Tuesday, February 05, 2013

Conducting Plastic Experiments

With the increasing number of multi-colour RepRaps out there (Ahem! Including ours at RepRapPro Ltd...), it will also be possible to use them to print in mixtures of materials with different engineering properties.

Measuring the resistance of a short length of conducting
filament extruded from a standard hot end. The 1.75mm filament
used as input was made by rolling and can also be seen.

We have been experimenting with putting conducting particles in printing plastic, inspired by this paper:

They used carbon as a filler, which works nicely, but gives rather high resistances.  So we tried silver instead (expensive...) to get the resistance down.  This didn't work very well because of the shape of the silver particles.

The silver particles are flakes, and the carbon particles are fluffy and dendritic.  We hypothesised that the shape of the carbon allows lots of particles to touch each other (or at least to conduct between them synapse-like by quantum tunneling), but gives the high resistance.  So a mixture of the carbon to get the geometry of the particles right and silver to introduce short circuits between the carbon projections might work even better.

There was also a problem with the silver when it came to contact resistance at the surface.  It may be that the spiky carbon tends to stick out of the surface, whereas the flatter silver ends up with a plastic film over it.

It seemed like mixing different conducting fillers may be the way to go.  So we have tried experiments with a mixture of both silver and carbon, which seem to work better in combination than either on their own.

Full details can be downloaded from the RepRapPro Github repository here.

Finally, another possibility is mixing a magnetic powder into the filament, for which we don't care whether the particles touch or not, of course...


  1. Nickel is generally what they use in conductive pens, so that is also an element to try.

  2. If you had something both magnetic and conductive perhaps the particles would tend to stick together to produce a better connection.

  3. We have tried nickel in the past, but the problem with it is the oxide that forms on its surface, as with all metals (silver oxide is special in that it conducts, though poorly; but the film is very thin).

    I like the idea of using magnetism to "link up" conductors, though they might tend to clump instead. As always, experiment will tell us :-)

  4. We tried and failed with less suitable graphite powder¹. I am now trying some carbon nanotubes² that are much better conductors, make printing very hard. Do you have a source for small quantities of Black Perl 2000?
    What temperatures where you able to extrude at?


  5. I always wanted to try it out, and looks like I can try it out in about 6 months from now.

    I aim for aluminium powder for start:)

  6. For metal sintering, they use tiny metal spheres - created by praying molten metal into a vacuum and sorting out the particle sizes. Surely a sphere would work better than plates?

  7. Great work!!
    Yeah, it seems like the particulate geometry has a lot to do with conductivity.

    It seems like these people make particles in just about any size and composition:

    and nanotubes:

    Where you can buy in modest quantities for testing. Some of them are quite pricey....

  8. Cant help but thinking about if electric fields might help pull things together.

    The leads need to be attached, otherwise it will move the internal charges
    and have no E-field inside, since it is conductive enough. It will subsequently do

    (all back of envelope calculations, infinite sheets)Attaching leads, you get a current,
    assuming 100W/cm is ok, assuming 0.1Ωm, i get ~100V/m, supposing 10μm on
    'cube ends' of bits of silver/carbon gives forces of about 10^-17 N, well maybe
    ten-hundreds times that because that assumes uniform electric field, rather than the
    field zeroed inside bits of conductors. Compared to 2⋅10^-12 N of gravity on
    carbon of the same ends. So it wont work.(Note: if it did reduce resistance, power

    I suppose next up is to try AC, each unconnected filament is an RCL circuit.
    Two filaments oscilating in sync would always have the charges near to each other
    opposed, and attracting, probably the frequency is pretty high.
    Maybe a microwave might do it.

    I doubt the numbers are favorable, on the other hand the individual ends only need
    to move like 50μm. Depending on how much effort it is, it might be worth trying. Not sure
    how to calculate when exactly things start moving. In the solution, the vicosity
    is lower. (The LCR response i know how to calculate, not sure what a regular microwave
    amplitude is)

    All these effects presumably lead to conductivity depending on direction.

