c (Å) 7.548 (3) 6.721 (3) 6.721 (3) 
V (Å363.145 (22) 45.667 (18) 82.70 
ρ (g/cm32.413 3.337 3.335 
χ2 1.89 2.41 2.09 
Li1 (0.0,0.0,0.0) (0.0,0.0,0.0) (0.0,0.0,0.0) 
Li2 (1/3,2/3,1/4) (1/3,2/3,1/4) (1/3,2/3,1/2) 
Li3 … … (2/3,1/3,0.243) 
O1 (1/3,2/3,0.853) (1/3,2/3,0.860) (0.0,0,0.610) 
O2 … … (1/3,2/3,0.114) 
dO−O (Å) 1.557 (5) 1.477 (1) 1.478 (2) 
dO−Li1 (Å) 2.101 (2) 1.871 (0) 1.791 (1) 
dO−Li2 (Å) 1.962 (1) 1.778 (2) 1.779 (4) 
dO−Li3 (Å) … … 1.780 (3) 
dLi1−Li2 (Å) 2.604 (1) 2.332 (1) … 
dLi1−Li3 (Å) … … 2.490 (5) 
dLi2−Li3 (Å) … … 2.190 (5) 

All observed diffraction data in Fig. 3 were fitted equally well with those in Fig. 4, yielding the pressure dependent unit cell volumes and the 3rd order Birch-Murnaghan equation of state (EOS) fit as presented in Fig. 5. The inset shows the pressure-dependent lattice parameters and the c/a ratio. The structure data at the ambient pressure were from the previous Föppl’s work.9 The compression curve shows a gradual volume change across the phase I to II transition without any apparent volume discontinuity. Therefore, it is fitted with a single EOS model, yielding Bo = 73.16 (2.71) GPa and B′ = 4.36(0.22). The relative compression ratios of ‘a’ and ‘c’ from the ambient pressure to 63 GPa are, respectively, 10.8% and 12.1%, exhibiting a nearly isotropic compression behavior. The isotropic compression behavior of hexagonal crystals is actually unusual, especially for layer structures with a non-ideal c/a ratio as in Li2O2. The c/a ratio drops gradually from 2.43 at ambient to ∼2.39 at 30 GPa, which then remains unchanged to 63 GPa. On the other hand, the change of “c” between 30 GPa and 40 GPa is not as gradual as that of “a,” which results in a wide variation for the c/a ratio. Such variation may be attributed to the change in chemical bonding between Li+ and O22 within the LiO2 layer which can introduce a structural distortion in the graphitic layer.

FIG. 5.

The unit cell volume of pure Li2O2 phases at ambient temperature as a function of pressure. The smooth line is the third-order Birch–Murnaghan EOS fit to the data, using B0 = 73.16 GPa, B′ = 4.36, and V0 = 66.4 Å3. The inset shows the pressure-induced changes in the lattice parameters. The blue vertical dashed line signifies the onset of the iso-structural phase transition at ∼35 GPa.

FIG. 5.

The unit cell volume of pure Li2O2 phases at ambient temperature as a function of pressure. The smooth line is the third-order Birch–Murnaghan EOS fit to the data, using B0 = 73.16 GPa, B′ = 4.36, and V0 = 66.4 Å3. The inset shows the pressure-induced changes in the lattice parameters. The blue vertical dashed line signifies the onset of the iso-structural phase transition at ∼35 GPa.

Close modal

The crystal structure of Li2O2 in both P63/mmc and P-6 models can be understood in terms of a double layer structure of graphitic Li+ and O22 layers alternating with intercalated Li+ layers, as illustrated in Fig. 6. In these models, the graphitic lithium atom is more closely bound with oxygen atoms than the intercalated lithium. For example, at 63 GPa, the P63/mmc model gives two oxygen-lithium distances: dOLi1intra=1.78 Å and dOLi2inter=1.87 Å, whereas the P-6 model gives four oxygen-lithium distances: dO1Li2intra=1.78 Å, dO2Li1intra=1.78 Å, dO1Li3inter=1.90 Å, and dO2Li1inter=1.84 Å. Both the models give the same oxygen-oxygen distance of dO−O = 1.48 Å, reminiscent to an oxygen-oxygen single bond. Thus, the Li2O2 structure can be considered as Li++LiO2, which makes Li2O2 highly ionic. On the other hand, within the graphitic layer, three Li+ ions and three O22 ions form a hexagonal ring of alternating Li+ and O22 ions (or LiO2) bound by more covalent bonding. The intralayer separation of Li+ to the center of O22 is only ∼1.62 Å, underscoring a strong intralayer covalence. The nearest intermolecular O⋅ ⋅ ⋅ O distance is found to be ∼2.48 Å between two adjacent graphitic layers, whereas that within the graphitic layers is ∼2.80 Å. Note that the interlayer oxygen-oxygen distance of 2.48 Å is comparable with that of ε-O2, 2.20-2.34 Å, giving rise to the O2 packing configuration similar to ε-O2 (or (O2)4),23 and probably the observed lattice vibrational modes similar to those in ε-O2 is shown in Fig. 1.

