Received: Added at production Revised: Added at production Accepted: Added at production

DOI: xxx/xxxx

ORIGINAL ARTICLE

Experimental Pathways for Detecting Double Superionicity in

Planetary Ices

Kyla de Villa1 Felipe González‐Cataldo1 Burkhard Militzer1,2

1
Department of Earth and Planetary Science, Abstract
University of California, Berkeley, California,
USA The ice giant planets Uranus and Neptune are assumed to contain large amounts of planetary ices
2
Department of Astronomy, University of such as water, methane, and ammonia. The properties of mixtures of such ices at the extreme
California, Berkeley, California, USA pressures and temperatures of planetary interiors are not yet well understood. Ab initio computer
simulations predicted that a number of ices exhibit a hydrogen superionic state and a doubly superi‐
Correspondence
onic state [DOI 10.1038/s41467‐023‐42958‐0]. Since the latter state has not yet been generated
Corresponding author Kyla de Villa.
with experiments, we outline here two possible pathways for reaching and detecting such a state
Email: [email protected]
with dynamic compression experiments. We suggest X‐ray diffraction as the principal tool for de‐
tecting when the material becomes doubly superionic and the sublattice of one of the heavy nuclei
melts. That would require a temperature of ∼3500 K and pressures greater than ∼200 GPa for
H3 NO4 , which we use as an example material here. Such conditions can be reached with experi‐
ments that employ an initial shock that is followed by a ramp compression wave. Alternatively, one
may use triple‐shock compression because a single shock does not yield sufficiently high densities.

KEYWORDS
superionic, high pressure, planetary ices, Uranus, Neptune, ice giant planets, X‐ray diffraction, shock
Hugoniot curve, density functional theory, molecular dynamics

Recently, crystal structure prediction methods have been applied to

1 INTRODUCTION
infer stable crystalline phases of H‐C‐N‐O materials which may form

The ice giants Uranus and Neptune orbit our sun at distances of 19 from H2 O, NH3 , and CH4 at the extreme pressures of planetary in‐

and 30 astronomical units, respectively, making them difficult to visit teriors 7,8 . An investigation of the high‐temperature behaviors of the

by spacecraft. Consequently, they are relatively unexplored, with open hydrogen bearing materials predicted through these structure searches

questions remaining regarding their interior structures, thermal profiles, revealed the prevalence of the hydrogen superionic phase of planetary

dynamic processes, and even composition. Following the rise of exo‐ ices at planetary interior conditions 9,10,11,12 . This phase is characterized

planet detection methods in the last 30 years, such questions now also by the rapid diffusion of H ions through a stable lattice of the heav‐

extend to super‐Earth and sub‐Neptune sized exoplanets, which domi‐ ier ions in the material, and has been observed computationally and

nate currently discovered exoplanets 1 . Although the exact proportion indirectly through experiments for the H‐C‐N‐O chemical space for wa‐

of rock to ice in these planets is uncertain, it is typically assumed that ter, ammonia, ammonia hydrides, and synthetic Uranus. Computational

Uranus and Neptune are composed of at least 50% planetary ices ‐ work has revealed a second superionic phase in many H‐C‐N‐O materi‐

mixtures of water, methane, and ammonia accreted in large volumes als: double superionicity, in which a second element (the lightest of the

by these planets during their formation 2 ; but the compositional space heavy species C, N, and O) also diffuses superionically, simultaneously

possible for icy exoplanets is far broader. with the superionic H ions 12 . Doubly superionic behavior of H‐C‐O ma‐

At the present time, the properties of mixtures of planetary ices terials was also recently reported, at pressures of 100‐500 GPa with

are relatively unexplored at extreme conditions, with most studies fo‐ simultaneously diffusing H and C atoms 13 .

cused on pure materials, binary mixtures, or specific stoichiometric Ion motion in the doubly superionic phase is demonstrated in Fig. 1
3
mixtures such as “synthetic Uranus,” with a H:C:N:O ratio of 27:7:4:1, for H3 NO4 in the P21 21 21 phase, which shows the oxygen ions limited

an approximate representative of proto‐solar elemental ratios 4,5,6 . to vibrating in their potential wells, while the N and H ions diffuse widely

