But the gradient won't be across a dozen kilometers - in fact if it were then the vast majority of the vented CO2 would be lost to interplanetary space rather than getting anywhere close to Mars. To be effective this system would have to be designed to resist a laminar flow distribution, so that virtually all the gas leaves the funnel at *exactly* the same speed. Any molecule moving at a speed off by even 0.001% will never reach Mars. We're trying to throw a dart at a bullseye moving at 24.1 km/s, hundreds of millions of miles away, with zero possibility of fine-tuning the path after launch.
As a matter of fact, even the initial thermalized lateral motion of the gas molecules would pretty much render it impossible to hit Mars - We're talking months to years of transition time, even a few dozen m/s of lateral motion will cause the gas cloud to expand so rapidly that I doubt more than a tiny fraction of a percent would even make it to Mars's Hill's sphere, much less the planet itself. You would have to find a way to completely eliminate all thermalized motion, essentially creating a giant highly collimated particle accelerator with negligible inter-molecular interactions rather than an air gun.