High-Pressure Synthesis and Study of NO⁺NO₃⁻ and NO₂⁺NO₃⁻ Ionic Solids

Abstract

Nitrosonium-nitrate NO⁺NO₃⁻ and dinitrogen pentoxide NO₂⁺NO₃⁻ ionic crystals were synthesized by laser heating of a condensed oxygen-rich O₂-N₂ mixture compressed to different pressures, up to 40 GPa, in a diamond anvil cell (DAC). High-pressure/high-temperature Raman and X-ray diffraction studies of synthesized samples disclosed a transformation of NO⁺NO₃⁻ compound to NO₂⁺NO₃⁻ crystal at temperatures above ambient and pressures below 9 GPa. High-pressure experiments revealed previously unreported bands in Raman spectra of NO⁺NO₃⁻ and NO₂⁺NO₃⁻ ionic crystals. Structural properties of both ionic compounds are analyzed. Obtained experimental results support a hypothesis of a rotational disorder of NO⁺ complexes in NO⁺NO₃⁻ and indicate a rotational disorder of ionic complexes in NO₂⁺NO₃⁻ solid.

1. Introduction

Application of high pressure can convert simple molecular structures of elemental and heteronuclear systems into nonmolecular solids with extended or “infinite” atomic lattices [1]. Examples of such extended solids are nonmolecular nitrogen [2–4], metalic oxygen [5], polymeric form of carbon dioxide [6], and carbon monoxide [7]. Such pressure-induced molecular-to-nonmolecular phase transitions are one of the ways to reduce a total energy of the molecular crystal via local (polymerization) or complete (metallization) delocalization of intramolecular electrons between adjacent molecules. Another alternative of a more stable form of chemical bonds perturbed by pressure and temperature variation is ionization with the onset of charge-transfer interactions. The first reported example of the neutral-to-ionic solid transformation concerns the organic charge-transfer compound tetrathiafulvalene-chloranil [8, 9]. In this material, the pressure or temperature variation promotes either neutral or ionic form, which is determined by the electrostatic energy contribution to the total energy of the donor-acceptor molecules. The similar neutral-to-ionic transformation has been reported in several simple molecular systems as well. The formation of H₂⁺H₂⁻ dimmer at 150 GPa in phase III of solid hydrogen [10] was proposed to explain the discontinuous changes of the Raman and infrared vibron frequencies [11, 12]. The nitrogen dioxide dimmer, N₂O₄, transforms to the ionic NO⁺NO₃⁻ complex by temperature variation at ambient pressure [13, 14] and by pressure variation after laser heating [15, 16]. Recently, the same ionic solid was synthesized from N₂O and N₂O₄ molecular crystals by laser heating at high pressures [17–21]. From the latter studies, one can infer a possibility of a direct synthesis of NO⁺NO₃⁻ crystal from oxygen and nitrogen reactants. In particular, the primary dissociation of N₂O on N₂ and O₂ at high temperatures and pressures below 30 GPa was evident from the Raman spectra [17, 18]. On the other hand, X-ray diffraction studies showed that nitrosonium nitrate is denser than other nitrogen-oxygen assemblages [19]. Direct confirmation of the O₂+N₂→NO⁺NO₃⁻ reaction has been done during high-pressure synthesis of CrO₂ from Cr and O₂, where addition of nitrogen into the Cr+O₂ reactants environment resulted in the formation of NO⁺NO₃⁻ compound [22].  Recent studies confirmed the transformation of a compressed mixture of O₂ and N₂ to NO⁺NO₃⁻ induced either by laser light [23] or by X-ray radiation [24]. The latter study also showed that at low pressures, the ionic form of dinitrogen pentoxide NO₂⁺NO₃⁻ is rather stable than nitrosonium nitrate.

