Thermodynamic modelling can reliably predict hydrated cement phase assemblages and chemical compositions, including their interactions with prevailing service environments, provided an accurate and complete thermodynamic database is used. Here, we summarise the Cemdata18 database, which has been developed specifically for hydrated Portland, calcium aluminate, calcium sulfoaluminate and blended cements, as well as for alkali-activated materials. It is available in GEMS and PHREEQC computer program formats, and includes thermodynamic properties determined from various experimental data published in recent years. Cemdata18 contains thermodynamic data for common cement hydrates such as C-S-H, AFm and AFt phases, hydrogarnet, hydrotalcite, zeolites, and M-S-H that are val-Cemdata18 includes a comprehensive selection of cement hydrates commonly encountered in Portland cement (PC) systems in the temperature range of to 100°C, including calcium silicate hydrate (C-S-H), magnesium silicate hydrate (M-S-H), hydrogarnet, hydrotalcite-like phases, some zeolites, AFm and AFt phases, and various solid solutions used to describe the solubility of these phases. Solubility constants have generally been calculated based on critical reviews of all available experimental data and from additional experiments made either to obtain missing data or to verify existing data. Additional solubility data were measured and compiled using temperatures ranging from 0 to 100°C in many instances, as documented in [9, 12, 27, 28]. Numerous solid solutions among AFm and AFt phases, siliceous hydrogarnets, hydrotalcite-like phases, C-S-H, and M-S-H have been observed and are included in Cemdata18. Several C-S-H solid solution models, as well as two models for hydroxide-hydrotalcite are available in Cemdata18. The CSHQ model from [11] and the OH-hydrotalcite end member with Mg/Al = 2 are well adapted for PC. Although the CSHQ model is able to describe the entire range of Ca/Si ratios encountered, it is best used for high Ca/Si C-S-H, as it still lacks the ability to predict aluminium uptake, which is of less importance for Portland cements than for blended cements. For alkali activated binders, the calcium (alkali) aluminosilicate hydrate (C-(N-)A-S-H) gel model, with lower calcium but higher aluminium and alkali content than in the C-S-H type phase which exists in hydrated PC, and a Mg-Al layered double hydroxide with variable Mg/Al ratio, are available. This paper summarises Cemdata18, which includes the most important additions to the Cemdata07 and Cemdata14 databases in recent years. It also discusses the relevance and implications of these additions, and compares Cemdata07 and Cemdata18, accounting for their main differences. Summaries