FIG. 6.

Crystal structure of Li2O2 in P63/mmc, showing (a) a double layer structure of Li+ and LiO2 along the a-axis and (b) the graphitic layer structure consisting of three Li+ and three O2 ions along the c-axis.

FIG. 6.

Crystal structure of Li2O2 in P63/mmc, showing (a) a double layer structure of Li+ and LiO2 along the a-axis and (b) the graphitic layer structure consisting of three Li+ and three O2 ions along the c-axis.

Close modal

The pressure-induced structural change may indicate a possibility of the 2nd order isostructural phase transition that leads to chemical decomposition. Figure 7 plots the pressure-dependent changes of the structural order parameter ‘z’ of the oxygen position and the oxygen-oxygen distance. Note that the oxygen-oxygen distance decreases in about 5.6%, from 1.56 Å at ambient to 1.48 Å at 63 GPa, which occurs rather smoothly. On the other hand, there is a slight change in the slope of the pressure-dependent oxygen position shift (Oz) at around 35 GPa. It is also important to note that this occurs at the onset pressure where new modes corresponding to ε-O2 emerge in the Raman spectra in Fig. 1, as well as fluctuations in the lattice parameters are observed in Fig. 5 inset. However, the change in slope of the oxygen z-coordinate is relatively small, and the refinement value is not sufficient to argue for any first-order phase transition.

FIG. 7.

The pressure-induced changes of oxygen order parameter (Oz) for Li2O2 in P63/mmc in O (13, 23, z) plotted together with the nearest oxygen–oxygen distance (dO−O). The inset shows the unit cell of Li2O2 in P63/mmc. The vertical bar signifies a subtle change in the slope of the pressure-dependent Oz change at around 35 GPa, which likely indicates a second order, isostructural phase transition.

FIG. 7.

The pressure-induced changes of oxygen order parameter (Oz) for Li2O2 in P63/mmc in O (13, 23, z) plotted together with the nearest oxygen–oxygen distance (dO−O). The inset shows the unit cell of Li2O2 in P63/mmc. The vertical bar signifies a subtle change in the slope of the pressure-dependent Oz change at around 35 GPa, which likely indicates a second order, isostructural phase transition.

Close modal

We conjecture a topological packing (or pairing) of peroxy oxygen ions along the c-axis to be responsible for the proposed iso-structural phase transition.26 The consequence of such pairing of peroxy ions is the strengthening of peroxy bonds, which will assist in the subsequent chemical decomposition to O2 and Li metals. This conjecture can be further argued in terms of the densification of the lattice upon decomposition as illustrated in Fig. 8. Note that a substantially smaller molar volume of 2Li + O2 than that of Li2O+12O2 clearly favors the conjectured decomposition mechanism. It is believed that the kinetics associated with the chemical decomposition retards its occurrence where the lattice develops substantial level of chemical bonding between the peroxy ions at the pressure range of 30-50 GPa.

FIG. 8.

The pressure-induced changes of molar volumes of Li2O2 (blue line and solid squares), obtained from the present ADXD study, plotted in comparison with those of 2Li + O2, and Li2O+12O2 mixtures, Li2O, and O2. The data for O2, Li2O, and Li are produced from the previous studies by Fujihisa et al.,24 Lazicki et al.,11 and Guillaume et al.,25 respectively.

FIG. 8.

The pressure-induced changes of molar volumes of Li2O2 (blue line and solid squares), obtained from the present ADXD study, plotted in comparison with those of 2Li + O2, and Li2O+12O2 mixtures, Li2O, and O2. The data for O2, Li2O, and Li are produced from the previous studies by Fujihisa et al.,24 Lazicki et al.,11 and Guillaume et al.,25 respectively.

Close modal

This material is based upon work supported by the U.S. Department of Homeland Security, Science and Technology Directorate, Office of University Programs, under Grant Award No. 2013-ST-061-ED0001, Defense Threat Reduction Agency, under Grant Award No. HDTRA1-12-01-0020), and National Science Foundation, Division of Materials Research under Grant No. 1203834. The views and conclusions contained in this document are those of the authors and should not be interpreted as necessarily representing the official policies, either expressed or implied, of the U.S. Department of Homeland Security, Defense Threat Reduction Agency and/or National Science Foundation. The CDAC has supported the beamtime for the powder x-ray diffraction experiments at the HPCAT. The HPCAT is operated in support of DOE-NNSA (Grant No. DE-NA0001974) and DOE-BES (Grant No. DE-FG02-99ER45775).

