Carbon dioxide has a number of advantages:
1) it is inert, and does not corrode (Oxygen compressed is very corrosive, and can not be used around greased fittings, or plastics).
2) it is very cheap, it is the industrial byproduct of many processes.
3) When compressed, carbon dioxide becomes liquid. Air, and nitrogen, do not. The carbon dioxide pressure above the liquid is determined only by temperature. At room temp, CO2 stabilizes at about 800PSI, so the bottles for CO2 liquified are about 1200PSI. The test rating on many bottles is 2000 pSI for liferafts.
An equivalent bottle of Nitrogen would be about 6000PSI, and the bottle would have to be constructed much heavier.
4) CO2 bottles such as this are safe to puncture, and in fact are not valved. The release mechnism is a small puncture device, the escaping gas actuates a piston to operate a full open puncture. To do the same on a N2 bottle at 6000 PSI, a puncture thru the thicker container would be dangerous. There is a lot of difference in 6000PSI and 800 PSI.
A note on the compression ratios. Because you are dealing with storing a liquid, the talk about "compression"ratio is meaninless. What is meant of course, how much gas will the cannister provide, and "compression"ratio is a way of describing how much gas (greater than the cannister) is made. Well, that depends on the pressure in the liferaft.
For many emergency liferafts, they inflate at about 2Psi guage, so conversion of the liquid to gas is about 1000:1 ratio. For more rigid lifeboats, the pressure is about 10 PSI, so the ratio is about 200:1. YOur number of 600:1 is probably typical of many escape boats at about 3-4 PSI. It does not take much air to float a tube. OF more concern, is leaks and leak control, as the cylinders are one time use.
Now your statement the compression ratio of air is much less: Air is not liquified. The "compression"ratio will depend on its pressure. But for safe practice (storage and opening), one could depend on about storing a volume of air as a gas about 1/100 the amount as storage as a liquid. A cup of liquid will make 160 liters ( a "compresson" factor of about 620:1) whereas a gas stored at the same pressure (800PSI) would be about 400:1. These ratios depend on pressure of the outlet gas, of course.
All in all, CO2 is the preferred gas: it liquifies, then stores at a reasonalble pressure, is very cheap, and is non corrosive.
Thank you, fantastic reply once again.
Do you happen to know the technical name for the device that spins round to pierce the Co2 bottle when the painter is pulled?