Contrib. Plasma Phys. 2024;00:1–8 wileyonlinelibrary.com/journal/ © 2024 Copyright Holder Name 1

2 DE VILLA et al.

throughout the cell. Figure 2 also illustrates the doubly superionic mo‐ 288 atom cells of H3 NO4 ‐P21 21 21 were geometrically optimized to
tion of H and N in H3 NO4 ‐P21 21 21 , by showing the distinct diffusion desired pressures of 150‐550 GPa, and were then simulated in the NVT
pathways that each ion type adopts as ions travel through the cell. H ensemble equilibrated with the Nose‐Hoover thermostat 21,22 , using a
and N ion pathways rarely overlap, as H ions primarily orbit O ions at 0.5 fs timestep. H3 NO4 ‐P21 21 21 was heated at fixed density from 0 K
short distances, whereas N ions maximize their distance from O ions. up to temperatures beyond the melting points in 500 K increments us‐
9
In their seminal paper, Cavazzoni et al. reported results from den‐ ing the heat‐until‐it‐melts approach 23 , and thus the melting line shown
sity functional molecular dynamics simulations that predicted water here represents just an upper bound to the melting temperature.
and ammonia become superionic at conditions of high pressure and 144 atom cells of a 1:1 liquid H2 O and HNO3 mixture were prepared
temperature in the interiors of Uranus and Neptune. Superionic wa‐ using the Packmol software 24 , followed by geometry optimization at
ter may contribute to the generation of their magnetic fields because pressures from 1 atm to 100 GPa. These cells were similarly heated
the hydrogen nuclei carry an electrical charge so that their diffusion up to 2000 K at fixed density with NVT simulations performed until
contributes to the material’s electrical conductivity. Earlier works had energy equilibration was reached. Upon sufficient pressurization the
relied on more approximate theoretical methods to predict that water H2 O‐HNO3 mixtures exhibited amorphous solid and amorphous supe‐
14
ice exhibits a superionic phase, by applying defect theory in 1985 , rionic phases. Further work is required to constrain the low pressure
and using classical interatomic potentials in 1988 15 . Indirect experi‐ boundaries between the solid, superionic, doubly superionic, and liquid
mental confirmation of the existence of superionic phases of planetary phases of this mixture.
ices has only been achieved in recent years. Raman spectroscopy has Electronic structure calculations were completed using the Perdew,
been employed for both water and ammonia to show the disorder‐ Burke, and Ernzerhof (PBE) exchange‐correlation functional 25 and a
ing of O‐H and N‐H bonds, interpreted to signify the onset of proton pseudopotential of the projector‐augmented wave (PAW) 26 type using
16,17
diffusion . Ionic conductivity, combined with compressibility mea‐ a plane‐wave cutoff of 1100 eV with core radii of 0.8, 1.5, and 1.1 Å for
surements derived from X‐ray diffraction (XRD) have been used to hydrogen, nitrogen, and oxygen, respectively. The Brillouin zone was
18
argue the observation of the onset of the superionic phase . Simi‐ sampled with the Γ‐point only which is sufficient for the supercells
larly, the superionic phase of water has been experimentally verified by used in our DFT‐MD simulations. Electronic states were populated ac‐
showing a difference between electronic conductivity derived from ab‐ cording to a Fermi‐Dirac distribution using the Mermin functional 27 to
sorption and reflectivity measurements, and total conductivity derived incorporate the excited electronic states at finite temperature within
from impedance measurements and density functional theory calcula‐ the Kohn‐Sham formalism 28 . These parameters were tested for conver‐
tions, which can only be explained by the high ionic conductivity of gence within 1% for energy, pressure, and components of the stress
the superionic phase 11 . Most recently, in situ XRD measurements of tensor.
shock compressed water have shown the formation of body‐centered
cubic (bcc) and face‐centered cubic (fcc) oxygen sublattices at the tem‐
perature and pressure conditions associated with superionic water, 2.2 Shock Hugoniot Calculations
19
experimentally confirming the high‐pressure stability of this phase .
In this work, we propose a pathway for the experimental detection of Shock wave experiments are the experimental method of choice when

the doubly superionic phase of the planetary ice H3 NO4 with symme‐ materials at megabar pressures and elevated temperatures need to

