Something that strikes me from time to time is the question whether it would not be possible to make fans more efficient. The places I mostly run into them are for cooling, so for example in computers or on hot summer days.

One thing that is immediately apparent about fans is the noise they produce. This noise is in part due to the motor when it is moving but this source is mostly negligible, thinking of motors with magnetic bearings nowadays. So most of the sound a cooling fan makes stems from the blades cutting throught the air and moving it around and about quite a lot. This sound can be reduced a little bit by designing the fan blades in a good way but still the big question remains:

Isn’t there a better way to make air move in a steady flow than using blades that swish through the air and push it forward?

Let’s think of the benefits for a second.

1. Having no moving parts mean that less things can break

2. In big fans, the blades are necessarily also big and a source of danger for unobservant fingers

3. The noise. Think about how silent the world could be when we needed no moving blades to make air or water move.


In my third point I already meantionend another application of the bladeless fan. It could be used to propel boats, making them quieter and hopefully more efficient.


Now, the benefits are clear but the big question remains. What possibilities do we have to make it happen? There is one technology which works in salty water and achieves something like that but I think it’s quite inefficient still and not so easy to implement. I’m talking about the magneto hydrodynamic effect, or MHD.

In short, MHD works by running a current through seawater (it’s salty, hence it’s a good conductor) and applying a magnetic field along the axis of the thruster. Because a lorentz force acts on any current flowing through magnetic field lines, the conducting material, sea water or ions, is pushed out the back and propels the ship. Alternatively, this principle can also be used to pump water. Read more about it on Wikipedia:


The question i’m asking myself now is, whether it wouldn’t be possible to come up with an efficient design for air which uses similar technology. There are some problems which arise when trying to apply this principle.

1. Air is a very bad conductor. Simply running a current through it will not work or require ionizing it first.

2. Having a big magnet close to the hard drives in a computer is still a bad idea. SSDs should not be affected though.

3. Air is comprised of many different components which might behave differently.


Let’s have a look at the average composition of air (Wikipedia):

78% N2

21% O2

~1% Argon

There are some small components like carbon dioxide and other things but the biggest parts we have to deal with are oxygen and nitrogen in molecular form.  The only possibility i’m seeing in using this composition to our advantage is using the molecules somehow. If the molecules were polar, like water for example it might be possible to do something with the tiny dipoles.

At 20°C, the saturation Humidity Ratio of water to air ist 0.014659, which is abou 1.5% of the weight of the air is water. This is already a very small number but since we almost never have 100% humidity the average figure will be much lower, I guess about 50% of that or 0.75% of the weight of air is contributed by water. That is not much but might be sufficient to get some airflow if the water molecules could actually be used.

The only chance there is to grab those water molecules and get them moving is in my eyes to use their dipole moment somehow. Let’s take a look at that.


A dipole, like the water molecule consists of one positive and one negative charge. Once these move in a magnetic field they are also subjected to the Lorentz force and thus deflected in a certain direction. If the molecule was made to “wobble” in resonance (by applying a high frequency electric field to the molecules), the two charges would move back and forth. Since their electric charge is opposite and their direction of movement is opposite at all times, the resulting Lorentz force would always point in the same direction and with the charges separating the whole molecule would move in one direction and then move back again when the charges contract again. The whole molecule would start some wobbly back and forth motion, synchronized with the motion of the charges.

There are only some minor problems.

1. The distance between the two charges is very very small. It is on the order of 3.9pm (picometres). That is less than the hydrogen atom radius at 25pm, so it is very small!

2. The molecule would wobble back and forth with the resonant frequency and so no forward motion would be generated.

3. Since the molecule is so small, the resonance frequency would be very high. About 1600GHz, which is a lot (microwave radiation is at 2.49GHz)


Solving this high frequency problem is a difficult technical task and the comparison with the microwave oven raises another question of possible heating of the water vapour instead of just moving the molecules (due to energy absorption).


Another possibility would be to use a much lower frequency for the electric field and use the effect, that a dipole aligns itself along the electric field lines. Applying a static field (like in a Capacitor) will make the water molecules align along the field, rotating if necessary. It is this rotation which might be used to actually propel them forwards.


Changing the polarity of the electric field would cause the dipole to flip into the opposite orientation and if a magnetic field is also applied to the region, a Lorentz force would result. The only remaining problem is now, that we don’t know in which direction the dipoles flip. Just switching the field would make some dipoles rotate clockwise and some dipoles would rotate counter clockwise, thus eliminating any effect there was by averaging the motion over many molecules.

