Computer-aided molecular design using the signature molecular descriptor: An application to design novel chemical admixtures for concrete.

摘要

The development of new chemical admixtures for concrete is normally an experimental endeavor in that the molecular scaffolds of existing admixtures are modified and tested. This approach is time consuming, incremental and typically expensive. Alternatively, a computer-aided molecular design (CAMD) approach is proposed that uses the Signature molecular descriptor. CAMD is the application of computer-implemented algorithms that are utilized to design molecules with optimally predicted properties such that they can be tested and evaluated for efficacy. After creating and refining a quantitative structure-property relationship (QSPR) model for a property of interest, a structure enumeration algorithm is employed to generate structures outside of the original training set that have optimally predicted properties. In this research, the novel CAMD approach is shown in application for designing two types of chemical admixtures are of interest: (1) water-reducing admixtures (WRAs) and (2) shrinkage-reducing admixtures (SRAs).;WRAs have been developed to reduce the required water content of concrete while maintaining a flowable cementitious material, described by its rheological properties. A parallel-plate rheometer was used to measure the effect of eighteen low molecular weight organic chemicals on rheology by directly obtaining yield stress and viscosity data. Workability, measured by a mini-slump cone test, is reported for comparison. Chemical structures were examined and analyzed for their effect on flow properties and set time. Sugar acids proved to be the most effective in reducing yield stress, viscosity, and increasing workability. Sugar alcohols with more hydroxyl groups and longer carbon backbone chains decrease plastic viscosity. Results suggest that mini-slump cone test is unreliable for measuring workability, and that more complex rheology-based protocols have to be employed to better translate cement paste workability to application targets. The CAMD methodology using the Signature molecular descriptor was used in order to identify new chemicals with high workability, low yield stress and viscosity, and short set times that might serve as a new generation of WRAs. Beginning with a training set of the eighteen tested compounds, the technique combines the atomic Signatures in novel ways to identify non-intuitive compounds with optimal predicted properties.;For SRAs, the property of interest is the surface tension of compounds in aqueous solutions as this property is related to shrinkage in concrete. By evaluating the initial surface tension reduction of amines and glycol ethers in solution with water, a number of structure-property conjectures associated with the effect of these compounds were developed. Accordingly, 14 compounds were identified and utilized as a training set for the CAMD algorithm and new compounds were identified. The surface tension reduction for the newly identified compounds was experimentally evaluated. Furthermore, the effect of five newly identified compounds on autogenous and drying shrinkage as well as the compressive strength of the cement paste were tested and compared to two commercial SRAs. Results for both types of shrinkage indicate that the designed compounds perform similar to commercial admixtures, yet have different chemical functionalities and better strength development.