try group P21 21 21 by presenting computed Hugoniot curves and X‐ray be studied 29 . The resulting shock Hugoniot curve can be calculated

diffraction patterns based on first principles simulations of shock con‐ theoretically as long as the equation of state of relevant materials is

ditions. We hope that by applying similar experimental approaches as known with sufficient precision 30,31,32 . At higher temperatures beyond

those applied to characterize the properties of superionic water at ex‐ ∼2×105 K, where nonideal mixing effects have shown to be small 33 ,

treme temperatures and pressures, this novel phase can be observed one may derive the Hugoniot curve from a linear combination of the

experimentally for the first time. EOSs of oxygen 34 , nitrogen 35 , and hydrogen under the ideal mixing
approximation 36,37,38,39,40 . Here we employ DFT‐MD simulations to
compute the equation of state (EOS) of H3 NO4 directly.
The principal Hugoniot curve was determined by solving for energy,
2 METHODS
pressure, volume, and temperature conditions that satisfy the Rankine‐
Hugoniot relationship for shock compression 41 :
2.1 Molecular Dynamics Simulations
1
(E – E0 ) + (P + P0 ) × (V – V0 ) = 0.
We considered all H‐bearing structures predicted to be stable by the 2
H‐C‐N‐O quaternary structure searches executed by Naumova et al. 7 Here E0 , P0 , and V0 represent the initial state of the sample being
and Conway et al. 8 . Simulations were performed using density func‐ compressed, which we assume to be a homogeneous liquid H2 O‐HNO3
tional molecular dynamics (DFT‐MD) as implemented in the Vienna ab mixture at room temperature and pressure. P0 was taken to be 1 atm.
initio simulation package (VASP) 20 . As H2 O and HNO3 are liquids at room temperature and pressure, V0
Experimental Pathways for Detecting Double Superionicity in Planetary Ices 3

a) First 0.1 ps b) ~15 ps, all N ions shown c) ~15 ps, only 2 N ions shown

F I G U R E 1 Trajectory visualizations for doubly superionic H3 NO4 ‐P21 21 21 at 4500 K with a density of 6.033 g/cm3 . Individual spheres for
each species (O in red, N in blue, and H in white) represent the positions of ions at different time steps, illustrating their motion throughout the
simulation. (a) First 0.1 ps of the simulation, demonstrating initial lattice sites of nitrogen and oxygen ions together with H diffusion. Even on such
a short timescale, H ions diffuse rapidly. (b) ∼15 ps trajectory showing rapid and widespread N diffusion, whereas oxygen ions vibrate around their
potential wells, remaining confined to their lattice sites without diffusing. (c) Same as (b), now with the trajectories of only 2 N ions shown, with
one in blue and one in green. For (b) and (c), H trajectories are hidden for ease of visualization of the N trajectories.

was calculated by assuming a linear mixing of their volumes. E0 was ob‐ and ρ∼2.396 g/cm3 , and a third shock starting from T∼1500 K and
tained by constructing a cell of 18 H2 O + 18 HNO3 molecules for a ρ∼3.486 g/cm3 . The initial conditions for a 1:1 liquid mixture of H2 O
24
total of 144 atoms using the Packmol software , followed by geometry and HNO3 and the requisite conditions for the double and triple‐shocks
optimization. are detailed in Table 1. These conditions can be easily achieved by laser
platforms such as NIF 43 and the OMEGA Laser Facility 44 . While the
initial state of the H3 NO4 is liquid, both ramp compression and the
3 RESULTS AND DISCUSSION secondary or tertiary shock will pressurize the material into a solid state.

3.1 Shock Hugoniot Curves

3.2 X‐Ray Diffraction Spectra
A number of planetary ices have now been suggested to exhibit a doubly
superionic phase at the extreme pressures and temperatures of plane‐ X‐ray diffraction (XRD) is the principal method for detecting phase

tary interiors 12,13

. One such compound, H3 NO4 has been predicted 7,8
, changes at high pressure 45 . Millot et al. 11 employed this technique

to have several stable structures, with a C21 structure exhibiting dou‐ to determine that superionic water ice assumes a face‐centered cubic

ble superionicity at a density of 5.644 g/cm3 , and a P21 21 21 structure structure at high pressure as was predicted with DFT‐MD simulations by

exhibiting double superionicity across a density and pressure range of Wilson et al. 10 . However, Millot et al. relied on indirect arguments about

at least 4.331‐6.203 g/cm3 and 150‐600 GPa (see Figure 3). The wide the electrical conductivity to argue that a hydrogen‐superionic state had

pressure range over which H3 NO4 ‐P21 21 21 exhibits double superionic‐ been reached because hydrogen atoms scatter X‐rays weakly compared

ity makes this material a promising candidate for experimental studies. to the heavier oxygen nuclei that come with many more localized core