  9. The oxide on the surface doesn't tend to be a real problem for conduction, due to tunnelling, as you've mentioned. Oxidation can transform finely-divided metals into oxides pretty quickly, though. Graphite forms a galvanic couple with silver, and tends to corrode it; with a complete coating of polymer, so that no moisture is present, this might not be a problem, but at high powder loading moisture might percolate through.

    You might want silver nanowires instead of straight-up silver powder.

    A quick web search turned up a method to synthesize them within a liquid, which might well be a route to incorporating them into polymer:

    DOI: 10.5772/39491

    Thankfully, polystyrene is very easy to dissolve, and AFAIK is compatable with ABS. A high loading of silver nanowires would probably give it even better impact resistance (on a volume basis) than ABS.

  10. 3M makes a conductive polyurethane. You might be able to use that as a binder, which would decrease the resistance between conductive particles.

    You have to contact 3M to get details as the information is not available on their site. When I contacted them in 2009, they were willing to ship me 1kg of sample if I paid for the shipping.

  11. Low aspect ratio particles will require a high concentration because they are unable to rely on percolation effects like the higher aspect ratio materials. For spherical particles ET only occurs through tunnelling which, loses about 0.8 orders of magnitude per angstrom of tunneling distance. Spherical particles have to be within 1nm or so to have any decent conductivity whereas carbon nanotubes in most polymers reach a decent conductive state in many polymer systems at up to 5% incorporation. I would love to chat with Adrian about this as I am preparing an undergraduate exercise on this topic.

  12. Hi Rhys!

    My name is Amy and I'm a Materials Science Student from Univ of Pennsylvania and working with a group of 4 other students on a project where we're trying to print conductive materials from a Makerbot

    So far, we've had limited success testing conductive inks, paints, solders and we really could use your help in continuing with our best course of action.

    I've been scouring the web to find an email or direct way to contact you other than the blog and I couldn't find anything so I decided the best way was to submit a comment!

    Let me know if you're interested in learning about our project/swapping valuable knowledge! We think we could both benefit from an exchange from ideas.

  13. Hi Rhys!

    My name is Amy and I'm a Materials Science Student from Univ of Pennsylvania and working with a group of 4 other students on a project where we're trying to print conductive materials from a Makerbot

    So far, we've had limited success testing conductive inks, paints, solders and we really could use your help in continuing with our best course of action.

    I've been scouring the web to find an email or direct way to contact you other than the blog and I can't find anything.

    Let me know if you want to learn more about our project. I believe an exchange of ideas would really benefit us both.

  14. Hi Amy Wy,

    There is a project on Kickstarter currently (The EX¹ - rapid 3D printing of circuit boards) and the team seems to have found a solution to the problem of printing circuit boards on any material. They are using two solutions that react to form a circuit. They use an inject type printer. You might contact them to see if they could share some of their secrets. It would be great to have a public domain solution.

  15. I've been doing some experiments over the years in novel circuit design.

    I've just bought and built a Prusa Mendel and I'm looking at conductive printing but it looks like everyone has pretty much given up in anything but semiconductive prints.

    @nop head:
    I came across a really old technique used in the early days of radio that can be used to make conductive paths through conductive powders. The device is called a Coherer.

    It consists of a small insulating tube with conductive ends filled with a powdered metal. This can be ferrous or not, but it must be conductive and fine. To use the device, it is tapped gently to disperse the powder and raise the resistance to practical infinity against a low voltage potential applied across it.

    To make it conduct, it needs a sharp EM discharge close to it. The static fields induced in the particles make them rearrange themselves and align to the field present in the contacts. This also sticks them together to make a conductive path and lower the resistance to as little as a few Ohms.

    I have successfully laid out paths of copper, brass, nickel and iron powders in an open channel and made them conduct using the discharge from the 'snapper' in an electronic lighter but it has to be very close and usually needs several attempts moving around the path to get a conductive trace.

    At this point it is very delicate and with a little knock on the desk it falls back into a nonconductive random state. Useless as a circuit but I am working on a way to keep the particles together. Sintering is one idea, as is capillary soldering. Vapour deposition might work, but trying to electroplate doesnt.
    I did try creating the powder traces on a wet glue surface with some success when it was dry.

    Does this spark any ideas? ;)