A considerable volume of experimental studies has been carried out in order to ascertain the high-pressure structure of NO⁺NO₃⁻ ionic crystal. Based on the energy-dispersive X-ray diffraction studies, aragonite-type structure was proposed for the crystal polymorph of NO⁺NO₃⁻ at pressures above 5 GPa and two space groups, 𝑃2₁𝑐𝑛 and 𝑃𝑚𝑐𝑛, equally well accounted for the observed Bragg reflections [17, 19]. Angle-dispersive X-ray diffraction study revealed the monoclinic 𝑃2₁/𝑚 structure of NO⁺NO₃⁻ at low pressures [24]. Several other space groups were also proposed for the low-pressure phase of NO⁺NO₃⁻ [23], however with the limited X-ray diffraction data quality and a nondecisive conclusion about the space group. At the same time, no experimental information is available on crystal structure of NO₂⁺NO₃⁻ at high pressures.

In this work, we have explored the question of stability of high-pressure/high-temperature forms of nitrogen oxides obtained by laser heating of oxygen rich O₂-N₂ mixture. In addition to the expected NO⁺NO₃⁻ ionic crystal, dinitrogen pentoxide NO₂⁺NO₃⁻ ionic solid is formed at pressures below 30 GPa. We report the results of high-pressure/high-temperature Raman and X-ray diffraction studies on the synthesized compounds. The conversion of NO⁺NO₃⁻ into NO₂⁺NO₃⁻ at high temperatures and pressure below 9 GPa is documented. Structural properties of NO⁺NO₃⁻ and NO₂⁺NO₃⁻ ionic compounds as a function of pressure and temperature are discussed

2. Experimental Methods

Investigated samples consisted of a mixture of the high-purity (99.999%) oxygen and nitrogen (99.99%) gases condensed at ambient pressure and liquid nitrogen temperature. Diamond anvil cells (DACs) were merged and closed in the liquid phase of studied O₂-N₂ assemblages. The small crystals of ruby in the gasket hole provided a pressure calibration by standard fluorescence technique. A predominant content of oxygen in condensed O₂-N₂ mixtures was ensured by adjusting the partial pressures of oxygen and nitrogen gases during the condensation process. After a compression to a desired pressure, an O₂-N₂ mixture was heated by Nd:YLF (1067 nm) laser. Temperature estimations of the heated area during laser heating relied on observed hot spot intensities [25]. Several experiments with O₂-N₂ mixture covered the range of pressure before laser heating from 10 GPa to 40 GPa, and the estimated temperature varied in the range from 1300 to 2000 K. Raman spectra from samples before and after laser heating were collected at Bayreuth Geo-Institute using Dilor XY systems for the Raman measurements with the 5145 Å Ar-ion laser excitation line and the incident laser power in the range 200–250 mW. The Dilor XY Raman spectrometer was calibrated using the Γ₂₅ phonon of diamond-structured Si (𝐹𝑑-3𝑚) and provided the data collection in the 50–3000 cm⁻¹ range. The peaks were analyzed using the TOPAS-Academic software [26] with pseudo-Voight function describing a peak profile. We estimate a resolution of 1 cm⁻¹ for Raman peak position.

In situ high-pressure X-ray diffraction measurements of O₂-N₂ mixture were done at ID30 beam line of European Synchrotron Radiation Facility (ESRF, Grenoble, France) before and after heating by Nd:YAG laser. A procedure of fitting of outgoing thermal radiation (the energy distribution of its emitted light) to a Planck radiation formula yielded the temperature of the heated area of approximately 2000 K. Diffraction patterns were collected using focused monochromatic (𝜆=0.3738 Å) X-ray radiation with an image plate detector (MAR345). One-dimensional 2𝜃 dependences of the X-ray diffracted intensities were obtained by integration of two-dimensional diffraction images using the ESRF Fit2D software [27]. Small crystals of ruby provided a pressure calibration of the sample by fluorescence technique.