Moreover, this stoichiometry can be achieved using 1:1 molar ratios of electrons to scatter off of.

liquid H2 O and HNO3 (nitric acid), which are miscible liquids at room The prospects of directly detecting the transition to a doubly superi‐

temperature and pressure . 42 onic state with X‐ray diffraction measurements are much better because

With this motivation, we propose two possible shock compression one of the heavier atomic species becomes disordered and one can

pathways for reaching and detecting the doubly superionic phase of expect the XRD peaks associated with that species (nitrogen in the fol‐

this material, shown in Figures 3 and 4. As these figures show, the prin‐ lowing example) to disappear. We computed the XRD spectra shown in

cipal Hugoniot curve for H3 NO4 does not reach sufficient density to Fig. 5 by averaging over MD trajectories of H3 NO4 performed with 288
46
sample the doubly superionic phase. However, such conditions could atom supercell configurations at a density of 6.033 g/cm3 , which cor‐

be reached by employing a single shock followed by ramp compression. responds to a pressure of 500 GPa at 0 K. One finds that multiple peaks

Figure 3 shows that ramp compression starting at different conditions appear very closely together because the orthorhombic unit cell has the

along the principal Hugoniot curve would allow each phase of H3 NO4 dimensions a=2.805 Å, b=5.582 Å, and c=5.700 Å. Because 2a ≈ b ≈ c,

to be reached at high pressure. the peaks with the Miller indices (2 0 0), (0 4 0), and (0 0 4) appear

Alternatively, double or triple‐shock experiments may be performed very close together as they represent Bragg planes with very similar d

to reach the doubly superionic phase of H3 NO4 (see Fig. 4). From the spacings.

principal Hugoniot, we propose a second shock starting from T∼750 K

4 DE VILLA et al.

FCC O Lattice Only Superionic H Isosurface Superionic N Isosurface Superionic H+N Isosurfaces

F I G U R E 2 Isosurfaces showing the regions in which superionic ions are most likely to be found in H3 NO4 ‐P21 21 21 at 4000 K with a density of
6.033 g/cm3 . Superionic H is shown in gray, superionic N is shown in blue, with O ions given as red spheres. N and H ions travel through different
pathways through the fcc oxygen sublattice, with their pathways rarely crossing. H ions primarily orbit oxygen ions at short distances, whereas N
ions maximize their distance from O ions. Top Row: 48 oxygen ion sublattice; Middle Row: fcc oxygen unit cell; Bottom Row: one layer of an fcc
oxygen unit cell.

Stage T (K) P (GPa) ρ (g/cm3 ) E (eV/atom) Us (km/s) Up (km/s)

Initial State 300 0.0001 1.358 –5.496 – –
Principal Shock 750 18.170 2.396 –5.179 5.448 2.360
Secondary Shock 1500 86.355 3.468 –4.474 9.593 2.963
Tertiary Shock 4000 383.992 5.325 –1.989 15.687 5.472
T A B L E 1 Sample conditions at different stages of a triple‐shock experiment, corresponding to the suggested compression pathway shown in
Fig. 4. The conditions listed for the principal shock are the requisite starting conditions for the secondary shock, and those listed for the secondary
shock are the requisite starting conditions for the tertiary shock. The state listed for the tertiary shock is within the predicted boundaries of the
doubly superionic phase. These conditions are represented in Figure 4 as cyan hexagons.

