Formation of Geopolymers Using Sodium Silicate Solution and Aluminum Orthophosphate

This paper reports the formation and structure of fast setting geopolymers activated by using three sodium silicate solutions with different modules (1.6, 2.0 and 2.4) and a berlinite-type aluminum orthophosphate. By varying the concentration of the aluminum orthophosphate, different Si/Al-ratios were established (6, 3 and 2). Reaction kinetics of binders were determined by isothermal calorimetric measurements at 20 °C. X-ray diffraction analysis as well as nuclear magnetic resonance (NMR) measurements were performed on binders to determine differences in structure by varying the alkalinity of the sodium silicate solutions and the Si/Al-ratio. The calorimetric results indicated that the higher the alkalinity of the sodium silicate solution, the higher the solubility and degree of conversion of the aluminum orthophosphate. The results of X-ray diffraction and Rietveldt analysis, as well as the NMR measurements, confirmed the assumption of the calorimetric experiments that first the aluminum orthophosphate was dissolved and then a polycondensation to an amorphous aluminosilicate network occurred. The different amounts of amorphous phases formed as a function of the alkalinity of the sodium silicate solution, indicate that tetrahydroxoaluminate species were formed during the dissolution of the aluminum orthophosphate, which reduce the pH value. This led to no further dissolution of the aluminum orthophosphate, which remained unreacted.

Potentiometric measurement of H+ concentrations for aluminum orthophosphate titration curves

A potentiometric titration method of measuring accurate H+ concentrations is demonstrated for aluminum orthophosphate solutions. A calibration procedure has been used which permitted the glass electrode to be used as a concentration probe instead of as a quasi-activity probe. The random error "background noise" has been reduced by the use of Fortran computing for chemical calculations, data handling and sorting, and averaging by least squares poly nomial fits. The distribution of experimental errors along the sample titration curve is indicated by estimates for its various regions.