I had two failed experiments with zip lock baggies. They may have been the store brand type, and not as good as a nationally advertized brand.
The first leaked at the seam within ten minutes. It may have started leaking right at the outset and I just missed it, because there was no sign of melting that would account for it to start to split, nor was the bag hot to touch.
The second bag, I hung with a pair of clothes-pin type clamps, to the edge of the cover, of my grill, that was propped open, but partially closed. The bag was about 3/4th full of water, and not supported in any way. An hour into the experiment, the water was only 140F and the back, top, of the bag melted and tore away so that all that was holding the pocket of water was the front half of the bag.
Problems with the experiment's design:
1. the heat was not evenly distributed. The bag of water was heated from below, but the front of the bag was hanging adjacent to the outside ambient air (about 65F) while the back of the bag was in the more enclosed space (like an oven). I don't know??? 300F ??
2. probably cheap plastic bags... better quality should be attempted.
Observation:
1. Just like in the paper cup experiment, (the cup caught on fire and burned down to the water-line), the plastic bag maintained it's integrity where it was in contact with the water it was holding, however, the plastic, not in contact with the water, if not closely monitored, can loose it's integrity.
2. water is heavy and the plastic bag is a little challenged in supporting the water's weight by the lip of the bag with clips.
More experimentation is necessary.
Better quality bags
more even distribution of heat
external support of bags (damp/wet cloth?, sitting in the coals?, lying on a rock, next to the fire?)