Alright, I have some results from my experimentation with freezing conditions. Cells were grown up in a starter until they reached stationary phase. They were then kept in the refrigerator for a few days. I aliquoted 30 billion cells into 9 different 50ml conical tubes, centrifuged them, and removed the supernatant. The cells themselves took up a volume of 15-20ml, so I added 20ml of fresh wort with glycerol added at a concentration of 15%, 30%, or 50%. Since 20ml of wort+glycerol was diluted 2-fold by the volume of the cells, the final concentrations of glycerol were more like 8%, 15%, and 25%. Cells were mixed thoroughly and then frozen at either -20C, -80C, or in liquid nitrogen. For cells frozen at -20C and -80C, the tubes were placed in a styrofoam cooler filled with room temp 70% isopropanol, then put directly into the freezer. The idea was to cool the cells very slowly and protect them being frozen too quickly, which can cause formation of ice crystals that can physically damage and kill the cells. It also prevents osmotic shock during the freezing that can kill the cells. The cells frozen in liquid nitrogen were thrown directly in and were frozen in ~10 seconds. I anticipated that this would result in poor viability. All cells were thawed as quickly as possible in a 37C water bath. Here are the results:
Starting viability before freezing: 78%
-20C 15% glycerol: 80% viability
-20C 30% glycerol: 50% viability
-20C 50% glycerol: 44% viability
-80C 15% glycerol: 70% viability
-80C 30% glycerol: 73% viability
-80C 50% glycerol: 50% viability
Liquid N 15% glycerol: 0% viability
Liquid N 30% glycerol: 0% viability
Liquid N 50% glycerol: 9% viability
Viability was measured 3 different was: 1) manual counting on hemocytometer using trypan blue to identify dead cells, 2) automated counting using Vi-Cell machine and trypan blue, 3) CFU counts on YPD plates. Both the Vi-Cell and the CFU gave pretty similar numbers, but I trust the CFU more. Thats because trypan blue was not a great way to identify dead cells. Dead cells dont stain dark blue with trypan, but rather just look less bright, which is very hard to discriminate between live cells. In fact, by eye I counted ~75% viability in samples that were 0% viable by CFU. The Vi-Cell machine was better at discriminating than my eyes, but since it still uses trypan blue, I doubt its 100% accurate.
Conclusions: Freezing at -20C with 15% glycerol gave the best viability. Increasing concentrations of glycerol was actually detrimental at -20. Good viability was also achieved at -80C and the effect of glycerol concentration was less apparent. Flash freezing in liquid nitrogen is bad. CFU is the best way to measure viability.
Future experiments will be needed to measure viability after long-term storage at -20C vs -80C. Future experiments will also be needed to measure viability after freezing in a home freezer. However, I think freezing slowly is essential, and freezing the cells in a styrofoam cooler filled with 70% isopropanol is something any homebrewer can do. Storing the cells like this should also help protect them from the freeze/thaw of the defrost cycle of home freezers.