3. Results and Discussion

3.1. High-Pressure Synthesis and Phase Relations of NO⁺NO₃⁻

The Raman spectra of the samples at high pressures and room temperature before laser heating contained the well-known features of the O₂-N₂ solid mixture. The intense oxygen vibron at about 1575 cm⁻¹, the nitrogen stretching mode at about 2365 cm⁻¹, and the lattice modes of O₂-N₂ mixture in a low-frequency part of spectra characterized the Raman spectra collected at high pressure before laser heating (see the selected examples in Figures 1(a) and 1(b)). The pressure evolution of a characteristic splitting of O₂ and N₂ stretching modes as well as the pressure dependence of the relative intensities of the O₂ and N₂ vibrons were compared with the available studies of O₂-N₂ mixture [28, 29]. We estimate the oxygen concentration in all studied O₂-N₂ samples to be within the range from 70% mol up to almost 100%. We did not control precisely the oxygen content in the O₂-N₂ mixture since the results of the laser heating experiments were independent on the variation of oxygen concentration. The compression to pressures above 30 GPa and subsequent laser heating in the range of temperatures from 1300 K to 2000 K of the O₂-N₂ mixture resulted in the formation of nitrosonium-nitrate, NO⁺NO₃⁻, as compared to similar spectra of the earlier studies on NO⁺NO₃⁻ [15–18]. The Raman peaks of this ionic compound can be easily separated from the peaks of nontransformed O-N mixture by comparing the Raman spectra collected from the nonheated (Figure 1(c)) and heated regions of the sample (Figures 1(d) and 1(e)). In the high-frequency range of the Raman spectrum, NO⁺NO₃⁻ ionic crystal exhibits the peaks associated with the intramolecular vibrations of the NO₃⁻ and NO⁺ groups. The intense 𝜈(NO⁺) stretching mode of NO⁺ ion at 2268 cm⁻¹ and the Raman bands of NO₃⁻ molecular complex 𝜈₁ (1112cm⁻¹), 𝜈₂ (825cm⁻¹), 𝜈₃ (1470cm⁻¹), 𝜈₄ (750cm⁻¹), and overtone of 𝜈₂ (1650cm⁻¹) are identified in Figure 1(d). The lower frequency range of the Raman spectra exhibits the peaks that can be related to the lattice vibrations of the NO₃⁻ and NO⁺ ions and to libration, that is, rotation, of these ionic groups. It is interesting to note almost identical resemblance of the lower part of the Raman spectrum in  Figure 1(d) with the respective part of the Raman spectra reported by Yoo et al. [18], Song et al. [20], and  Somayazulu et al. [17]. This similarity indicates an invariance of a crystal structure of NO⁺NO₃⁻ produced from different starting forms of N-O system.

Figure 1 

Vibrational modes assignment for O2-N2 mixture before laser heating at (a) 15 GPa and (b) 39 GPa; (d) NO+NO3− ionic crystal at 32 GPa  and (e) its lattice modes at 30 GPa after laser heating. Raman spectrum of untransformed O2-N2 mixture after laser heating at 32 GPa (c) is shown for comparison.

High-pressure polymorphism of NO⁺NO₃⁻ can be inferred from X-ray diffraction studies [17, 19, 24]. This conclusion corroborates with the Raman spectroscopy data of this work and previous studies [18, 21], where the pressure dependence of mode frequencies of NO⁺NO₃⁻ exhibits the change of a slope at ~5 GPa (Figure 2). This singularity delimits fairly well the pressure ranges of observation of monoclinic [24] and orthorhombic [17, 19] phases of NO⁺NO₃⁻. One more change of a slope can be identified in Figure 2 at ~22 GPa indicating an eventual structural alteration in NO⁺NO₃⁻ system at higher pressures.

(a)

(b)

Figure 2 

Pressure variation of the vibrational frequencies obtained on decompression of NO+NO3− (solid squares) and N2O4 (opened squares) molecular crystals.

A symmetrization of high-pressure NO⁺NO₃⁻ crystal structure with decreasing pressure has been pointed out as a possible general phase transformation path [18]. The asymmetry of stretching mode of NO⁺ ion (~2270 cm⁻¹) that visibly vanishes on decompression at pressures about 5 GPa (compare respective insets in Figures 1(d) and 3(a)) can be a consequence of such a symmetrization. In addition to the asymmetry of the 𝜈(NO⁺) stretching mode, we could clearly detect two weak peaks in the vicinity that were not reported in previous studies. These two peaks are identified in the insets of Figures 1(d) and 3(a) at ~2200 cm⁻¹ as 𝜈^′_NO⁺ and 𝜈^′′_NO⁺. The obtained frequencies for the latter peaks and their pressure behavior suggest that 𝜈^′_NO⁺ and 𝜈^′′_NO⁺ modes are not the overtones or combination bands. The intensities of these peaks are weakly affected by the pressure variation; as a result, no correlation between pressure behavior of intensities of these peaks and symmetrization of stretching mode of NO⁺ ion could be seen. A plausible assumption about the origin of these peaks could be an orientational (dynamic or static) disorder of NO⁺ ionic groups. The orientational disorder may result in metastable orientations of NO⁺ ions in the structure of NO⁺NO₃⁻ with slightly different stretching frequencies of NO⁺ ions.

