You raise some good points that I will attempt to address.
If the DX ground loops are installed correctly, they are not in contact with dirt! All of my loop installations [pit and trench systems] are encased in screenings [crushed granite]. Vertical and diagonal loops are grouted with a thermal conductive grout. In the cooling mode, heat rejected into the Earth can cause the dirt to dry and separate from the loop leaving a void resulting in poor heat transfer. The screenings will not dry and separate from the loops resulting in better heat transfer. The additional mass also aids in the transfer process.
You are correct in stating that plastic pipe has a poor energy transfer rate. This is why water source systems require 4-5 times more loop length than DX. Longer loops require more loop field and excavating resulting in higher install costs. As far as localized freezing and heating of the Earth, it has the same effect on water source systems as it does on DX with the exception of open loop systems. But with open-loop systems, you have higher pumping costs, the water had to be filtered and treated in some cases, plus you have to dump it in an unobjectionable area.
As for loop life, I know that most plastic loops carry a 50-year warranty. However, I can not tell you how many I know of that are leaking and require continuo maintenance. Some even have water-feeding attachments on them to keep the system full of water. This leads to the transfer solution being diluted causing premature corrosion of the heat exchanger. As for the copper loops, copper is a noble metal that comes from the Earth and is the oldest metal used by man. A pH test must be performed on the soil to determine its acidity level before the system can be installed. A pH between 5.5-11.0 requires no protection, as this will not harm the copper. If it is outside this range, a cathodic protection system will protect the copper from corrosion. I have had to abandon many water loops due to leakage, but I know of only one DX system that had a loop failure. With better than a 30 year track record, I believe that speaks for itself.
As for efficiencies, I believe this to be correct. The compressors in these systems are not under anymore load than a typical system [air or water source]. The loops have been engineered with this in mind. If there were such a load on the compressor, the efficiency would drop off.
What I have a problem with is how water source manufacturers are deceiving consumers by advertising super high EER numbers. As you are aware, this is achieved by 2 speed systems and the rating is for low speed. But most consumers do not know this. When you look at the EER numbers at full capacity, they are at or below that of DX systems. Second, water source systems typically have a major drop off in heating capacity. For example, Waterfurnace's "Premier E Series Dual Capacity" 5 Ton heat pump [E060] has a cooling capacity of 60,000 Btuh and only 41,800 Btuh heating capacity. The EER is 17.0 and the COP is 3.9. That equates to 3.5 Tons of heating capacity from a 5-Ton system. American Geothermal's 5-Ton system has a heating capacity of 61,000 Btuh with a COP of 3.7. The cooling capacity is 64,000 Btuh with an EER of 16.6. This gives the Amgeo DX system 32% more heating capacity than this comparable unit. Check it out for yourself, this is directly from ARI's website.
As for DX falling off, it is all I install in terms of geothermal and I install several each year.