"mass" of burning steel wool drops then rises. Burning oil or rising hot air?

Question

In a demonstration, a small wad of extra-fine steel wool is placed on a digital scale and set alight. Why does the measured mass drop initially?

Do you think the low pressure applied to the entirety of the area of the plate would be significant?

The mass is expected to rise as oxygen is gathered from the air. It does, of course, when one looks at the final results. BUT, while the wool is actively burning, the mass reading drops.

I have two hypotheses: 1) the steel wool is coated with oil(?) to reduce rusting before use and the lost mass is this oil being driven off by evaporation or combustion into CO₂ gas.

Alternatively, and more fun, 2) the release of energy heats the air above the scale. The hot air rises, causing a low pressure above the plate. The scale doesn't actually measure mass but rather net downward force on the plate and assumes that the force is entirely due to the gravitational force. A column of rising hot air breaks that assumption.

The explanation presented as "correct" is that the steel alloy has ~2% carbon and that is what is lost as CO₂.

Do you think the low pressure applied to the entirety of the area of the plate would be significant?

Answer 1

I think that might depend on the sensitivity of the balance (not a scale) being employed. It may not even be detected if not sensitive enough, and even if sensitive enough, it probably wouldn't be a significant contribution relative to the actual change in mass of the 2% carbon. But, it's an interesting concept to at least consider.