1. Introduction
Historical mortars have two main components: the binder,
which is usually lime or hydraulic lime, and the aggregate [1],
which in most cases consists of a locally available sand [2].
Knowledge of the grain-size distribution of the aggregate in
these mortars, which plays an important role for restoration
applications, is invaluable [3]. Combined with the volume
fraction of aggregate used, this is necessary for designing
compatible mortars. Furthermore, the sieving curve of a sand is
a fingerprint of the geological formation from which the sand is
taken, so that it is an important tool for tracing the origin of the
raw materials [4]. A generally applied method of establishing
the grain-size distribution of the sand is to attack chemically the
mortar sample with an appropriate acid [5]. The binder is
dissolved and the insoluble residue is used to construct the
sieving curve. However, part of the aggregate may not be
resistant to the acid attack and readily dissolve [6,7]. The large
amount of sample required [8] is an additional restriction of this method. A different approach is required to obtain reasonable
results when only a limited amount of sample is present. Image
analysis of thin sections can overcome the limitations of the
acid-attack method. Traditional optical or scanning electron
microscope (SEM) images can be processed by image analysis
methods to allow measurement of the sectioned grains. Wellestablished
stereological calculations are used to convert these
2D data to the required 3D values [9,10]