Osmosis Demystified — How It Works

©2003 Darel Rex Finley. This complete article, unmodified, may be freely distributed for educational purposes.

“Osmosis” is the process by which small molecules automatically cross a semi-permeable membrane, compensating for a difference in the concentration of those molecules on either side of the membrane. But how do the molecules know to do that? What makes it happen? I have found the standard explanation to be a bit fuzzy and mystical; hence this tutorial.

First, the standard explanation of osmosis:

A tank contains water. A quantity of sugar-water is introduced at the right end of the tank.
Moving about randomly, the sugar molecules scatter throughout the tank until they are fairly evenly distributed. This can be interpreted as the water moving from an area of high water concentration (the left side of the tank) to an area of lower water concentration (the right side of the tank). Water naturally moved “across the concentration gradient,” from high to low. (Notice that the sugar did the same — it moved from an area of high sugar concentration to an area of low sugar concentration.)
Now we perform the experiment again, but this time with a semi-permeable membrane separating the sugar-water from the pure water. The membrane has tiny pores in it which are large enough for water molecules to pass through, but too small for sugar molecules. What happens this time?
The water flows through the membrane into the right side of the tank — this flow is called “osmosis.” The membrane swells under the pressure of the invading water — this is known as “osmotic pressure.”
What made the water do that? The standard explanations are:
A. Water wants to move across its concentration gradient, from high to low, and this requires it to cross the membrane.   or   B. Molecules on both sides are bumping into the membrane. Because there is a higher concentration of water molecules on the left side of the membrane, water molecules will attempt to pass through a pore more often from left-to-right than right-to-left, resulting in a net flow of water from left-to-right.

Sounds good, doesn’t it? But something’s wrong.

First, A. Water does not “want” to cross a concentration gradient. In the non-membraned tank, it did so simply because all the molecules were free to move around any which way, and their random movements scattered them randomly. But with the membrane, the sugar molecules are trapped on the right side, so why wouldn’t the water just flow randomly through the membrane’s pores in either direction, causing no osmotic pressure?

And, B. Yes, the concentration of water is lower on the right side, but what does that mean to any individual pore? If a sugar molecule is not nearby the pore, the water concentration is the same on both sides of the pore, and if a sugar molecule is close enough to the pore to prevent water molecules from escaping right-to-left, it is also blocking water molecules from entering left-to-right. When a sugar molecule blocks a pore, it blocks water flow in both directions, indiscriminately.

When the sugar molecule is not blocking the pore, the water concentration is the same on both sides of the pore.
When a sugar molecule is blocking the pore, water cannot pass through in either direction. Water molecules trying to get through the pore from the left may beat against the sugar molecule, but water molecules are also beating against the sugar molecule from the right.

Obviously, another explanation is required. This is it:

Let’s put in a non-permeable membrane. It has no pores, and nothing passes through it. Although the membrane is elastic, it is not swelling in either direction, so we know that the force of the water molecules beating from the left is exactly counterbalanced by the force of (fewer) water molecules and some sugar molecules beating on the membrane from the right. In other words, there is no absolute pressure differential between the two sides.

Now we snap our fingers and the membrane changes to a magical material that is impervious to sugar, but does not interfere with water molecules at all! Water molecules pass right through it, and it through them, as if they weren’t even in the same universe.

What happens? The water pressure on both sides disappears completely, because water cannot affect the magic membrane, but the force of sugar molecules beating from the right is still there. This unbalanced pressure causes the membrane to swell to the left! And as it swells, the membrane passes magically through water molecules, encompassing more water on the right side of the membrane. The water did not “move in” and swell the membrane — the sugar molecules pushed the membrane to the left, causing it to encompass more water. When the membrane resists stretching further, and stays in its swollen state, it is the sugar molecules that are holding it there by beating against it from the right — the water molecules can do nothing to the membrane, because they pass right through it with no effect.

Of course, no such magical material exists. But a good approximation is a membrane with small holes that allow water to pass through. In that case, the solid parts of the membrane experience a balanced impact of water molecules from the left, and water and sugar molecules from the right. The pores experience no water pressure in either direction, but do receive sugar pressure from sugar molecules bouncing off the pore’s opening (unable to fit through). The effect is the same as with the magical material described above, although it is weaker since only the pores have the “magical property.”

Mystery solved — osmosis explained entirely by random movements of molecules, without reference to mysterious “tendencies” to cross gradients! Water’s tendency to cross its concentration gradient is properly interpreted as an effect of osmosis, not as its cause. Osmosis is caused by the large molecules bouncing against pores through which they cannot fit.

