2 EXPERIMENTAL

A flowchart of the three experimental protocols is presented in the appendix to this paper.

2.1 GYPSUM

In the first place, the behavior of pure gypsum powder (calcium sulfate dihydrate 98%, Aldrich) was studied when dried at ~0% RH at room temperature, and at 40�C, 60�C, and 105�C. The ~0% RH was obtained using a laboratory desiccator containing Drierite. This product, based on specially processed calcium sulfate, is capable of taking up 14% of its weight in water. In practice, it is estimated that the actual relative humidity is below 5% for standard glass laboratory desiccators.

The drying temperatures were set on a commercial laboratory drying oven. The samples, weighing ~3 g, were left at these conditions until the relative change in weight was less than or equal to 0.003%. Two of these samples, those dried at 0% RH and at 105�C, were then conditioned for 20 days at 75% RH to determine their moisture absorption.

2.2 GYPSUM-CONTAINING MIXTURES

At the same time, mixtures of gypsum with powdered calcite (calcium carbonate, certified ACS, Fisher Scientific) or silica (floated, Fisher Scientific, particle size approximately 240 mesh) were dried at ~0% RH, 40�C, 60�C, and 105�C to determine the effect a second component would have upon the dehydration of gypsum. For this purpose, 5% by weight of gypsum was mixed with each of the powdered materials. Pure calcite and pure silica samples were also dried as controls. All samples, weighing approximately 3 g total, were left to dry until the relative change in weight was less than or equal to 0.003%.

Of the above mixtures, those dried at 60�C, as well as the corresponding sample of pure gypsum, were redried at 105�C for 11 days after having been kept in a closed container in laboratory conditions for about a month. After drying they were re-exposed to ambient conditions in the laboratory, and their weight increase was monitored. It was found that after four days the weight of the samples changed (on the order of 0.05%) as a function of the fluctuations of relative humidity in the laboratory, and the weight monitoring was discontinued.

2.3 GYPSUM-AND SALT-CONTAINING MIXTURES

A third set of samples was prepared by mixing key components of a real mortar. Approximately equal amounts of silica and calcite—about 1.5 g of each—for the control were mixed and labeled as C. The set was completed by adding 5% of gypsum (~0.15 g) to some samples (C + G); 2% of sodium chloride to others (C + H for halite); both 5% gypsum and 2% sodium chloride (~0.06 g) to a third group (C + G + H); and both 5% gypsum and 1% sodium sulfate (~0.03 g) to the fourth group (C + G + T for thenardite). All weighings were carried out to 0.01 mg. The actual number of millimoles contained in the samples were calculated from these weighings.

These samples were first conditioned at 75% RH for two weeks and then dried at 40�C or at 105�C, or in a desiccator at approximately 0% RH, until the relative change in weight was less than or equal to 0.003%.