Crystal Structure of N₂O₃

The crystal structures of many "textbook" compounds known for a long time still have to be determined. One example is $N$₂O₃, which has been discovered by Gay-Lussac. In spite of several attempts the crystal structure was unknown up to now.

Indigo-blue $N$₂O₃ exists as a pure phase only in the solid state, melting at approximately $-100oC$. Reed and Lipscomb discovered that single crystals become polycrystalline due to a phase transformation when cooled below $-125oC$. Therefore, extensive powder diffraction experiments, using the modified Guinier technique, were performed in order to optimize crystal growth conditions. The samples, prepared from $NO$₂ and NO, showed a very complex behaviour (Figure 1). Depending on temperature and thermal treatment we were able to identify six different powder diagrams besides those of $N$₂O₄.

Figure 1: Investigation with the modified Guinier technique of $N$₂O₃ containing samples which were sealed in X-ray cappilaries under vacuum (I) and 1 bar NO (II), respectively. Arrows indicate directions of temperature changes (numbers in $oC/h$; dotted lines for T changes without X-ray observations, full lines for simultaneous X-ray diffraction). Cooling always starts from room temperature, k and m denote the diagrams of cubic and monoclinic $N$₂O₄, respectively. Diagrams A to D are observed together with those of $N$₂O₄ in contrast to diagrams E and F.

Above approximately $-150oC$ the intense powder lines of the cubic diagram k of $N$₂O₄ occur with all samples. After quenching to $-190oC$, diagram k is observed at $-160oC$ when the amorphous sample (I) is slowly heated up. A little later, diagram A of $N$₂O₃ appears in addition. Obviously, $N$₂O₄ crystallizes easier than $N$₂O₃. Therefore, it is surprising that with both a quenched or a slowly cooled sample (II), respectively, the diagrams (E,F) after crystallization of the amorphous phase do not show any lines of $N$₂O₄. At $-150oC$ in both cases the cubic diagram of $N$₂O₄ (k) together with diagram B of $N$₂O₃ is developed. These results are best explained with the existence of at least one hitherto unknown oxide below $-150oC$, the composition of which has to be between $N$₂O₃ and $N$₂O₄. Diagrams A to D obviously belong to different modifications of $N$₂O₃.

Figure 2: Projection of the orthorhombic unit cell of $N$₂O₃ (phase B). The distances $[pm]$ and angles $[o]$ are: N1-N2 = 189.06(6), N1-O1 = 111.96(6), N2-O2 = 120.87(6), N2-O3 = 120.57(5); O1-N1-N2 = 105.12(4), O2-N2-O3 = 128.61(5), Co2-N2-N1 = 119.55(4), O3-N2-N1 = 111.84(3)

A single crystal of $N$₂O₃ was grown after the compound had been enriched locally in the sample by zone melting. Probably this is the reason why a crystal of phase B grew, in spite of the fact that in powder diagrams phase A was reproducibly observed in the same temperature range. Figure 2 shows a projection of the structure on (100) of the orthorhombic unit cell. As had been proposed early and supported by IR-, Raman- and NMR-experiments, $N$₂O₃ has to be described as nitroso-nitro compound. There are only slight changes in geometric details determined in the solid state compared to the molecular shape as revealed from microwave spectra in the gas phase. Especially the very long N-N distance (189.1 pm) is rather close to the value (186.4 pm) in the gas phase, and represents a Pauling bond order of about 0.2. A formal description as nitrosyl nitrite is also supported by the N-O distances and the O-N-O angle (N1-O1=112 pm compared to 115 pm in NO and 106.5 pm in $NO+,N2-O\approx 121\; pm$ and O-N-O=128.6$o$ in comparison with 118 pm and 134.3$o$ in $N$₂O₄ or 124 pm and $115o$ in $NO$₂-, respectively). The $N$₂O₃ molecule is almost planar; on the average the atoms are only 1.5 pm off the least squares plane. However, as the angles around N2 are adding up to exactly $360o$, this deviation is mainly due to the position of O1, which lies 7.8 pm above the plane through N1, N2, O2 and O3 (mean deviation 0.3 pm). The nitrosyl group is therefore tilted by $3.7o$ around the N-N bond relative to the nitrite unit. The planarity of $N$₂O₄ was explained as a consequence of bonding interaction between oxygen atoms. This explanation seems also reasonable for $N$₂O₃ as the O-O distance in $N$₂O₃ is even shorter than in $N$₂O₄, despite the longer N-N distance. Recently published quantum mechanical calculations of the electronic structure of $N$₂O₃ were performed for the planar geometry without further explanation. The calculated N-N distances are 10-15% shorter than the experimental value for all basis sets tested.

(A. Simon, J. Horakh, A. Obermeyer, H. Borrmann)

From the yearbook of the institute ("Wissenschaftlicher Tätigkeitsbericht") 1991

webmaster