If air is a mixture, why don't the gases separate?

[AMan24]

If you try to mix water and sand, the sand will mix around and eventually fall to the bottom. Sand and water can't make a solution, so they separate. However salt and water can make a solution, and they don't separate. If air is a mixture, why don't the gases separate?

1) So, pretty much what I'm asking is, is there a term for a solution of gases?
2) Is it just called a solution?
3) Or am i completely off track and is it something totally unrelated?

4) I was also thinking, why don't they separate because of density?

 

[russ_watters]

The gases in the atmosphere do separate somewhat, but because the atmosphere gets pretty mixed-up by wind, they don't separate that much.

 

[Dale]

The gases in the atmosphere do separate somewhat.

I didn't know that! That is interesting. Do they just separate by density? There is no surface tension so I would imagine that any such separation would be gradual.

 

[anorlunda]

I didn't know that! That is interesting. Do they just separate by density? There is no surface tension so I would imagine that any such separation would be gradual.

I believe free hydrogen in the atmosphere is an example. Most of it rose to the fringes of space and disappeared long ago.

 

[Chestermiller]

At the laboratory scale, the main mechanism is rapid diffusion resulting from the high kinetic energy of the molecules. Even over most of the atmosphere, small scale turbulence combined with diffusion provides most of the story. Only high up in the atmosphere, where the mean free path of the molecules is much larger, does gravitational segregation play much of a role.

Chet

 

[sophiecentaur]

Only high up in the atmosphere, where the mean free path of the molecules is much larger, does gravitational segregation play much of a role.

Why wouldn't gravity have an equal effect all the way up (statistically)? It seems to me that, with or without collisions, the distributions of heavy and light molecules would have the same effect on each other, despite the random motions. Is there something 'different' at work when gases are well intermixed?

 

[Chestermiller]

Why wouldn't gravity have an equal effect all the way up (statistically)? It seems to me that, with or without collisions, the distributions of heavy and light molecules would have the same effect on each other, despite the random motions. Is there something 'different' at work when gases are well intermixed?

Hummm. Good question. I'm not sure. Maybe it has something to do with the gravitational force on each molecule staying about the same, while the frequency of collisions with other molecules decreases as the altitude increases.

 

[lightarrow]

Why wouldn't gravity have an equal effect all the way up (statistically)? It seems to me that, with or without collisions, the distributions of heavy and light molecules would have the same effect on each other, despite the random motions. Is there something 'different' at work when gases are well intermixed?

What do you mean with "the distributions of heavy and light molecules would have the same effect on each other"? I would say diffusion can't happen in absence of collisions.

--
lightarrow

 

[anorlunda]

Why wouldn't gravity have an equal effect all the way up (statistically)? It seems to me that, with or without collisions, the distributions of heavy and light molecules would have the same effect on each other, despite the random motions. Is there something 'different' at work when gases are well intermixed?

Even though we do not normally speak of buoyancy with respect to uncontained gasses, that is what is happening here on a molecular level. Lighter gas molecules are more buoyant than heavy ones. That is why a helium balloon rises. Pop the balloon and the helium gases rise anyhow.

 

[Chestermiller]

Even though we do not normally speak of buoyancy with respect to uncontained gasses, that is what is happening here on a molecular level. Lighter gas molecules are more buoyant than heavy ones. That is why a helium balloon rises. Pop the balloon and the helium gases rise anyhow.

This definitely doesn't sound correct to me. Are you saying that, in a sealed room containing helium in air, the helium will all segregate near the ceiling, and the air will stratify below?

Chet

 

[anorlunda]

This definitely doesn't sound correct to me. Are you saying that, in a sealed room containing helium in air, the helium will all segregate near the ceiling, and the air will stratify below?

Chet

Yes, that's what I'm saying.

Edit: Of course the boundary between the helium and air will not be sharp because of diffusion, but the layer should be there.

 

[russ_watters]

This definitely doesn't sound correct to me. Are you saying that, in a sealed room containing helium in air, the helium will all segregate near the ceiling, and the air will stratify below?

Not completely, but it has to separate a little, doesn't it?