We find the XRD spectra for 500 K (solid phase) and 3000 K (hy‐ 3000 K. Because hydrogen atoms scatter X‐rays only weakly, we at‐
drogen superionic phase) in Fig. 5 to be very similar. The peak heights tribute the reduction in peak height to the thermal motion of the heavy
are reduced by a modest amount for the hydrogen superionic state at nuclei due to the Debye‐Waller effect 47 . We predict that it will be dif‐
ficult to detect this reduction directly with XRD measurements as we
Experimental Pathways for Detecting Double Superionicity in Planetary Ices 5

F I G U R E 3 Phase diagram for H3 NO4 showing proposed compression pathways using a single shock followed by ramp compression. The
principal Hugoniot curve is shown in green. Isentropes starting at different conditions along the principal Hugoniot curve allow for different phases
of H3 NO4 to be reached. Diamonds represent simulations performed starting at 0 K with a P21 21 21 structure of H3 NO4 . Purple circles represent
simulations that were started from a 1:1 liquid mixture of H2 O and HNO3 .

have seen for H2 O 11 . If one increases the temperatures from 3000 K to the pressure‐temperature field where the H3 NO4 ‐P21 21 21 structure is
4500 K, the first three peaks in the XRD spectrum in Fig. 5 all disappear strongly favored over other structures.
because the nitrogen atoms become disordered as the material transi‐
tions to a doubly superionic state. In our view, this is the most promising
way to detect a doubly superionic state with laboratory experiments. 4 CONCLUSIONS
Many other peaks that are associated with an ordered arrangement
of oxygen atoms remain present at 4500 K while their peak widths We presented results from ab initio computer simulations of the material
increase because the of thermal motion of the nuclei. H3 NO4 in the P21 21 21 structure over a broad range of pressures and
If the temperature is increased further to 6000 K, the material melts temperature from 150‐700 GPa and 500‐7000 K. We discussed path‐
and the computed XRD spectrum exhibits the typical broad peak of a ways for identifying the transition from a solid to hydrogen superionic to
liquid. The peak maximum is centered at the (1 2 2) peak, which is the a doubly superionic and finally to a liquid state with experimental meth‐
strongest reflection in all three lower‐temperature spectra. This makes ods. We showed that XRD measurements are a promising approach to
this peak essential for detecting that a transition to a solid or superi‐ directly detect the transition from a hydrogen superionic to a doubly su‐
onic phase has occurred during a shock experiment, which we assume perionic state. Our calculations show that during this transition, three
started from a mixture of two liquids at ambient conditions. prominent peaks in the XRD spectrum disappear because the nitrogen
The H3 NO4 ‐P21 21 21 structure is predicted to be thermodynami‐ atoms become disordered.
cally stable at pressures greater than 200 GPa at zero temperature We conclude that dynamic compression experiments that employ
while its stability range at high temperature has not yet been mapped an initial shock that is followed a ramp wave are the preferred ex‐
out carefully. Given this uncertainty, one may consider conducting perimental technique to generate the conditions of T∼3500 K and
double or triple‐shock experiments 48,49,50 , ramp compression measure‐ P>200 GPa where H3 NO4 is predicted to assume a doubly superi‐
ments 51,52,53 or employ diamond cells as a precompression module to onic state. Alternatively triple‐shock experiments may be employed
54,55,56,57,58 while a single shock will not yield sufficiently high densities. At the
increase the initial density of the sample to probe deeper in
6 DE VILLA et al.

F I G U R E 4 Equations of state of principal, secondary, and tertiary shocks of H3 NO4 starting from a 1:1 mixture of H2 O and HNO3 . Top left:
pressure/temperature phase diagram showing different shock pathways through doubly superionic phase. Top right: density/temperature phase
diagram. Bottom left: density/pressure phase diagram. Bottom right: particle velocity vs. shock velocity for single, double, and triple‐shocks. For all
plots, green lines represent the principal shock, red is secondary shock, and blue is tertiary shock. Cyan hexagons show the locations of the points
indicated in Table 1.