One possibility around that would be to use at least 4 electrodes instead of just 2. By adding them perpendicularly to the first two electrodes, we have an effective way to control the direction of rotation of the dipoles.

By adding more poles and rotating the field accordingly, the motion becomes more controlled but that should not be necessary. The only thing needed now is a small calculation of the actual effect which is to be expected. We still need to know how long the molecules need to rotate to the new position and what field would be necessary to chieve the rotation of all molecules of air humidity.

As calculated earlier, the maximum content of water in air is about 1.5%. We use half this value to estimate the content of water vapour to 0.75% as an average value. Taking a tube with a distance between the two electrodes (the length of the faces of the cube) of 2cm and a length of 10cm (that seems like a reasonably small tube which would be useful for testing purpposes or cooling small things).

The volume in the tube is then 40cm³ and the corresponding weight of the air is (at approx. 1.2kg/m³ at sea level) approx. 4.8g, thus having a water content of 0.036g or 36mg. Not an awful lot but we’ll see where that gets us. At 18g/mole for H2O this should amount to 1.2e21 water molecules.

Each water molecule has an electric dipole moment of 1.85D (Debye) and in total 7.44e-9 C*m. To counteract this charge and hence align all molecules to an external electric field, with a distance of 2cm between the capacitor plates we would need a charge of 3.72e-7 C.

Now let’s calculate the capacity of our simple plate capacitor. It should be 8.85e-13 F (Farad). The voltage we need on those plates is the charge needed on the plates divided by the capacitance of our capacitor. That amounts to 420KV. Hmpf. That doesn’t sound like a safe voltage. Hope I calculated that right. Bringing the plates close together would help thing but only linearly. To reduce this to a somewhat manageable 1KV we would need to reduce the size (diameter) of the tube by the root of 420 (the amount of water and thus dipole charge depends on the diameter with a power of two and the capacitance cancels out the size of the plates duo to the symmetric distance / area). So, in total we arrive at a diameter of 1mm. Pretty small but not too bad. We could later attach many of such tubes in parallel and should still be able to create a respectable airflow (but friction will play a greater role.


So, the only thing missing is the question of how much a water molecule will be displaced in one whole rotation (or possibly also a quarter rotation at first since the rest derives from that).

The charge is said to be 10e, separated by a distance of 3.9pm (see here). That means, that in a quarter rotation, a charge of 20e moves by half that distance perpendicularly to the magnetic field (which can be fixed, the charges always move the same distance regardless of orientation). This is equivalent to 1e moving by 39pm. Now on to the magnetic field.

It seems that when calculating the Lorentz force, another problem comes to light that I didn’t think of. The charges cross the magnetic field lines in different directions when they make their round trip. This would lead to oppositely directed Lorentz forces and thus a net zero effect. The only way around that which comes to my mind now would be to change the direction of the magnetic field also on these turning points but as magnetic fields go, when they are strong enough, reversing them takes some time or involves very large currents and might be technically difficult to handle.

The remaining calculation I will do in a second part of this post. If you actually stayed with me until here, respect. I would be happy to discuss this topic, let me know what you think and whether I made some grave mistakes somewhere!




5 thoughts on “How about making air flow without moving parts

  1. zapp

    Good post, but I don’t think reliance on moving water molecules is a good idea — your cooling method would become less effective in drier environments such as deserts.

    Check out and this practical application of it:

    1. jens

      You’re of course right that drier air would impede the airflow with such technique. Thank you also for the links. I was not aware that such thing has actually been built already. I considered ionizing the air also but rejected this idea because I thought it was to impractical. It turns out that it was not 😉

  2. John5505

    I am interested in the idea and have a few ideas of my own but don’t have time or knowledge to flesh them out. I would appreciate any help understanding the following ideas.

    1. Can a certain frequency of ultrasonic/subsonic waves move air efficiently move air?
    2. Is the air around capacitors/inductors in a computer charged enough to move it with weak magnetic field?

  3. jens

    1. Sonic waves do not move any air, as it is only a compression wave. Air molecules move back and forth around their current position based on the amplitude (volume) of the sound
    2. I don’t think that the air around capacitors on a mainboard is charged at all and would probably not be affected by a magnetic field

  4. Half Time Full Time

    This is my first time visit at here and i am really impressed to read everthing at single place.

Leave a Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.