Persistence and Eventual Demise of Oxygen Molecules at Terapascal Pressures

Abstract

Computational searches for structures of solid oxygen under pressures in the multi TPa range have been carried out using density-functional-theory methods. We find that molecular oxygen persists to about 1.9 TPa at which it transforms into a semiconducting square spiral-like polymeric structure (I41/acd) with a band gap of ∼3.0 eV. Solid oxygen forms a metallic zig-zag chain-like structure (Cmcm) at about 3.0 TPa, but the chains in each layer gradually merge as the pressure is increased and a structure of F mmm symmetry forms at about 9.5 TPa in which each atom has four nearest neighbors. The superconducting properties of molecular oxygen do not vary much with compression, although the structure becomes more symmetric. The electronic properties of oxygen have a complex evolution with pressure, swapping between insulating, semiconducting and metallic. Among the simple molecules studied at high pressures, oxygen has attracted particular attention due to its fundamental importance and intriguing properties. [1][2][3][4][5] For example, oxygen is the only known elemental molecule that exhibits magnetism, which adds substantial complexity to its phase diagram. When liquid oxygen is cooled at ambient pressure it undergoes a sequence of transitions to the γ, β and α solid phases at 54.39, 43.76 and 23.88 K, respectively [6]. Upon increase of pressure the monoclinic α-O 2 (C2/m) phase transforms into an orthorhombic δ-O 2 (F mmm) phase at about 3 GPa and to -O 2 at about 10 GPa. The structure of -O 2 has only recently been solved by x-ray diffraction (XRD) studies of single crystal [3] and powder [4] samples, although its vibrational spectrum was reported more than 20 years earlier [7]. The unit cell of -O 2 has C2/m symmetry and contains four O 2 molecules forming O 8 units. Both α and δ-O 2 are antiferromagnetic and the magnetic collapse at the δ-transition was predicted using molecular dynamics simulations [8]. This breakdown of the long-range antiferromagnetic order at about 8 GPa was recently observed in a neutron scattering experiment [2]. Density functional theory (DFT) studies have found another chain-like structure to be energetically slightly more favorable than the O 8 structure [9,10], although it has not been observed in experiments.Recent experiments [5] have shown that insulating -O 2 remains stable up to about 96 GPa before undergoing a continuous displacive and isosymmetric transition to the ζ phase, in agreement with earlier predictions [10]. The metallic ζ phase [11,12] has C2/m symmetry [2] and superconducts at temperatures below 0.6 K [1]. Manybody perturbation theory GW calculations [13,14] have suggested that the metal-insulator transition occurs at lower pressures than the measured -ζ transition pressure of 96 GPa. The above information, however, covers only a small part of the phase diagram and rather little is known about pure oxygen at pressures above 100 GPa.A previous DFT study reported that oxygen molecules persist to at least 250 GPa [10]. It is interesting to speculate about the highest p...

Density functional theory in the solid state

“…Through the application of AIRSS, unexpected phase transitions have been predicted in iron [57], carbon [58] and oxygen [59] at pressures up to 1 PPa. Aluminium was predicted to undergo a transition to a complex, incommensurate, host guest phase at around 3 TPa [60].…”

Cagelike Diamondoid Nitrogen at High Pressures

“…The extremely strong triple N≡N bond dissociates into three weaker single N-N bonds at a modest pressure >150 GPa [1][2][3]. In contrast, the molecular dissociation for oxygen is shown to happen at 1920 GPa [4][5] and for hydrogen at 500 GPa [6], even though these molecules have much weaker intra-molecular bonds. Because of the exceedingly large difference in energy between the single N-N and triple N≡N bonds, singly-bonded polymeric nitrogen has the potential to become an excellent high-energy-density material for energy storage, propellants, and explosives.…”

“…Because density functional calculations usually lead to a considerable underestimation of the energy gap, the actual band gaps are expected to be much larger. The insulating state is the result of the complete localization of valence electrons, similar to the monatomic O 4 phase of oxygen [4][5]. The increase of the N 10 -cage band gap with pressure is the result of the competition of two effects: (i) on one hand, the compression and the consequent shortening of the bond length will widen both valence and conduction bands, and therefore tends to reduce the gap; (ii) on the other hand, the stronger coupling of the sp 2 (or sp 3 ) orbitals at the neighboring N atoms will lower the energy of the bonding states (valence bands) and increase the energy of the anti-bonding states (conduction bands), therefore leads to an increase of the gap.…”

First-principles equation of state and electronic properties of warm dense oxygen

“…X-ray diffraction and optical experiments reveal that oxygen condenses to a molecular solid with a rich phase diagram made up of at least ten different structural phases [1][2][3][4][5][6] . Static compression experiments on the solid have been performed up to 1.3 Mbar and 650 K 1 .…”