Figure 3 

Raman spectra of (a) NO+NO3− ionic crystal at 2.8 GPa, (b) a mixture of NO+NO3− ionic and N2O4 molecular crystals at 2 GPa, and (c) N2O4 molecular crystal at 1.7 GPa, showing the phase transition of N2O4 molecular compound from ionic to neutral form. The assignment of the internal vibrational modes in (c) corresponds to symmetrical N2O4 dimer [14].

In order to obtain an additional insight into structural properties of NO⁺NO₃⁻ at high pressures, one can confront the structural information obtained from X-ray diffraction studies with Raman spectroscopy data. Specifically, in case of a known or assumed crystal structure, the correlation analysis [30, 31] allows establishing the number, symmetry, and spectral activity of the external and internal optics modes. Since no such study has been done so far for NO⁺NO₃⁻, we present briefly the main results of correlation analysis for low-pressure monoclinic phase (𝑃2₁/𝑚) whose structure was established from X-ray diffraction measurements [24]. We also discuss the implications of correlation analysis under assumption of orthorhombic phases 𝑃2₁𝑐𝑛 and 𝑃𝑚𝑐𝑛 with aragonite-type structure as a most plausible structural models of high-pressure phase of NO⁺NO₃⁻ inferred from earlier X-ray diffraction studies [17, 19].

Table 1 summarizes the result of the correlation analysis for the monoclinic structure of NO⁺NO₃⁻. The correlations between molecular and site group species of NO₃⁻ ions result in splitting of 𝜈₃ and 𝜈₄ modes, all of them being infrared and Raman active. Indeed, our Raman spectra (Figure 3(a)) as well as Raman spectra of previous studies [17, 20, 23] clearly show a splitting of 𝜈₄ fundamental. Besides, somewhat broad and asymmetric 𝜈₃ band detected in our study suggests a poorly resolved doublet. This doublet was clearly resolved in the study by Sihachakr and Loubeyre [23]. One can observe, consequently, six internal optic modes originated by NO₃⁻ ions and one NO⁺ stretching mode in both Raman and IR spectra of monoclinic NO⁺NO₃⁻ crystal. This prediction is in excellent agreement with our Raman and previous Raman and IR studies.

Table 1 

Correlation between point group, site group, and factor group symmetry species and their Raman and IR activities for the NO⁺ and NO₃⁻ molecular ions in the monoclinic, 𝐶²_2ℎ (𝑃2₁/𝑚), NO⁺NO₃⁻ crystal.

Superposition of translational and rotational motions of NO₃⁻ and NO⁺ ions originates the external optic modes of NO⁺NO₃⁻. In case of 𝑃2₁/𝑚 space group, one can find the following irreducible representations of the lattice vibrations and libration modes [30, 31]:Γ^Lattice_NO⁺_NO3⁻=4A_g^(R)+2A_u^(IR)+2B_g^(R)+4B_u^(IR),Γ^Libration_NO⁺_NO3⁻=2A_g^(R)+3A_u^(IR)+3B_g^(R)+2B_u^(IR), (1)

The spectral activity of each symmetry species is indicated by a superscript (R) for Raman and (IR) for infrared active modes. Table 2(a) summarizes the total irreducible representation of NO⁺NO₃⁻ crystal under assumption of 𝑃2₁/𝑚 space group. Full factor group for external vibrations of monoclinic NO⁺NO₃⁻ contains 11 Raman active and 8 IR active translational and libration modes. Raman spectra obtained in our and in the previous studies [17, 18, 20] exhibit only 8-resolved peaks associated with the external modes at high pressures and room temperatures; this number reduces to 6 peaks at pressures below 5 GPa (see Figures 1(d), 1(e), and 3(a)). The discrepancy between the number of observed and predicted external Raman modes can be attributed to a thermal overlapping of the peaks. Indeed, all 11 peaks are resolved in the low-frequency range of low-temperature Raman spectra of NO⁺NO₃⁻ [21].