1.
H.
Jakob
,
S.
Leininger
,
T.
Lehmann
,
S.
Jacobi
, and
S.
Gutewort
, “
Peroxo compounds, inorganic
,”
Ullmann’s Encycl. Ind. Chem.
26
,
293
(
2011
).
2.
D. E.
Clark
,
Peroxides and Peroxide-Forming Compounds
(
Texas A&M University
,
2000
).
3.
J. C.
Oxley
,
Terrorism Political Violence
5
,
30
(
1993
).
4.
R.
Engelke
,
S. A.
Sheffield
, and
L. L.
Davis
,
J. Phys. Chem. A
104
,
6894
(
2000
).
5.
M. V.
Aleinikova
,
Yu. N.
Zhuravlev
, and
D. V.
Korabelnikov
,
Russ. Phys. J.
55
,
495
(
2012
).
6.
M. K. Y.
Chan
,
E. L.
Shirley
,
N. K.
Karan
,
M.
Balasubramanian
,
Y.
Ren
,
J. P.
Greeley
, and
T. T.
Fister
,
J. Phys. Chem. Lett.
2
,
2483
(
2011
).
7.
T.
Ogasawara
,
A.
Débart
,
M.
Holzapfel
,
P.
Novák
, and
P. G.
Bruce
,
J. Am. Chem. Soc.
128
,
1390
(
2006
).
8.
K. C.
Lau
,
D.
Qiu
,
X.
Luo
,
J.
Greeley
,
L. A.
Curtiss
,
J.
Lu
, and
K.
Amine
,
Energies
8
,
529
(
2015
).
9.
H.
Föppl
,
Z. Anorg. Allg. Chem.
291
,
12
(
1957
).
10.
L. G.
Cota
and
P.
de la Mora
,
Acta Crystallogr., Sect. B: Struct. Sci.
B61
,
133
(
2005
).
11.
A.
Lazicki
,
C. S.
Yoo
,
W. J.
Evans
, and
W. E.
Pickett
,
Phys. Rev. B
73
,
184120
(
2006
).
12.
K.
Kunc
,
I.
Loa
, and
K.
Syassen
,
Phys. Rev. B
77
,
094110
(
2008
).
13.
M.
Moakafi
,
R.
Khenata
,
A.
Bouhemadou
,
H.
Khachai
,
B.
Amrani
,
D.
Rached
, and
M.
Rérat
,
Eur. Phys. J. B
64
,
35
(
2008
).
14.
Yu. N.
Zhuravlev
,
D. V.
Korabelnikov
, and
M. V.
Aleinikova
,
Phys. Solid State
54
,
1518
(
2012
).
15.
H. K.
Mao
,
J.
Xu
, and
P. M.
Bell
,
J. Geophys. Res.
91
,
4673
, doi:10.1029/JB091iB05p04673 (
1986
).
16.
A. P.
Hemmersley
,
FIT2D V9.129 reference manual V3.1. ESRF Report 98HA01T
,
1998
.
17.
A.
Boultif
and
D.
Louër
,
J. Appl. Crystallogr.
37
,
724
(
2004
).
18.
S.
Desgreniers
and
K.
Lagarec
,
J. Appl. Crystallogr.
27
,
432
(
1994
).
19.
A. C.
Larson
and
R. B.
Von Dreele
, “
General structure analysis system (GSAS)
,”
Los Alamos National Laboratory Report LAUR 86-748
,
2000
, pp.
1
221
.
20.
H. H.
Eysel
and
S.
Thym
,
Z. Anorg. Allg. Chem.
411
,
97
(
1975
).
21.
F. A.
Gorelli
,
L.
Ulivi
,
M.
Santoro
, and
R.
Bini
,
Phys. Rev. B
63
,
104110
(
2001
).
22.
J. Y.
Chen
,
M.
Kim
,
C. S.
Yoo
,
D.
Dattelbaum
, and
S.
Sheffield
,
J. Chem. Phys.
132
,
214501
(
2010
).
23.
L. F.
Lundegaard
,
G.
Weck
,
M. I.
McMahon
,
S.
Desgreniers
, and
P.
Loubeyre
,
Nature
443
,
201
(
2006
).
24.
H.
Fujihisa
,
Y.
Akahama
,
H.
Kawamura
,
Y.
Ohishi
,
O.
Shimomura
,
H.
Yamawaki
,
M.
Sakashita
,
Y.
Gotoh
,
S.
Takeya
, and
K.
Honda
,
Phys. Rev. Lett.
97
,
085503
(
2006
).
25.
C. L.
Guillaume
,
E.
Gregoryanz
,
O.
Degtyareva
,
M. I.
McMahon
,
M.
Hanfland
,
S.
Evans
,
M.
Guthrie
,
S. V.
Sinogeikin
, and
H.-K.
Mao
,
Nat. Phys.
7
,
211
(
2011
).
26.
M.
Dunuwille
, Ph.D dissertation,
Washington State University
,
2015
.