present time, it remains unclear whether double‐shock experiments REFERENCES

would be able to reach a doubly superionic state because its field of sta‐ 1. Fulton BJ, Petigura EA, Howard AW, et al. The California‐Kepler
survey. III. A gap in the radius distribution of small planets. The
bility in pressure‐temperature space has not yet been determined with
Astronomical Journal. 2017;154(3):109.
sufficient accuracy. 2. Malamud U, Podolak M, Podolak JI, Bodenheimer PH. Uranus
and Neptune as methane planets: Producing icy giants from
AUTHOR CONTRIBUTIONS refractory planetesimals. Icarus. 2024;421:116217.
BM conceived of the project. KDV performed the simulations. KDV, 3. Radousky H, Mitchell A, Nellis W. Shock temperature measure‐
ments of planetary ices: NH3, CH4, and ‘‘synthetic Uranus’’. The
FGC, and BM analyzed the data and wrote the manuscript.
Journal of chemical physics. 1990;93(11):8235–8239.
4. Nellis W, Hamilton D, Holmes N, et al. The nature of the interior
CONFLICT OF INTEREST of Uranus based on studies of planetary ices at high dynamic
The authors declare no potential conflict of interests. pressure. Science. 1988;240(4853):779–781.
5. Lodders K. Solar system abundances of the elements. In:
ACKNOWLEDGEMENTS Goswami A, Reddy BE., eds. Astrophysics and Space Science Pro‐
ceedings, , Berlin: Springer‐Verlag, 2010ch. Principles:379–417.
This material is based on work supported by the Department of Energy, 6. Chau R, Hamel S, Nellis WJ. Chemical processes in the deep
National Nuclear Security Administration under award DE‐NA0004147. interior of Uranus. Nature communications. 2011;2(1):203.
Computational resources at the National Energy Research Scientific 7. Naumova AS, Lepeshkin SV, Bushlanov PV, Oganov AR. Un‐
usual chemistry of the C–H–N–O system under pressure and
Computing Center and the Livermore Computing Center were used.
implications for giant planets. The Journal of Physical Chemistry
A. 2021;125(18):3936–3942.
DATA AVAILABILITY STATEMENT 8. Conway LJ, Pickard CJ, Hermann A. Rules of formation of
The data that support the findings of this study are available from the H–C–N–O compounds at high pressure and the fates of plan‐
corresponding author upon reasonable request. etary ices. Proceedings of the National Academy of Sciences.
Experimental Pathways for Detecting Double Superionicity in Planetary Ices 7

25
Solid

(1 2 2)
T = 500 K
Superionic

(0 0 4), (2 0 0), (0 4 0)
20 T = 3000 K
Doubly superionic
X-ray intensity (arbitrary units)

T = 4500 K
Liquid
(0 1 2), (1 1 0) T = 6000 K
15

(2 0 4), (0 4 4), (2 4 0)

(1 2 6), (3 2 2), (1 6 2)
10
(0 2 0)

(1 0 2)

010 20 30 40 50 60 70 80 90
Scattering angle 2 (degrees)
F I G U R E 5 Computed X‐ray diffraction spectra of H3 NO4 ‐P21 21 21 for a density of 6.033 g/cm3 . Assuming a wavelength of 1 nm, the spectra
were calculated by averaging configurations of MD trajectories for the solid phase at 500 K (dashed blue), the hydrogen superionic phase at 3000 K
(black), the doubly superionic (H+N) phase at 4500 K (red), and the liquid phase at 6000 K (orange). The Miller indices are shown for all major
peaks. From the solid to the superionic phase, peak intensities decrease only slightly, which is primarily due to thermal motion of the heavy nuclei
because hydrogen is a weak X‐ray scatterer. During the transition from the hydrogen superionic to doubly superionic phase, the first three peaks
disappear because the nitrogen nuclei become disordered. The strongest (1 2 2) reflection persists because the oxygen nuclei remain bound to
their lattice sites. When the entire structure melts at 6000 K, this peak marks the maximum of the broad peak of the liquid.

