A new model for the formation of calcium silicate hydrate (C-S-H)

Abstract

Models used to describe the hydration of beta-dicalcium (C2S) and tricalcium silicate (C3S) pastes are generally based on the notion that the dimeric calcium silicate hydrate (C-S-H) that begins to form at the end of the induction period is not a stable phase. It is implied that its formation simply marks the beginning of a rather lengthy equilibration process. This statement is based on the fact that the arrangement of Ca2+, OH− and (Si2O7)6- ions in the C-S-H at this time has no long range order (C-S-H is X-ray amorphous) and the fact that the dimers present in the C-S-H will seemingly polymerize and form dreierketten of the 3n-1 type with time. It has been suggested that the localized layers of C-S-H that contain these dreierketten are structurally related to to-bermorite that has been modified in various ways, including omission of bridging tetrahedra and partial replacement of silicate ions by OH− groups.As an alternative, it is suggested that the dimeric C-S-H that forms soon after setting and hardening may actually be a metastable phase in its own right, a rigid gel precursor phase whose stability is related to its calcium content. Should calcium concentrations fall below the level needed to stabilize the phase (a phenomenon that can be initiated by rising temperatures and/or portlandite precipitation), the dimeric C-S-H will undergo a phase change forming nano-sized fragments of tobermorite/jennite-like C-S-H. The proposed model differs from current models in that it proposes a relatively rapid equilibration followed by a much slower diffusion-controlled phase change process. Although both models predict the same outcome, i.e. a mature paste sample will normally contain a mixture of dimer and dreierketten, the new model gives a raison d'être for the presence of consistently large amounts of dimer throughout the entire hydration process.