Gypsum is chemically classed as slightly soluble in water. However it's robustness can be improved in a number of ways. These have to do partly with changing the chrystal morphology and partly to do with removal of excessive void spacings between the chrystals using additives. There are other ways to improve it but these are trade secrets.
Where the chrystal formation makes large loosely packed chrystals e.g. plasterboard then the material remains relatively soft and prone to cleaving when large quantities of water are present. However gypsum floor screeds are engineered to use tiny tighly packed chrystals which are bedded around a matrix of fine aggregate e.g. sand or ground glass. This imparts very significantly greater strength and less propensity towards cleaving due to the addition of water. If a gypsum screed is left saturated for a very lengthy period i.e. quite a number of years then some small amount of swelling may be observed at the molecular level but for all practical purposes the only thing that happens to a gypsum screed when it gets wet is that it gets wet. As long as it is re dried this does not cause an issue. It is very difficult to get a floor covering to stick to a wet screed. Again there are things which can be done to improve the robustness of the bond between the adhesive and the screed e.g. use a gypsum based adhesive. These screeds began life in a big way in the late 80's early 90's but this was before the advent of gyspum based adhesives so there were the obvious issues. To put this into context for you I have been on and off involved with a recurring issue at a major hospital complex in Norflok where Gyvlon screed was placed over 17 years ago. The area was smoothed using a gypsum based smoothing compound and then a vynil floor on top. They have intermittent problems with the floor flooding and have had at least 4 occasions where large scale immersion has occurred. On each occasion they have lost the floor coverings due to the emulsification of the flooring adhesive. However on each occasion the smoothing compound has remained firmly adhered to the screed and the screed has remained unaffected apart from being wet.
Where the chrystal formation makes large loosely packed chrystals e.g. plasterboard then the material remains relatively soft and prone to cleaving when large quantities of water are present. However gypsum floor screeds are engineered to use tiny tighly packed chrystals which are bedded around a matrix of fine aggregate e.g. sand or ground glass. This imparts very significantly greater strength and less propensity towards cleaving due to the addition of water. If a gypsum screed is left saturated for a very lengthy period i.e. quite a number of years then some small amount of swelling may be observed at the molecular level but for all practical purposes the only thing that happens to a gypsum screed when it gets wet is that it gets wet. As long as it is re dried this does not cause an issue. It is very difficult to get a floor covering to stick to a wet screed. Again there are things which can be done to improve the robustness of the bond between the adhesive and the screed e.g. use a gypsum based adhesive. These screeds began life in a big way in the late 80's early 90's but this was before the advent of gyspum based adhesives so there were the obvious issues. To put this into context for you I have been on and off involved with a recurring issue at a major hospital complex in Norflok where Gyvlon screed was placed over 17 years ago. The area was smoothed using a gypsum based smoothing compound and then a vynil floor on top. They have intermittent problems with the floor flooding and have had at least 4 occasions where large scale immersion has occurred. On each occasion they have lost the floor coverings due to the emulsification of the flooring adhesive. However on each occasion the smoothing compound has remained firmly adhered to the screed and the screed has remained unaffected apart from being wet.