2021;118(19). 15. Demontis P, LeSar R, Klein ML. New high‐pressure phases of ice.
9. Cavazzoni C, Chiarotti G, Scandolo S, Tosatti E, Bernasconi M, Physical Review Letters. 1988;60(22):2284.
Parrinello M. Superionic and metallic states of water and ammo‐ 16. Goncharov AF, Goldman N, Fried LE, et al. Dynamic ioniza‐
nia at giant planet conditions. Science. 1999;283(5398):44–46. tion of water under extreme conditions. Physical review letters.
10. Wilson HF, Wong ML, Militzer B. Superionic to Supe‐ 2005;94(12):125508.
rionic Phase Change in Water: Consequences for the 17. Ninet S, Datchi F, Saitta AM. Proton disorder and superi‐
Interiors of Uranus and Neptune. Physical Review Letters. onicity in hot dense ammonia ice. Physical Review Letters.
2013;110(15):151102. doi: 10.1103/PhysRevLett.110.151102 2012;108(16):165702.
11. Millot M, Hamel S, Rygg JR, et al. Experimental evidence for 18. Sugimura E, Komabayashi T, Ohta K, Hirose K, Ohishi
superionic water ice using shock compression. Nature Physics. Y, Dubrovinsky LS. Experimental evidence of superionic
2018;14(3):297–302. conduction in H2 O ice. The Journal of Chemical Physics.
12. de Villa K, González‐Cataldo F, Militzer B. Double superionic‐ 2012;137(19):194505.
ity in icy compounds at planetary interior conditions. Nature 19. Millot M, Coppari F, Rygg JR, et al. Nanosecond X‐ray
Communications. 2023;14(1):7580. diffraction of shock‐compressed superionic water ice. Nature.
13. Murayama D, Ohmura S, Kodama R, Ozaki N. Manuscript in 2019;569(7755):251–255.
prep. Personal communication from 2024 Gordon Conference 20. Kresse G, Furthmüller J. Efficient iterative schemes for ab initio
on High Pressure, noted with authors’ permission. total‐energy calculations using a plane‐wave basis set. Phys. Rev.
14. Ryzhkin I. Superionic transition in ice. Solid State Communica‐ B. 1996;54:11169–11186. doi: 10.1103/PhysRevB.54.11169
tions. 1985;56(1):57–60. 21. Nosé S. A unified formulation of the constant temperature
molecular dynamics methods. The Journal of chemical physics.
8 DE VILLA et al.

1984;81(1):511–519. 45. Gorman MG, Elatresh S, Lazicki A, et al. Experimental ob‐