I know due to Dalton's law we sometimes treat gases as completely separate(and therefore each occupying the full volume uniformly), but I don't think that is the reality for this context.

See: http://wordpress.mrreid.org/2014/08/01/the-composition-of-Earth's-atmosphere-with-elevation/

Now, virtually all of the variation is above 100km, which makes that hard to read, so I'd like to find some more detail of what those minor gases are doing down low.

 

[anorlunda]

From https://en.wikipedia.org/wiki/Helium

In the Earth's atmosphere, the concentration of helium by volume is only 5.2 parts per million.[76][77] The concentration is low and fairly constant despite the continuous production of new helium because most helium in the Earth's atmosphere escapes into space by several processes.[78][79][80] In the Earth'sheterosphere, a part of the upper atmosphere, helium and other lighter gases are the most abundant elements.​

The same applies to hydrogen. That is why we only find trace amounts of hydrogen and helium in the atmosphere today. It rose and drifted away into space in primordial times.

 

[nasu]

There are some caves rich in carbon dioxide and this gas collects closer to the bottom of the cave.
But I suppose that if the source of CO2 will stop the bottom layer will disperse after some time.
http://www.showcaves.com/english/explain/Speleology/Dogs.html

 

[Chestermiller]

Not completely, but it has to separate a little, doesn't it?

I know due to Dalton's law we sometimes treat gases as completely separate(and therefore each occupying the full volume uniformly), but I don't think that is the reality for this context.

See: http://wordpress.mrreid.org/2014/08/01/the-composition-of-Earth's-atmosphere-with-elevation/

Now, virtually all of the variation is above 100km, which makes that hard to read, so I'd like to find some more detail of what those minor gases are doing down low.

I'm not saying that there isn't a slight effect in a sealed room. But I am saying that, in a sealed room at atmospheric pressure, the molecular agitation would be adequate to guarantee that the amount of segregation would be virtually undetectable. There certainly wouldn't be a mostly helium layer adjacent to the ceiling.

Chet

 

[russ_watters]

Here's a NASA paper on the subject. Just skimmed, but the abstract mirrors our discussion and there is a graph on page 21:

https://www.google.com/url?sa=t&sou...UiWpNzQ5wHr49NbIA&sig2=yiLQRKqbkanOOF3CZVHG0Q

 

[anorlunda]

Also propane. Boats that allowed propane to leak to the low points in the bilge have exploded years after the leak was stopped. Despite limited air circulation, the propane remains layered in the low points.

 

[russ_watters]

I'm not saying that there isn't a slight effect in a sealed room. But I am saying that, in a sealed room at atmospheric pressure, the molecular agitation would be adequate to guarantee that the amount of segregation would be virtually undetectable. There certainly wouldn't be a mostly helium layer adjacent to the ceiling.

Chet

We're agreed. I did not intend to imply the effect would be significant for a small container. Whether it is significant for the lower atmosphere depends on your definition of "significant". But since that issue was the thrust of the OP's question, I think it is significant enough to mention.

 

[Chestermiller]

We're agreed. I did not intend to imply the effect would be significant for a small container. Whether it is significant for the lower atmosphere depends on your definition of "significant". But since that issue was the thrust of the OP's question, I think it is significant enough to mention.

I didn't see anything in the OP's question about air over vertical atmospheric distances on the order of 10's of km. Maybe it's just my perspective as a ChE to think small, on a scale on the order of meters rather than km. (Although I do have some actual experience as an atmospheric scientist).

Chet

 

[nasu]

Also propane. Boats that allowed propane to leak to the low points in the bilge have exploded years after the leak was stopped. Despite limited air circulation, the propane remains layered in the low points.

I read claims of a similar effect in wine cellars. The CO2 will come from the fermentation, in these cases.

 

[russ_watters]

I didn't see anything in the OP's question about air over vertical atmospheric distances on the order of 10's of km. Maybe it's just my perspective as a ChE to think small, on a scale on the order of meters rather than km. (Although I do have some actual experience as an atmospheric scientist).

