Molecular orbital study of apatite (Ca5(PO4)3OH) nucleation at silica bioceramic surfaces

Silica bioceramics are used as prosthetic bone and dental implants because they promote apatite formation at their surfaces when immersed in simulated body fluids of composition similar to human blood plasma. Apatite formation occurs in stages, but the reaction pathway remains unresolved. We have used molecular orbital calculations to model the interactions of Ca2+, H2PO4-, HPO42-, and H2O with bioceramic surface sites represented by Si7O12H10, Si4O8H8, and Si3O6H6 to determine the reaction sequence for apatite nucleation. Predicted reaction energies, vibrational frequencies, and Si-29, P-31, and Ca-43 NMR shifts were used as probes of the stable intermediates during reaction progress. Energies and vibrational frequencies were calculated using effective core potentials and valence double-xi basis sets. Most NMR shifts were calculated using the 3-21G* basis, and using the 6-31G* basis for selected clusters. With increasing pH above the point of zero charge of silica, (pzc approximate to 3) the stability of the calcium surface complexes on the Si 3-ring ([Si3O6H6]) is predicted to be outer-sphere < inner-sphere < surface-hydroxide. Comparison of calculated Si-29 and P-31 NMR shielding trends of Si 3- and 4-rings (Si3O6H6 and Si4O8H8, respectively) with experimental trends for bioceramic reacting with SBF suggest that (i) the partially deprotonated 3-ring is the active surface site, because it promotes calcium ion dehydration as calcium adsorbs, (ii) formation of [SiO-Ca-OPO3H] bonds is energetically preferred over direct Si-O-P bonds, and (iii) an acidic precursor with bidentate Ca > OPO3H bonds nucleates rapidly (minutes to 1 h) at the bioceramic surface. The precursor oligomer is modeled by the complex [Si3O6H5CaHPO4(H2O)(3)](-) The actual precursor is proposed to contain two or three calcium hydrogen phosphate units with a stoichiometry such as [Si3O6H5CaHPO4(H2O)(n)](-), where n = 2 or 3. The complex [Si3O6HSCaHPO4(H2O)(3)](-), young bone, and bioceramic reacting with simulated body fluid share unique infrared/Raman bands at 631 and 1125-1145 cm(-1), which are distinct from the bands observed in crystalline apatite and mature bone. Predicted Ca-43 NMR shifts provide avenues for future experiments.