Now, this is potentially a good point you mention. They estimated that the tubing would pass about .3 cubic centimetres (.3 ml) of gas into the solution per hour. This gas would raise the conductivity of the top N2 gas layer, resulting in more heat loss from the solution. But results showed a heat gain instead. I'm supposing that you are thinking that 1) the hydrogen in a KCO3 solution might alter (lower) the specific heat very significantly while not doing so with a NaCO3 solution, or 2) the hydrogen would undergo some unknown but conventional chemical reaction with KCO3 which was very exothermic, but not with NaCO3.
Both 1) and 2) are not supported by evidence. A third option, hydrino chemistry, is.