A NEW APPROACH TO THE DESALINATION PROCESS OF ARCHAEOLOGICAL POTTERIES
Keywords:
Desalination, optimization, rinse water, agitation, response surface methodology (RSM), archaeological potteriesAbstract
As salt crystallization is one of the most serious damages in historical porous materials (e.g. potteries, bricks,
etc.), desalination is highly important process in conservation. Since this process is irreversible, and object
has direct vicinity with water, preventing the misrepresentation of valuable archaeological and archaeomet
rical data of historic object will be substantial. On the other hand, a considerable amount of water is used in
this process because of a huge number of excavated potteries in archaeological sites, therefore, controlling
the amount of rinse water consumption is extremely desirable. As this process is a multivariable system, a
new approach based on a mathematical design of experiment, response surface methodology (RSM), was
used as a n aid in determining the significance of the various parameters and optimization of this process.
The input (independent) variables in the experimental design were immersing time, agitation rate, firing
temperature (as void fraction indicator), type and concentration of surfactant. For each variable, five levels
were selected in a batchwise pilot tests. Electrical conductivity (EC) and Ca2+ concentration of rinse solutions
were measured as responses of experiments. It was found that the most effective parameters in the immers
ing desalination process are: firing temperature of specimen, immersing time, and agitation rate. The exper
imental results revealed that desalination process in optimized conditions results in speed up the operation
time up to less than 8% and saving up to 50% water consumption compared to conventional method at the
same desalination achievement. The accuracy of the modeling was validated with triplicate experiments. It
was found that the average Ca2+ concentration and EC value in rinse solution fit approximately 90% with
RSM predicted data.