One small point with regard to dwell time and RMV. The curve of respiration is a sine wave - slow initial velocity, fast mid-portion, and a trail-off at the end.
Think of breathing your max, say 40 full chest-fulls pet minute. Your can't sustain that, I agree. But let's look at the numbers. If you're a huge guy, 2.5 liters per breath x 40 = 100 lpm, which equals 3.5 CFM. You're above the testing limit, but as
@tbone1004 points out, that's not sustainable, and it's therefore not a stress to the scrubber as tested.
But let's look at the dynamic of it. Let's presuppose hard work, but only half of that max, i.e., 50 lpm (1.75 CFM). Easily below scrubber max. But what is the max flow? The 50 lpm is both in and out, so at a minimum, the 50 liters is going out during only half the minute, which makes the minimum expiratory flow 100 lpm. And due to the acceleration and deceleration phases of each breath, max flow may only be occurring during 60% of each cycle. In other words, if 85% of each breath is packed into the middle 60% of each cycle, the max flow is now not 50 lpm, not 100 lpm, but 141 lpm or nearly 5 CFM.
Now do the same calculations at 132 feet, or 5 ata. Five times as many molecules are passing thru the scrubber, irrespective of WOB.
You can begin to see how you might actually get CO2 breakthru at depth with a workload, irrespective of the added CO2 stress of increased gas density, and potentiality lesser volume movement.
Now during CE testing, the artificial lung is supposed to mimic all this, but I don't know that for a fact. So it is not inconceivable that despite a scrubber capability of handling "3CFM for 120 minutes with lots of added CO2", that high gas flow in mid breath coupled with decreased granule efficiency from thick gas at depth might allow some CO2 breakthru.