Is the R index accurate to assess the preferred orientation of portlandite in cement pastes?

Abstract

The evaluation of portlandite orientation through the R index obtained from X-ray diffraction (XRD) pattern has become popular in recent years, and many works related this index with the mechanical properties of the composite. Such index is defined by 1.35 times the ratio between the intensities of the 001 and 101 reflections XRD pattern. However, most of these works neither present specific strategies to minimize preferred orientation due to sample preparation nor account for experimental contribution in the background. This paper pointed some factors that affect XRD data collection and interpretation, and discussed their effects on R index determination. Powder XRD was conducted using different sample holders and collection times, in addition to in-situ measurements for fresh paste. The results showed that sample preparation itself results in important intrinsic variability, which can be mitigated in Rietveld refinement but not in R index analysis. In addition, the sample type played an important role in portlandite assessment: paste slice kept the random distribution of portlandite crystals but damaged ettringite, while powder press induces preferred orientation perpendicularly to the pressing direction. Powder backloading only partially reduced the preferred orientation issue. In-situ sample can keep the original matrix arrangement, but such analysis is difficult for long periods. Furthermore, a corrected R index is proposed, in order to consider the contribution of the background. Overall, the use of the R index for portlandite orientation assessment must be carefully addressed to avoid misinterpretations.