Table 2 

Total irreducible representation and spectral activity of the monoclinic and aragonite-type NO⁺NO₃⁻ crystal for (a) 𝐶²_2ℎ (𝑃2₁/𝑚), (b) 𝐶⁹_2𝜈 (𝑃𝑛𝑎2₁), and (c) 𝐷¹⁶_2ℎ (𝑃𝑛𝑚𝑎) space group.

Analogously, a site group correlation analysis under assumption of orthorhombic structure of NO⁺NO₃⁻ with 𝑃2₁𝑐𝑛 and 𝑃𝑚𝑐𝑛 space groups leads to six internal optic modes originated by NO₃⁻ ions and one NO⁺ stretching mode, all being active in both Raman and IR spectra. The respective total irreducible representations are summarized in Tables 2(b) and 2(c). The factor group symmetry species are labeled using standard axial settings (space groups 𝑃𝑛𝑎2₁ and 𝑃𝑛𝑚𝑎, resp.). The correspondence between the crystallographic axis and irreducible representations of factor groups of 𝐶₂𝜈⁹ and 𝐷¹⁶_2ℎ in different settings can be found elsewhere [32]. As the result of the 𝐶₁ site symmetry of NO₃⁻ and NO⁺ ions, 41 Raman active and 30 IR active external vibrations are expected in case of 𝑃2₁𝑐𝑛 space group. In case of 𝑃𝑚𝑐𝑛 space group of NO⁺NO₃⁻ crystal, either 𝐶₁ or 𝐶_𝑠 site symmetry can accommodate four NO₃⁻ and four NO⁺ ions. Table 2(c) shows the results of correlation analysis assuming 𝐶_𝑠 site symmetry for both ions. The choice of this site symmetry is due to the fact that crystallographic positions occupied by ions in aragonite-type crystal have the same 𝐶_𝑠 symmetry [33]. External vibrations in this case consist of 12 Raman and 7 IR active translational modes, and of 10 Raman and 10 IR active libration modes.

A considerable discrepancy between the number of experimentally observed and predicted external Raman active vibrations in case of orthorhombic phase of NO⁺NO₃⁻ can reflect a structural peculiarities of high-pressure polymorph of NO⁺NO₃⁻. As it was shown for aragonite, CaCO₃, and isomorphic KNO₃ [34], one can expect a considerable reduction in a number of external modes if a structural model for a compound is a slightly distorted variant of a structure with higher symmetry. In such a distorted structure, certain vibrational modes can be associated with a motion of ions in planes that have a symmetry closely coinciding with a higher base symmetry. Consequently, these modes should be determined using the high-symmetry model. In particular, if a higher-symmetry base structure has a reduced translational symmetry in such a plane, one can expect to observe only lattice and libration modes originated by ions motion which is translational invariant with respect to a reduced unit cell [34]. It is worth noting, in this regard, that the established structure of the low-pressure phase of NO⁺NO₃⁻ has the monoclinic unit cell [24] with two molecular units per unit cell in contrast to four molecular units of the proposed aragonite-type structure [17]. It might be possible that this monoclinic structure is a descendent of the higher-symmetry base structure of the high-pressure phase of NO⁺NO₃⁻ with reduced unit cell.

It is worthwhile to mention, in conclusion of this section, the pressure-induced transformation of ionic NO⁺NO₃⁻ to neutral N₂O₄ molecular crystal that occurs at pressures below 2 GPa (Figures 3(b) and 3(c)). We found that the pressure of this transformation is highly dependent on a specific pressure-time path followed by the sample. In some experiments on decompression of NO⁺NO₃⁻ at room temperatures, the ionic form could be retained up to a liquid phase.

