h a l f b a k e r y
Bone to the bad.
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
or get an account
I was thinking about how to make self-assembling nanostructures,
like they do with DNA etc, and wished it could be done on a larger
Now, imagine a rod (made of, say, brass). On each end, you put a
magnet. The magnet is shaped like one third of a cone (ie, if you
looked at it along
the axis of the rod, the magnet would look like
one third of a circle). The third-cone is magnetized in such a way
that one of the flat "cut" faces is a south pole, and the other is a
So, three of these "third-of-a-cone" magnets would tend to stick
together to form a complete cone. The complete cone would not
have much of an overall magnetic field, since all the poles of the
three magnets are connected in a sort of circle. (If you put two
horseshoe magnets head-to-head, there's not much field.)
So, we have three of our magnetic "third-of-a-cones" stuck together
to form a complete cone. The rods will project outwards from this
cone in a tripod structure.
But, of course, the rods each have a third-of-a-cone magnet on their
other (free) ends too. So, other rods would tend to attach to these,
and so on.
The result should be that, if you took a bunch of these rods and
shook them enough, you'd create a sort of self-assembled pyramidal
Or maybe not.
But, I think with enough ingenuity, it should be possible to make
magnet-ended rods which would spontaneously self-assemble into a
variety of different lattices.
(And yes, I know about the ball-and-stick magnetic sets you can buy,
but with those, you have to assemble the structure yourself.)
Self-assembling tetra kite?
[mouseposture, Jun 04 2010]
World of Goo...
[RayfordSteele, Jun 04 2010]
||Why does this sound like the Wii game, 'World of
||I don't know, why does it?
||You will want a mechanical automatic fastener system
that initiates when two cone faces are in alignment. This
way you secure your direct hits and can continue to
shake off your misaligned magnets. The mechanical
equivalent of Van der Waals forces, perhaps.
||You know, Daseva, that is not such a stupid idea.
||However, I was hoping that a two-magnet group would be
very receptive to a third member, whilst a three-magnet
group would be effectively inert (ie, a closed circle, no
exposed poles). So, I was hoping that the whole thing
work without the gizmetry.
||In thinking about this, imagine that instead of the rods-
with-thirds-of-cones, we just had a bag full of magnets,
where each magnet is a one-third segment of a sphere,
with north and south poles on the "cut" faces. If you
shook the bag, I bet you'd get a bunch of spheres.
||//I was hoping that a two-magnet group would be very
receptive to a third member, whilst a three-magnet group
would be effectively inert//
I'm sure no one would ever have suspected you were a
||Do you think this might work for implementing a self-
assembling tetra-kite <link> ?
||I wonder if you could make a set of self-assembling Platonic solids this way. Make the edges out of brass rod, and the vertices out of your circularly-magnetised spheres. Slice up the spheres as appropriate and you're left with a pile of brass rods with strangely-shaped magnets attached to each end. However if you put them on a cloth and give them a good shake you end up with a beautiful set of the five perfect solids.
||I wonder what would happen if you shook up a thousand of them all at the same time.
||//I wonder what would happen if you shook up a thousand of
them all at the same time.//
||You'd probably generate a solution for either (a) the
travelling salesman problem (b) Goldbach's conjecture or (c)
the public transport system in the UK.
||I doubt that this hasn't been thought of before, but it's an interesting idea nonetheless, so I'll vote towards neutral.
||On the one hand, you have to shake the rods hard enough so that they sample enough positions to assemble properly. On the other hand, if you shake them too hard, the vibrations could tear the structure apart. I am wondering what the outcome would be if you put these rods in some sort of dense, viscous fluid-like substance (that is popular around these parts) and used currants to drive the random sampling.
||I think the best use of such self-assembling atoms would be to create a large number of slightly more complex structures, which would then be assembled intelligently. These intermediate units would be simpler and therefore harder to break by shaking than compared to a complex structure. Moreover, the shaking becomes less efficient (i.e. energy/atom assembled) as the finished product has "nucleated" and is growing ever larger, assuming your pool of atoms isn't sufficiently large.
||To make a poor analogy to a protein, the secondary structure of the object would be formed with the information contained in the primary structure (a bag of rods), while the tertiary structure (the finished piece) would be assembled intelligently, as if by the aid of some series of chaperones.
||I'm already signed up for the dance, please don't call me
for this one.