I guess any gas is as good as another ???

My first impression was that your diameters were very Very Close ! The difference lies in controlling the tapers of the inlett, and output. This is nearly impossible to control by thermo forming pipe without a constricting mold to limit the expansion.

Secondly: For our purposes (Co2 Injection) we require a greater gas intake velocity, and a smaller orifice. The smooth transition of the stretched taper is actually an impediment. We require a sharp edge to the throat, and an increased pressure differential to chop up smaller bubbles of gas. (Carefully study the Mazzei Illustration) pay close attention to the offset diameter of the opposing pressure well. Mazzei Injector Corp. - How a Mazzei Injector Works This forces the stream to turn/cut across the intake port ripping the bubbles from the orifice. If the intake jet has a sharp proud edge it will punctuate the portion of gas more effectively, but there is a very fine point of adjustment before it actually impedes the flow, and decreases the liquid velocity. (This is why I used an 1/8 bulkhead to sleeve the capillary tube) it's threaded and has a nut on it to raise or lower the intake within the throat. I'm just lucky that it reinforced the tube or it likely would have burst as Tom pointed out ! ;^)

Combine this with a lower flow rate to sustain the reaction, and a significantly smaller volume of gas, and the task becomes a formidable challenge.

We can decrease and meter the gas intake, but we still require a significant shear force to tear the bubbles from the orifice.

The process is a vast contrast to O2 or O3 injection.

Last point of observation is that if you accomplish these goals a reactor is almost un-necessary. Tom's greater point being that we may not want to contain the reaction at all, but distribute the confluent to the tank creating a gaseous interface of Co2 through the micro bubbles, and the plants.

I've tested dissolved Co2 against misting, and those plants that enjoyed the introduction of Co2 micro bubbles thrived by comparrison. (Tom makes a Very Good Point)

So long as we are designing a more efficient diffusion we may as well finish the job, instead of resuscitating the aborted shortcomings of prior methods, and systems.

I DO understand that you're trying to refine a simple production process of fabrication, but don't lose sight of the larger picture. You may very well succeed, only to show up a day late, and a dollar short of the ultimate goal !