tj,
While CO2 I suppose can be contaminated, as much as any gas, CO2 itself is not a contaminant. Its essentially a combination of elements in gaseous form. Carbon's alloying properties are well known, if not well understood in many circumstances, but what is often overlooked is oxygens beneficial properties as well.
Many think in terms of deoxidation, deoxidation, and therefore oxygen must be eliminated. But much research has been done that confirms that oxygen content can be critical in the developement of acicular ferrite in carbon steels and actually improve toughness properties. Much of the balancing of alloying elements such as Si, Mn, Ti, and others is done with an eye towards not reducing oxygen too much. Especially with SMAW electodes and SAW fluxes.
And consider this from a theoretical standpoint, there is in some instances concern over the strengthening and hardening of weld deposits using ER70S-6 (as opposed to S-2 or S-3) wires due to the higher Mn contained. Higher oxygen could actually relieve the weld deposit of Mn by oxidizing it and therefore improve toughness and ductility. This would come in handy if someone found themselves with a spec that required some hardness control and did not wish to adjust their procedures.
Just theoretical on my part right now, but I wonder if there has been research to confirm such thinking. My guess is there has. I just haven't found it yet.