3.2. High-Pressure Synthesis and Structural Properties of NO₂⁺NO₃⁻

Laser heating of oxygen rich O₂-N₂ mixture at pressures below 30 GPa produced a mixture of two ionic crystals: NO⁺NO₃⁻ and additional phase identified as ionic form of dinitrogen pentoxide, NO₂⁺NO₃⁻ (see selected examples in Figures 4(a) and 4(b)). In order to establish precisely the nature of the additional phase, we carried out a series of Raman and diffraction experiments at high pressures and temperatures. The identification of NO₂⁺NO₃⁻ was facilitated by the fact of a complete conversion of NO⁺NO₃⁻ ionic crystal to NO₂⁺NO₃⁻ at pressures below 9 GPa and temperatures above ambient temperature. This transformation was confirmed by both Raman and X-ray diffraction measurements (Figures 4(c) and 5(a)). The determined earlier hexagonal (𝑃6₃/𝑚𝑚𝑐, two formula units per unit cell) crystal structure of NO₂⁺NO₃⁻ [35] perfectly accounted for diffraction patterns obtained at pressures below 9 GPa (Figure 5(b)). The obtained Raman spectra of NO₂⁺NO₃⁻ also showed good agreement with the previous studies [36, 37]. The Raman bands of NO₃⁻ and NO₂⁺ molecular complexes are indicated in Figure 4(c). Under assumption of hexagonal crystal structure, the correlation analysis accounts for the presence in the Raman spectrum of 𝜈₁ (1070 cm⁻¹) and 𝜈₄ (736 cm⁻¹) bands of NO₃⁻ molecular complex, and 𝜈₁ (1403 cm⁻¹) and 𝜈₂ (505 cm⁻¹) bands of NO₂⁺ ion (see Table 3(a)). A very weak and broad band around 1370 cm⁻¹ (not shown in Figure 4(c)) was detected in most of collected Raman spectra of NO₂⁺NO₃⁻. We assigned this feature to the antisymmetric stretching mode 𝜈₃ of NO₃⁻. Extrapolation to zero pressure of the 𝜈₃ mode frequency (Figure 6) gives a similar value (~1350 cm⁻¹) as the respective band frequency observed at ambient pressure in the previous studies [36, 37]. The multiple bands indicated in Figure 4(c) as 2𝜈₂ have been explained by Wilson and Christe [36] as a split overtone of the NO₂⁺ deformation mode. The pressure behavior of these bands is a conclusive evidence of Wilson’s assignment (see Figure 6).

Table 3 

Correlation between point group, site group, and factor group symmetry species and their Raman and IR activities for the NO₂⁺ and NO₃⁻ molecular ions in the NO₂⁺NO₃⁻ crystal assuming (a) 𝐷⁴_6ℎ (𝑃6₃/𝑚𝑚𝑐) and (b) 𝐷¹⁷_2ℎ (𝐶𝑚𝑐𝑚) space group.

Figure 4 

Raman spectra and vibrational modes assignment for (a) O2-N2 mixture before laser heating,  (b) mixture of NO+NO3− and NO2+NO3− ionic solids after laser heating (vibrational modes of NO+NO3− are identified), (c) NO2+NO3− at 3.7 GPa and 500 K, (d) NO2+NO3− at 25 GPa and room temperature.

(a)

(b)

Figure 5 

(a) Evolution of the diffraction patterns of NO+NO3− and NO2+NO3− ionic solids on decompression at 500 K. The transformation of the phase mixture to pure NO2+NO3− is confirmed by the Rietveld refinement (b) of the hexagonal structural model (𝑃63/𝑚𝑚𝑐) of NO2+NO3− at 8.2 GPa and 500 K. The refined lattice parameters are 𝑎=5.1377(1); 𝑐=5.8930(1). The 𝑥, 𝑦, and 𝑧 fractional atomic coordinates for two nitrogen and two oxygen atoms in asymmetric unit are N(1)=0,0,1/4; N(2)=1/3,2/3,3/4; O(1)=0.1792,0.3583,1/4; O(2)=1/3,2/3,0.5631.

