An article on pharmaceutical engineering says: Direct steam contact with the surface of the object to be sterilized is required for the steam to transfer its stored energy to the object. Without direct steam contact to all surfaces, the item will not be sterilized. The amount of energy stored in steam is much higher than dry air or water at the same temperature. From the saturated steam table mentioned above, one can see that it takes 419 kJ/kg (180 Btu/lb) to heat water from 0°C to 100°C (32°F to 212°F). This is the enthalpy of water (hl). It takes an additional 2,257 kJ/kg (970 Btu/lb) to create steam at atmospheric pressure (100°C or 212°F). This additional energy stored in the steam is the enthalpy of vaporization (he), and is the key to steam sterilization. In order for the steam to transfer its stored energy, it must condense on the surface of the object being sterilized.

The energy that article talks about is thermal energy. If you pour dry powder into a vial, seal the stopper and heat the vial, the heat is conducted from the glass into the powder quite slowly because heat conduction in unmoving dry powder (with air between particles) is slow. After half hour heating, part of the powder remains relatively cold, unsterilized. The sterilization time begins when everything is heated enough - it's why the talk about "energy transfer".

Thermal energy can be transferred by condensation of steam (as explained in the quote above) or by conduction, radiation or convection. Conduction is indeed quite slow. Convection of dry air transfers thermal energy not enough fast. However, the author of that article talks about a "steam autoclave" implying that the water boiling in one part of the autoclave is not in direct contact with the items intended to be sterilized. If a vial is submerged in boiling water, convection in boiling water transfers thermal energy into the vial very, very much faster. If the vial contains a solution instead of a dry powder, the solution inside the vial also convects, the thermal energy is quickly transferred to every part of the solution, so if you submerge the vial in water boiling in a 15 PSI (or 100 kPa) pressure cooker then entire solution in the vial is quickly heated to 121°C.

I was asked "how quickly?". I did an experiment with viscous Castor oil (slow convection) and a digital thermometer with long thin sharp probe. I mixed 10 ml of solution with Castor oil and 2% benzyl alcohol in a vial, sealed and crimped the vial, pierced the rubber stopper with the probe (so that the thermo-sensitive end of the probe was at the center of the solution in the vial) and plunged most, but not all height of the vial into water boiling in a pot (so that the probe didn't touch the water). The numbers the thermometer indicated rose from 26°C to 95°C during 4.5 min. Then I cooled the probe in cold tap water to 16°C and plunged the probe directly into boiling water. The numbers rose to 95°C during 0.5 min (because the probe with metal coating has thermal inertia).

The XFS-260 autoclave I use is similar to a pressure cooker, but has two bowls. The inner bowl is made of thin stainless steel. I pour water into the outer bowl up to slightly above the bottom of the inner bowl, and pour water into the inner bowl too. So, vials are submerged in water boiling at 129°C (under pressure 24 PSI). Thermal energy is quickly transferred by convection into every part of the solution inside the vials.

Officially 121°C for 8 minutes is enough. Add 4 min for heat transfer into the vial. 30 min is overkill I added before the experiment to be sure that the temperature 121°C is reached inside the vial too. For 100°C the theory says longer time is required, but the heat-resistant pathogenic bacteria (tetanus, botulism) are anaerobic, oxygen containing in the air inside the vial is a poison for them, and the waterless solution is not nutritious for them (unlike canned meat - the best growth medium for pathogenic bacteria). For canned meat CDC says 110°C is enough.