When ZrO2 powders are prepared using any number of precipitation methodologies, the resulting structure before calcination is amorphous [19-22]. Detailed studies by Chadwick et al.  have shown that the amorphous gel has both OZr3 and OZr4 environments in approximately equal proportions, as evidenced from the two 17O NMR peaks at 405 ppm and 303 ppm, respectively. EXAFS results show that there is a well-defined oxygen shell with a coordination number of 7 at ~0.214 nm and a second shell with a much smaller coordination number at ~0.342 nm corresponding to the Zr-Zr correlation. The NMR and Zr K-edge EXAFS results unambiguously indicate that the short-range structure of the amorphous gel is monoclinic-like.
Once the amorphous gel is heated, 17O NMR shows that both OZr3 and OZr4 environments remain, although there is an increase in line width. There is also an increase in the isotropic chemical shift of the peaks, especially the OZr4 peak, which moves from 303 to 321 ppm as the gel starts to crystallize. At the crystallization temperature (approximately 360°C), EXAFS results show that there is a distinct change in the structure, with the oxygen correlation now better fit by two closely spaced shells and a large increase in the coordination number to 12 associated with the Zr-Zr correlation. This is likely because the particles are nanocrystalline.
After crystallization has occurred, there is an additional 17O NMR peak at 374 ppm that corresponds to OZr4 in tetragonal ZrO2. However, the NMR data show that although crystalline tetragonal ZrO2 is forming at the point of crystallization, oxygen is still present as part of the disordered ZrO2. Upon further heating, the tetragonal content increases significantly, but at no time is there complete elimination of the disordered ZrO2.
In addition, ‘H NMR results show that, even after crystallization, it is not correct to describe the sample composition as ZrO2. Data show that, after being heated to 300°C, the sample’s composition is ZrO142(OH)116. At 500°C, well above the crystallization temperature, it is still ZrO176(OH)048. The OH – content found below this temperature is not related to the usual surface hydroxylation upon exposure to the atmosphere, instead the hydroxyls are structural units within the sample. Above 700°C, the hydroxide content is no longer measurable and with subsequent heating the sample changes to the monoclinic structure. Hence, the reaction for the formation of zirconia by precipitation can be described as:
Zr4O(8_x)(OH)2x — ZrO(2_yl2)(OH)y(amorphous) —
ZrO(2_z l2)(OH )z (tetragonal, crystalline;monoclinic _ like, disordered) — ZrO2 (monoclinic, crystalline)
There is an initial metal hydroxide that becomes an amorphous oxide containing hydroxyls. With heating, some of these hydroxyls are lost, resulting in the formation of a mixture with more-ordered tetragonal and less-ordered monoclinic components. With further heating, the eventual crystalline product becomes monoclinic.
In contrast, an EXAFS analysis from amorphous zirconia films of nominal ZrO2 composition, as opposed to a hydroxide composition, found that the local structure in amorphous ZrO2 can be described by an eightfold Zr-O shell widely spread between 0.19 and 0.32 nm with a distinct peak at 0.216 nm consisting of four oxygen nearest neighbors. The average Zr-O coordination distance is 0.255 nm for all eight oxygen neighbors. The local structure also consists of a very broad Zr-Zr shell at about 0.41 nm with 12 next nearest neighbors .