(a)

(b)

Figure 6 

Raman shifts as a function of pressure for NO2+NO3− ionic crystal (a). Opened squares refer to ambient temperature data obtained on compression and solid squares represent the data obtained at 600 K on decompression. Green and red lines correspond to 𝜈3 mode pressure evolution at room temperature and at 600 K, respectively. The influence of temperature on the vibrational frequencies can be seen at a zoomed scale (b).

The important different feature in NO₂⁺NO₃⁻ Raman spectra of this study, however, was the presence of antisymmetric stretching mode 𝜈₃ (2260 cm⁻¹) of NO₂⁺ complex. Correlation analysis cannot explain this Raman band, which is silent under assumption of hexagonal structure of NO₂⁺NO₃⁻ (see Table 3(a)). A plausible explanation of the Raman activity of the 𝜈₃ band can be based on a bent structure of rotating NO₂⁺ ion. A reduction of NO₂⁺ ion’s symmetry from a linear to a bent one implies a reduction of symmetry of NO₂⁺NO₃⁻ crystal structure. As it was pointed out by Simon et al. [38], a very small deviation from hexagonal unit cell toward an orthorhombic one is needed in order to explain the low-temperature single-crystal diffraction data. Assuming the orthorhombic (𝐶𝑚𝑐𝑚 space group, four formula units per unit cell) for NO₂⁺NO₃⁻ crystal and a bent configuration of NO₂⁺ ion, one can deduce the Raman activity of all three normal vibrational modes of the NO₂⁺ cation observed in our Raman spectra (see Table 3(b)). As for NO₃⁻ ions vibrational modes, the symmetry reduction of the unit cell has to produce a splitting of the 𝜈₃ antisymmetric stretching vibration and 𝜈₄ in plane deformation mode, analogously to the case of nitrosonium-nitrate NO⁺NO₃⁻. Indeed, at pressures above ~20 GPa, we observed a splitting of 𝜈₄ mode (compare the respective insets of Figures 4(c) and 4(d)). In addition, the low frequency part of the collected Raman spectra suffered strong alterations at high pressures, indicating a considerable structural changes induced by pressure. These results support the suggestion of Simon et al. [38] about orthorhombic  nearly hexagonal structure of NO₂⁺NO₃⁻ and evidence an increase of an orthorhombic distortion with increasing pressure.

Another strong evidence of a bent structure of NO₂⁺ ion is a soft behavior of deformation mode 𝜈₂ as a function of pressure (Figures 6 and 7). Such behavior would reflect the reduction of the bending force constant as the bent O–N–O structure deforms toward a linear configuration with increasing pressure.

(a)

(b)

Figure 7 

Pressure evolution of the Raman spectra of NO2+NO3− at 600 K (a) and Raman spectra of 𝜈1(NO3−) region at 11.3 GPa, 8.4 and 5.6 GPa at ambient temperature (b). A shoulder in the vicinity of 𝜈1(NO3−) vibration mode indicates an orientational and/or site mobility of NO3− ionic complexes in NO2+NO3− ionic crystal.

The temperature effect on the internal vibrational modes of NO₂⁺NO₃⁻ is detailed for selected modes in Figure 6(b). A shift of 𝜈₂(NO₂⁺) Raman peak toward a lower frequency with increasing temperature was observed in high-pressure/high-temperature Raman experiments. This trend can be seen more pronouncedly for overtone bands of 𝜈₂ mode. Such a behavior may indicate a less bent geometry of NO₂⁺ ion in the expanded lattice of NO₂⁺NO₃⁻. It should be pointed out that the temperature increase results in downshifted Raman peak associated with antisymmetric stretching mode 𝜈₃ of NO₂⁺ ion as well. At the same time, all other vibrational modes of NO₂⁺NO₃⁻ are shifted toward a higher frequency with increasing temperature.

It is worth noting that a transformation to a bent configuration of NO₂⁺ ion can be viewed as a removal of degeneration in doubly degenerated Π_u bending mode of a linear ion. One of the components of the Π_u mode reduces 𝐷_∞ℎ symmetry to 𝐶_2𝜈, and other component becomes the rotation of the ion. Consequently, a rotation of NO₂⁺ ion in NO₂⁺NO₃⁻ crystal can be expected just from symmetry considerations. As was already discussed by Wilson and Christe [36], namely, this rotation may result in an average linear structure of NO₂⁺ ion detected by X-ray diffraction technique.

