not to argue fine points. but the interstitial soldered bond gap between the solder surfaces and the solder, as well as the the pore size/ surface roughness variation in the refrigerant conveyance lines, solder and motor case seals is huge compared to freon or ammonia molecule sizes.
I have sectioned the sealed joints, I have observed migration of material through the joints and yes I have access to an electron microscope and metallurgical lab equipment.
For practical purposes they are sealed, on an absolute scale......no they are not.
25 years is normally the end of practical life with the workmanship practices used on most commercial small refrigeration systems. Internal corrosion is an issue, motor wear is an issue.
Been in the lab, nuclear and insurance industries over 35 years, I have had opportunity to work on and examine dozens of refrigerant systems equipment failures.
I stand by my findings for boats. 6 months or more of lack of use every year on a recreational marine refrigeration system is not going to prolong the life of the commercial grade systems we use on boats.
I am sure there are some boats out there that have had fridges that are turned off 6 or more months a year for many years. It would be an interesting test. However, when working with ThermoKing units, there are statistics availble on a representative sample of tested units. These tests showed extended no operation periods is harder on truck/trailer sized units. These were averaged over mulitple model years, multiple locations. Note. these units are also exposed to high temps during the summer when out of use, and that plays a part. A boat is a fairly benign location temperature wise, and the lower the temp during off periods, the lower the internal gas pressure and leak rate. YMMV