The structure of the Ca–P solid phase in bone was first identified by deJong in 1926 as a crystalline calcium phosphate similar to geological apatite by chemical analyses and, most importantly, by X-ray diffraction . The X-ray diffraction data was confirmed a few years later .
These findings initiated a flurry of research on a more detailed chemical composition and crystal structure of both geological and synthetic apatites and of bone mineral, initially carried out principally by geologists, crystallographers and chemists, but later by biochemists and physiologists because of the clear potential of this new information to shed light on the biological and physiological functions of bone mineral and as indicators of disorders of the skeletal system.
It soon became clear that there were significant structural and chemical compositional differences between the many different geological hydroxyapatites, synthetic hydroxyapatites, and the apatite crystals found in bone and related skeletal tissues in addition to the very large size of the geological and many of the synthetic apatite crystals, compared with the extremely small particle size of bone mineral.
Further studies were directed in roughly three avenues: continued more careful and complete analytical compositional data of bone mineral, from which it was clearly established that the chemical composition of bone crystals in many ways did not correspond to the chemical compositions of stoichiometric hydroxyapatite. Indeed, the bone crystals were found to contain significant and varying amounts of carbonate and HPO4 ions. Much later it was discovered by a variety of techniques, including solid state NMR , Raman spectroscopy , and inelastic neutron scattering , that the biological bone apatites contain only a very small percentage of the total number of hydroxyl groups present stoichiometric hydroxyapatites.
Other studies clearly pointed out that a substantial fraction of the phosphate ions are situated on the surfaces of the bone mineral crystals and are mainly protonated and in a disordered environment , in contrast to the phosphate ions in lattice positions, which are unprotonated. Also, other uniquely protonated phosphate ions were identified in bone mineral by phosphorus 31 NMR spectroscopy which are not present in synthetic calcium phosphate apatites .
Structural studies were first carried out to determine crystal size by measuring the extent of X-ray diffraction peak broadening , which yielded crystal sizes varying from 31Å to 290Å. More detailed structural data were obtained by the then recently introduced field of electron microscopy and electron diffraction [9-11], which revealed that the bone crystals were thin plates, approximately 500Å long, 250Å wide, and 100Å thick. However, calculations from low angle X-Ray diffraction scattering studies [12-14] were more consistent with the conclusions that the bone crystals were very much smaller than those observed by Robinson et al by TEM, which were rods 250Å long and 50Å thick rods and not platelets. Crystal perfection, another structural property of the individual bone nanocrystals is also a function of the extent of mineralization, as demonstrated by the changes in their wide-angle X-ray diffraction patterns  [Figure 1].
More accurate evaluations of the size and shape of the bone crystals were later obtained by the development of techniques which removed essentially all of the organic matrix constituents of the bone samples prior to TEM without altering the bone crystals, thus permitting their examination by high resolution TEM and electron diffraction as well as 3D stereoscopic TEM, which firmly established that the bone apatite crystals were indeed nanosized, very thin, long platelets (nanocrystals) and not rods [16-18,]. This was further confirmed by atomic force microscopy [19, 20]. Further TEM compositional studies of the bone crystals from a wide variety of species as a function of “crystal age”(vide infra) revealed that the initial Ca–P solid phase deposited consisted of irregular, roughly round electron dense particles, some as small as ~10 Å . These data focused attention on the important physiological implications of the enormous surface area of the bone nanocrystals, especially during the early stages of new bone formation and calcification, and its role in ion homeostasis of the extracellular fluids. These data demonstrate another important structural and physical chemical characteristic of the bone crystals as a function of “crystal age” (vide infra), further complicating the problem of quantitatively describing the bone crystals in a particular sample of bone substance.
Great strides have been made in utilizing in vitro calcification experiments using purified native collagen fibrils and metastable solutions of Ca and P to gain information on the composition and structure of the initial Ca–P solid phase formed during heterogeneous nucleation. These have demonstrated that purified natively 3D-packed collagen molecules in collagen fibrils with an axial period of approximately 670Å are capable of heterogeneously nucleating apatite crystals in vitro [22-24].
It was hoped that this information on the potential chemistry and structure of the initial nucleation sites within the collagen fibril responsible for initiating crystallization might be obtained from the in vitro nucleation experiments.