Since he asked "why don't they separate because of density", I wanted to start off by congratulating him for his correct logic and telling him he's right that they do/should, before hitting him with the big "but".

 

[Chestermiller]

Also propane. Boats that allowed propane to leak to the low points in the bilge have exploded years after the leak was stopped. Despite limited air circulation, the propane remains layered in the low points.

So, in a room filled with stagnant air, the majority of the oxygen molecules will eventually be situated in the bottom 20% of the room, and the majority of the nitrogen molecules will be situated in the upper 80 % of the room (even though air is nearly an ideal gas, with the individual molecules traveling very rapidly in all directions)? How does that reconcile with your helium balloon explanation?

Chet

 

[nasu]

It does not follow. The difference in molecular weight between oxygen and nitrogen is much smaller than the difference between nitrogen and propane, CO2 or helium.
So if the diffusion is enough to equalize O2 and N2 in normal conditions it does not mean is enough in every case.
It would be interesting to have some quantitative estimate of the effect.

 

[DrStupid]

the majority of the nitrogen molecules will be situated in the upper 80 % of the room

That sounds plausible.

 

[Chestermiller]

That sounds plausible.

I sure don't want to try to breathe the air in the upper part of that room.

 

[Chestermiller]

It does not follow. The difference in molecular weight between oxygen and nitrogen is much smaller than the difference between nitrogen and propane, CO2 or helium.
So if the diffusion is enough to equalize O2 and N2 in normal conditions it does not mean is enough in every case.
It would be interesting to have some quantitative estimate of the effect.

In your professional judgement, do you really think there would be a significant effect in these other cases?

Chet

 

[russ_watters]

It would be interesting to have some quantitative estimate of the effect.

Have a look at the link I posted. The graph shows the concentrations of [molecular] oxygen and nitrogen basically constant up to 100km, but gases with much different molecular weights like H and O (monoatomic?) vary noticeably...though those don't exist at the surface.

All that said, I'm no longer certain of the mechanism of the variation. I've only skimmed the theory, but what I saw implies that the variation above 100 km can be entirely explained by Dalton's law and the different molecular weights, not by buoyancy:

Each constituent has a density vs altitude curve who's slope depends on molecular weight: heavy gases weigh themselves down and light ones spread themselves out.

 

[Chestermiller]

Have a look at the link I posted. The graph shows the concentrations of [molecular] oxygen and nitrogen basically constant up to 100km, but gases with much different molecular weights like H and O (monoatomic?) vary noticeably...though those don't exist at the surface.

All that said, I'm no longer certain of the mechanism of the variation. I've only skimmed the theory, but what I saw implies that the variation above 100 km can be entirely explained by Dalton's law and the different molecular weights, not by buoyancy:

Each constituent has a density vs altitude curve who's slope depends on molecular weight: heavy gases weigh themselves down and light ones spread themselves out.

Gases like atomic H and atomic O are in very low abundance because they are created and destroyed by very rapid photochemical reactions. So they can't be regarded on the same basis as inert tracer gases.

 

[nasu]

In your professional judgement, do you really think there would be a significant effect in these other cases?

Chet

My professional judgement tells me to prefer some quantitative parameters rather than just intuition.
I have no direct experience with gas separation.
I just said that the non-separation of N2 - O2 does not necessarily implies non-separation in all cases.

 

[russ_watters]

Gases like atomic H and atomic O are in very low abundance because they are created and destroyed by very rapid photochemical reactions. So they can't be regarded on the same basis as inert tracer gases.

Sure, below 100km it would be safe to say that O2 and N2 dominate and several others vary greatly or are limited due to chemical/thermodynamic processes (water, Ozone too).

Above 100km, it turns out that monoatomic oxygen dominates for a few hundred km, then monoatomic helium. This may be beyond the scope of the OP, but it fascinates me!

 

[Chestermiller]

My professional judgement tells me to prefer some quantitative parameters rather than just intuition.
I have no direct experience with gas separation.
I just said that the non-separation of N2 - O2 does not necessarily implies non-separation in all cases.