An orientational disorder of NO₃⁻ ionic groups in NO₂⁺NO₃⁻ crystal can be inferred from our Raman data as well. Numerous studies on nitrate salts showed that the totally symmetric vibration 𝜈₁ of nitrate ions may have a complex structure; in particular, an anomalous second component is present near the main Raman peak at slightly lower frequencies [39–41]. The appearance of the second component is generally explained by a disorder in nitrate ions position or orientations in the crystal lattice of nitrates [39]. The obtained Raman spectra of NO₂⁺NO₃⁻ exhibit a shoulder next to the symmetric vibration 𝜈₁ (see Figure 7), and analogously to nitrate salts this shoulder can be attributed to a rotational or site disorder of NO₃⁻ ions. It is interesting to note that Wilson and Christe [36] also observed an unexplained peak near the 𝜈₁ stretching mode of NO₃⁻ groups in Raman spectra of NO₂⁺NO₃⁻ collected at low temperatures and ambient pressures. It is possible that the origin of this peak is connected to a disorder of the nitrate ions.

It is worthwhile, in conclusion, to discuss briefly the question of relative stability of two ionic phases of nitrogen oxides, NO⁺NO₃⁻ and NO₂⁺NO₃⁻. Raman and X-ray diffraction data show unambiguously the instability of NO⁺NO₃⁻ ionic solid at pressures below 9 GPa and temperatures above 450 K (perhaps, even at lower temperatures). Figure 8 stresses this point showing the relations between ionic forms of nitrogen oxides and includes the melting line of NO₂⁺NO₃⁻ ionic crystal mapped from our Raman and X-ray diffraction experiments.

Figure 8 

Phase diagram showing the domain of stability of crystalline NO2+NO3− on decompression of a mixture of NO2+NO3− and NO+NO3−. A textured area indicates the eventual extension of the region of (meta)stability of NO+NO3− solid to low pressures.

On the other hand, a stability of NO⁺NO₃⁻ ionic crystal at pressures above 30 GPa and high temperatures was confirmed by numerous laser-heating experiments of our and previous studies. One can expect, consequently, a conversion of NO₂⁺NO₃⁻ to NO⁺NO₃⁻ at sufficiently high pressures. A big pressure region of coexistence of NO⁺NO₃⁻ and NO₂⁺NO₃⁻ phases observed in our study can be attributed to kinetics of the NO⁺NO₃⁻ to NO₂⁺NO₃⁻ transformation. In particular, high-kinetic barrier of NO⁺NO₃⁻ to NO₂⁺NO₃⁻ conversion can explain a stability of NO⁺NO₃⁻ at pressures below 9 GPa and ambient temperature. The high-pressure IR spectra of NO⁺NO₃⁻ [21] showed that a degree of ionicity of NO⁺NO₃⁻ decreases with decreasing pressures. It follows from our results that a reduction of ionicity of NO⁺NO₃⁻ toward a neutral form of nitrogen dioxide is intermediated by another stable ionic form of nitrogen oxide, NO₂⁺NO₃⁻. Most likely, interplay between electrostatic and intraionic covalent bonding energy contributions to a total energy defines a stable ionic form of nitrogen oxide at high pressures.

4. Conclusions

In summary, we have synthesized the ionic forms of nitrogen oxides, NO⁺NO₃⁻ and NO₂⁺NO₃⁻, directly from oxygen-rich O₂-N₂ mixture. A direct high-pressure synthesis of ionic forms of nitrogen oxide provides a convenient way of its study at high pressures and high temperatures. The Raman and X-ray diffraction studies showed that a sequence NO⁺NO₃⁻-NO₂⁺NO₃⁻-N₂O₄ of transformations can be realized with pressure-temperature variation underlying the transition from highly ionic to neutral form of nitrogen oxides. Our results evidence that ionic NO₂⁺NO₃⁻ phase is characterized by increased orientational and/or site disorder of ionic groups.