I looked into stratification of gasses by molecular weight for another
project I was working on, and the punchline is that it only happens at
close to the liquefaction temperature of the gas.
When a person goes in a confined space and dies because it's full of
carbon dioxide, it is basically never the case that the space filled
up because the CO2 is heavier than air. What really happens is that
the CO2 displaces the air via whatever process introduces it, and
there just isn't enough ventilation for it to diffuse out. That it's
heavier than air has little bearing; if the space was filled with
hydrogen the effect would be basically the same. Though spaces
unexpectedly full of hydrogen tend to explode before they have the
chance to suffocate anyone.
Citation: G. Badino – The legend of carbon dioxide heaviness. Journal
of Cave and Karst Studies, v. 71, no. 1, p. 100–107.
That said, you can get gasses to stratify based on molecular weight
when they are near the boiling point. It's actually possible to
separate oxygen isotopes with a sufficiently tall (few hundred feet)
column of very cold oxygen gas. I look forward to SpaceX switching to
enriched oxygen to further increase their LOX density. By throwing
away a mere 99.8% of atmospheric oxygen, you can create stable LOX-18
rather than traditional LOX-16.
In the context of a cryogenic tank with a mix of helium and oxygen or
nitrogen you're in a mixed situation where the O2/N2 may be close to
its boiling point in places, but the helium is still well above its
boiling point. In fact, as in most rocketry applications, it is so far
above its critical temperature that it's not totally accurate to call
it a gas.
So no, I don't think you're going to get the helium at the bottom of
the stack doing some weird gas-phase rollover where it suddenly wants
to be at the top of the tank. It would be interesting to have some
actual test data behind the idea that it will create a buffer layer
between the liquid and gas though. I like the idea, and have claimed
it on here when we had this exact discussion three or four times ago,
but some experimental proof would be nice. I'm sure a curious soviet
scientist ran ullage samples through a GCMS at some point.
Now, if you want a fun physical effect that could do surprising things
on a really big rocket that was filled with cryogen and left to sit
unusually long, like a Sea Dragon rocket that was filled and then
towed to launch position or a BFR that wasn't detanked after an abort
waiting for the next window, google the PDF "Stratification and
rollover in liquefied natural gas storage tanks". Includes the
important lesson that any cryogenic tank should have a vent system
sized to vent vapor at hundreds of times of the expected normal
boiloff rate.
Ben
On Wed, Aug 3, 2016 at 3:52 PM, Norman Yarvin <yarvin@xxxxxxxxxxxx> wrote:
On Wed, Aug 03, 2016 at 12:33:15AM -0400, Henry Spencer wrote:
Starting with a small amount of non-condensible gas (neon or helium) will
help greatly, because if you use a good diffuser so the incoming gas
doesn't stir the already-present gas too vigorously, you get a cold layer
of the non-condensible gas above the liquid surface, which tends to block
further condensation.
Wouldn't helium want to rise above the GOX or the N2 that you're using
as the bulk of the pressurant, rather than forming a layer? Even cold
helium is quite light.