The chemical and X-ray diffraction data identifying the nanocrystals nucleated in vitro by collagen fibrils, including the specific location of the nanocrystals in an orderly repeating pattern within the collagen fibrils [25-27], was later confirmed by wide and low angle X-ray diffraction studies . Much more quantitative in vitro nucleation experiments were carried out by Katz , who quantitatively measured the kinetics of in vitro calcification of native collagen fibrils and established the size of the apatite nuclei found in his in vitro nucleation studies (~ 20 ions under the specific conditions of his experiments). Various schema and experiments were devised from these in vitro studies of nucleation of the likely composition of the nuclei (first solid phase Ca–P formed during in vitro calcification) and the potential role of the phosphoryl groups and glutamic acid side chains of a particular noncollagenous matrix protein in bone (bone sialoprotein-BSP) [27, 29, 30, 32].
Very importantly, solid state 31P NMR analysis of the youngest sample of native bone available at that time identified for the first time, the presence of a seryl phosphate-calcium bond, presumably directly linking a prominent protein in the organic matrix of bone (BSP) with the mineral phase .
A number of other experiments were also devised and carried out to surmount the problems of how to isolate the very youngest bone mineral crystals in specimens of bone without introducing any contact of the specimens with water, which not only very rapidly dissolves the crystals, but also introduces physical and chemical compositional changes in the crystals which remain in the bone sample.
While the chemical composition and crystal structure, particle size, degree of crystallinity, etc., of the individual bone crystals in a specific sample of bone in a particular location of a particular bone is clearly a significant function of the age of the animal, a recently introduced new concept has pointed out that the most significant “time function” as far as it affects all of the compositional, structural, and interaction properties and characteristics of bone apatite crystals is not the age of the animal, but the age of the individual crystals (“crystal age”) viz., the elapsed time between the initial deposition of the crystals and their removal (resorbtion) from the tissue . The very significant chemical and structural changes in the solid mineral phase are primarily a function of “crystal age” and clearly directly related to the very rapid local turnover of bone substance and its crystals, in many instances over widely separated local regions of a particular bone.
Other difficulties in accurately determining the chemical composition and structure of the bone mineral at any one point of time in a specific location of a particular bone during it’s initial formation and subsequent maturation (as a function of age of the animal and especially “crystal age”) are also due to the very complex internal geometric disposition of the bone substance containing the crystals at all levels of the anatomical hierarchy, observable from light microscopy to high resolution transmission electron microscopy and atomic force microscopy. For example, it is clear from light microscopy and contact X-ray microscopy that there are great differences in the mineral content of individual, immediately adjacent osteons, for example, and in immediately adjacent individual lamellae of a single osteon. Indeed, transmission electron microscopy and electron diffractometry, and other high resolution techniques, have highlighted the compositional, structural and mineral content differences of the bone mineral substance due to the marked heterogeneity within various volumes of the bone substance at all anatomic levels observable from light microscopy to high resolution TEM, atomic force microscopy, and even higher levels of resolution.
It was also clear from the early TEM experiments that the exposure of samples of bone to water introduced significant dissolution of the bone nanocrystals as well as phase and compositional changes in many of the remaining crystals in the samples of bone. It was therefore clear that the preparation of samples of bone for microscopy at all levels of resolution, and their chemical and physical properties would all require the preparation and analyses of samples of bone without exposure to water.
The earliest such experiments were carried out on the very outermost layers of bone of the midshaft of the long bones of embryonic chick (and other species) using only non-aqueous solvents. The bone samples were powdered and density separated in non-aqueous solvents. The very lowest density fractions were first evaluated by wide angle X-ray diffraction, chemical composition and 31P NMR, revealing the presence of only very poorly crystallized apatite. There was no evidence of an initially “amorphous Ca–P solid phase” which had first been postulated by Posner et al, which, with time (higher density samples), underwent either a solid state phase change to very poorly crystallized apatite crystals or was first solubilized and then recrystallized as poorly crystallized apatite crystals [32- 36].
Further compositional and structural studies to ascertain the structure of the initially nucleated bone crystals were carried out on the intramuscular bones of various species of fish. The use of these bones containing highly oriented collagen fibrils provide many important advantages in identifying the location and nature of the first Ca–P solid phase formed very accurately at high resolution. Using high definition synchrotron-generated X-ray diffraction and FTIR, and the simultaneous measurement of the concentrations of calcium and inorganic phosphorus at the same sites  clearly established that the earliest Ca–P solid phase deposited during the calcification of the fish bones is a very poorly crystalline solid phase of apatite, similar to the results obtained from the earliest detectable Ca–P mineral phase found in chick and other species. Note that the earliest solid Ca–P mineral phase in the fish bones was obtained from freshly caught fish immediately frozen while alive in liquid nitrogen and examined using a temperature-cooled stage.