Reverse Osmosis: What is “reverse osmosis?” It’s just fancy term for pushing water through a very fine filter to purify it. That’s all.


UPDATE 2008.10.05 — Over the past year or two, a few persons have e-mailed me to suggest that this explanation may not apply to the case of a rigid membrane, and I have to agree that it doesn’t. For example, suppose you have a U-shaped tub filled with water, like so:

Now if you separate the two halves of the tub with a rigid, water-permiable membrane, and add equal volumes of water on one side and sugar-water on the other, then the sugared side will rise up a bit, and the plain water side will sink down. Clearly, water is, in this case, invading the sugared side by crossing through the membrane from the non-sugared side of the tank.

The explanation must be different, but it still isn’t, “water wants to cross the gradient” — water doesn’t “want” to do anything but bump around randomly against nearby molecules (and react to being bumped by them). Nor is the explanation, “water more frequently tries to pass from left-to-right because there’s a higher concentration of water on the left side” — that is dispelled above (see close-up diagrams of a single sugar molecule blocking and not-blocking a pore), and if it were true, the water level on the left side would sink all the way down to the top of the membrane, which it does not.

So how to explain it? I don’t think it can be explained without referring to water pressure, which is how hard the average molecule is pushing in whatever direction is is currently trying to go.

In the first diagram, the water pressure is greatest at the bottom of the tank, because the weight of all the water on top of the bottom water — with nowhere for the bottom water to go — increases the pressure of that bottom water. (The exact mechanics of that, in terms of molecules pushing against each other, will not be discussed here, although I think it would greatly help in completing this explanation.)

This is why the water level stays even in the first diagram. If you pour some extra water in one side, then the pressure at the bottom is greater on that side, and it enables that water to push into the other side until the levels are even on both sides.

Now to the second diagram. We will assume for the sake of the discussion that sugar water has the same weight per unit volume as the plain water. Still, the sugar water will rise. Why?

The best explanation I can offer (without getting into the exact mechanics of how molecules behave in a liquid, which I don’t pretend to understand) is that a portion of the pressure on the right side of the tank is being wasted with molecules that cannot go through the rigid membrane. On both sides of the tank, the water pressure is “wasted” when the molecules push against the bottom and far sides of the tank, and also when they push against the solid parts of the rigid membrane. But the total pressure exerted through the pores of the membrane (against the liquid in the other side of the tank) is reduced on the right side by the presence of molecules that can’t fit through a pore, and thus can’t invade the other side at all.

I’m aware that this explanation sounds more than a little bit like the second standard explanation (B) that I discredited above, so I’m still thinking about this one.

In the meantime, I should mention that the flexible-membrane explanation is the one most relevant to cellular biology, since cells have flexible, semi-permiable membranes, and do not rely on gravity to move water from one chamber to another.


Impossible Diagrams

If you do a Google image search on osmosis, you will see a lot of pictures that look like this (spotlighting added):

In these diagrams, the semi-permiable membrane goes all the way up to the top, yet the diagrams show uneven levels of liquid on the two sides of the membrane. This is not possible, because water would immediately leak through the membrane from the high side to the low side, as indicated by the arrow in the below figure:

So clearly, the authors of those diagrams didn’t understand what they were drawing.

An interesting question, however, is what would happen if you constructed the arrangement described by that diagram. Would the osmotic effect in the vicinity of the bottom of the tank cause water to move from left to right, while water spilled from right to left at the top? In that case, there would be a net flow of water in a circular path — perhaps a perpetual motion machine, which (for good reasons) is considered impossible.

I don’t know the answer to that one!


Update 2014.05.01 — Not exactly related to osmosis, but here’s an article that explains how dictionaries have, for the past century, incorrectly identified air pressure as the cause of the siphon effect, when actually gravity is responsible.


Update 2015.04.17 — Gregor Kikelj suggests an answer to the perpetual-motion scenario described above. Instead of water spilling right-to-left through the top part of the filter, a very small amount of water would seep through, then it would stop and stay in place. The osmotic tendency of the water to flow back through that top part of the filter (left-to-right) would prevent any more water from coming through, and the water’s tendency to bond to itself (a.k.a. surface tension) would keep it clinging to the left side of the top part of the filter, so it wouldn’t fall down into the main body of water in the left tank. Genius!


See also: Maxwell’s Demon — Three Real-World Examples


Osmosis links:

How Does Reverse Osmosis Work?
Model Of An Osmotic System

Reciprocal Links:

Ben Best’s Science Pages

Send me an e-mail!


Back to Tutorials.