22. Shuichi N. Constant temperature molecular dynamics methods. servation of open structures in elemental magnesium at ter‐
Progress of Theoretical Physics Supplement. 1991;103:1–46. apascal pressures. Nature Physics. 2022;18:1307‐1311. doi:
23. Wang S, Zhang G, Liu H, Song H. Modified Z method to calcu‐ 10.1038/s41567‐022‐01732‐7
late melting curve by molecular dynamics. Journal of Chemical 46. Militzer B. Supercell design for first‐principles simulations of
Physics. 2013;138(13). doi: 10.1063/1.4798225 solids and application to diamond, silica, and superionic water.
24. Martínez L, Andrade R, Birgin EG, Martínez JM. PACK‐ High Energy Density Physics. 2016;21:8–15.
MOL: A package for building initial configurations for molec‐ 47. Simon SH. The Oxford solid state basics. OUP Oxford, 2013.
ular dynamics simulations. Journal of computational chemistry. 48. Boehly T, Hicks D, Celliers P, et al. Properties of fluid deuterium
2009;30(13):2157–2164. under double‐shock compression to several Mbar. Physics of
25. Perdew JP, Burke K, Ernzerhof M. Generalized Gradient Approx‐ plasmas. 2004;11(9):L49–L52.
imation Made Simple. Phys. Rev. Lett.. 1996;77:3865–3868. doi: 49. Hansen L, Fratanduono D, Zhang S, et al. Melting of magnesium
10.1103/PhysRevLett.77.3865 oxide up to two terapascals using double‐shock compression.
26. Blöchl PE. Projector augmented‐wave method. Phys. Rev. B. Physical Review B. 2021;104(1):014106.
1994;50:17953–17979. doi: 10.1103/PhysRevB.50.17953 50. Militzer B, Ceperley D, Kress J, Johnson J, Collins L, Mazevet
27. Mermin ND. Thermal Properties of the Inhomogeneous S. Calculation of a deuterium double shock Hugoniot from ab
Electron Gas. Phys. Rev.. 1965;137:A1441–A1443. doi: initio simulations. Physical review letters. 2001;87(27):275502.
10.1103/PhysRev.137.A1441 51. Smith R, Eggert J, Jeanloz R, et al. Ramp compression of dia‐
28. Kohn W, Sham LJ. Self‐Consistent Equations Including Ex‐ mond to five terapascals. Nature. 2014;511(7509):330–333.
change and Correlation Effects. Phys. Rev.. 1965;140:A1133– 52. Lazicki A, McGonegle D, Rygg J, et al. Metastability
A1138. doi: 10.1103/PhysRev.140.A1133 of diamond ramp‐compressed to 2 terapascals. Nature.
29. Millot M, Zhang S, Fratanduono DE, et al. Recreating gi‐ 2021;589(7843):532–535.
ants impacts in the laboratory: Shock compression of 53. González‐Cataldo F, Godwal B, Driver K, Jeanloz R, Militzer B.
bridgmanite to 14 Mbar. Geophysical Research Letters. Model of ramp compression of diamond from ab initio simula‐
2020;47(4):e2019GL085476. tions. Physical Review B. 2021;104(13):134104.
30. Soubiran F, González‐Cataldo F, Driver KP, Zhang S, Militzer 54. Eggert J, Brygoo S, Loubeyre P, et al. Hugoniot data
B. Magnesium oxide at extreme temperatures and pressures for helium in the ionization regime. Physical review letters.
studied with first‐principles simulations. The Journal of Chemical 2008;100(12):124503.
Physics. 2019;151(21). 55. Brygoo S, Millot M, Loubeyre P, et al. Analysis of laser shock ex‐
31. González‐Cataldo F, Soubiran F, Peterson H, Militzer B. Path periments on precompressed samples using a quartz reference
integral Monte Carlo and density functional molecular dynam‐ and application to warm dense hydrogen and helium. Journal of
ics simulations of warm dense MgSiO 3. Physical Review B. Applied Physics. 2015;118(19).
2020;101(2):024107. 56. Militzer B. First principles calculations of shock compressed
32. Militzer B, Gonzalez‐Cataldo F, Zhang S, Driver KP, Soubiran F. fluid helium. Physical review letters. 2006;97(17):175501.
Phys. Rev. E. 2021;103:013203. 57. Militzer B, Hubbard WB. Implications of shock wave experi‐
33. Militzer B, González‐Cataldo F, Zhang S, Whitley HD, Swift DC, ments with precompressed materials for giant planet interiors.
Millot M. Nonideal mixing effects in warm dense matter stud‐ In: . 955. American Institute of Physics. 2007:1395–1398.
ied with first‐principles computer simulations. The Journal of 58. Militzer B. Path integral Monte Carlo and density func‐
Chemical Physics. 2020;153(18). tional molecular dynamics simulations of hot, dense helium.
34. Driver KP, Soubiran F, Zhang S, Militzer B. J. Chem. Phys.. Physical Review B—Condensed Matter and Materials Physics.
2015;143:164507. 2009;79(15):155105.
35. Driver KP, Militzer B. Phys. Rev. B. 2016;93:064101.
36. Pierleoni C, Ceperley D, Bernu B, Magro W. Phys. Rev. Lett..
1994;73:2145.
37. Militzer B, Magro W, Ceperley D. Characterization of the state
of hydrogen at high temperature and density. Contributions to
Plasma Physics. 1999;39(1‐2):151–154.
38. Militzer B, Ceperley DM. Phys. Rev. Lett.. 2000;85:1890.
39. Militzer B, Ceperley D. Path integral Monte Carlo simula‐
tion of the low‐density hydrogen plasma. Physical Review E.
2001;63(6):066404.
40. Militzer B, González‐Cataldo F, Zhang S, Driver KP, Soubiran
F. First‐principles equation of state database for warm dense
matter computation. Physical Review E. 2021;103:013203. doi:
10.1103/PhysRevE.103.013203
41. Zel’Dovich YB, Raizer YP. Physics of shock waves and high‐
temperature hydrodynamic phenomena. Courier Corporation,
2002.
42. ICSC 0183 ‐ nitric acid (> 70% in water). WebPage; .
43. Kritcher A, Döppner T, Swift D, et al. Probing matter at Gbar
pressures at the NIF. High Energy Density Physics. 2014;10:27–
34.
44. Nora R, Theobald W, Betti R, et al. Gigabar spherical
shock generation on the OMEGA laser. Physical review letters.
2015;114(4):045001.