Wouldn't you think that the magnitude of the gravitational contribution would be roughly proportional to the difference in molecular weights. So if the segregation of N2 and O2 is zilch, the segregation between air (mw 29) and CO2 (mw 44) at equilibrium would be about 15/4 = ~ 4 times zilch.

 

[sophiecentaur]

What do you mean with "the distributions of heavy and light molecules would have the same effect on each other"? I would say diffusion can't happen in absence of collisions.

--
lightarrow

My wording was really poor, wasn't it?
I was thinking of partial pressures and that the gradients would be expected to be different. Yes, the diffusion will be slowed by collisions (as in a porous medium) but I am sure that the final gradients should turn out to be different for different gases, and the same that you would get if each gas were in isolation around a less massive planet (?).

 

[nasu]

Wouldn't you think that the magnitude of the gravitational contribution would be roughly proportional to the difference in molecular weights. So if the segregation of N2 and O2 is zilch, the segregation between air (mw 29) and CO2 (mw 44) at equilibrium would be about 15/4 = ~ 4 times zilch.

Are you sure it is a linear effect?
Do you have some quantitative basis for this? That will be interesting.

 

[anorlunda]

Let me propose a laboratory experiment that I bet Chet would agree on after some thought.

Take a vertical cylinder container of height L and cross-sectional area A. Let R=L/A.

We will fill it with 1/3 He (1/3 by volume), 1/3 air, and 1/3 CO2.

Now, if there are stratified layers, then mixing will occur at the boundaries between layers because of thermal agitation, or diffusion, or turbulence. But mixing efficiency decreases with A.

So for small values of R, agitation dominates and there will be no measurable layering (zilch layering). But as R increases, mixing efficiency becomes arbitrarily small. Something else must begin to dominate. What candidates are there other than buoyancy? For high values of R, buoyancy dominates and there will be measurable stratified layers in the steady state.

In the lab, rather than directly modifying R, we could start with a pipe, L=10 m, A=$1cm^2$. Then place 2 gate valves to partition the cylinder into 3 equal volume regions. The valves would initially be fully open. The experiment would then be to very gradually close the valves over a period of weeks or months while measuring the proportions of He, air, and CO2 in the regions.

 

[Chestermiller]

Here is a simple calculation that quantifies the effect. I have air (MW =29) and He (MW =4) in a room at 1 atm and room temperature. Such a mixture can be treated as an ideal gas mixture. In an ideal gas mixture, the different gases behave as separate entities. Let:

p_ao= partial pressure of air at the floor
p_h0= partial pressure of He at the floor
L= height of room (nominally 3 m)
z = distance measured upward from the floor
T = absolute temperature
R = ideal gas constant
M_a=molecular weight of air
M_h=molecular weight of helium

For each gas in the ideal gas mixture, the barotropic equation tells us that:

$$\frac{dp}{dz}=-\frac{Mg}{RT}p$$
So as a function of height z in the room, the partial pressures of air and of helium are given, respectively, by:
$$p_a=p_{ao}\exp{\left(-\frac{M_agz}{RT}\right)}$$
$$p_h=p_{ho}\exp{\left(-\frac{M_hgz}{RT}\right)}$$
So the ratio of the partial pressures (and mole fractions) at the ceiling are related to the ratio of the partial pressures (and mole fractions) at the floor by
$$\frac{p_h}{p_a}=\frac{p_{ho}}{p_{ao}}\exp{\left(\frac{(M_a-M_h)gL}{RT}\right)}$$
The term in parenthesis in this equation is equal to ~ 0.0003.

Therefore, the ratio of the mole fractions at the ceiling is equal to the ratio of the mole fractions at the floor times about 1.0003 (i.e., a variation of 0.03%). That's the big stratified separation that occurs.

For a small typical value of the expression in parenthesis like 0.0003, the relationship becomes:

$$\frac{p_h}{p_a}=\frac{p_{ho}}{p_{ao}}\left(1+\frac{(M_a-M_h)gL}{RT}\right)$$

So, to answer nasu's equation in post #33, yes I do have evidence that the change is essentially linear in the molecular weight difference.

Chet