Printing Fiber-loaded Material

I used to work for one of the major plastic companies, and they had plenty of engineering plastics which were loaded with fibers to give enhanced structural properties. Some were loaded with glass, and some were even loaded with Kevlar.

I'd like to know if anyone out there is doing 3D printing with fiber-loaded plastics. If not, then why not? Is there some reason why this cannot be done?

Furthermore, whereas in traditional injection-molding the fiber orientation is determined by the way the injection wavefront travels, I'm thinking that 3D printing could offer the possibility of more control over how fibers are arranged across the part.

Can others please educate me on what the state of the art is in 3D printing of fiber-loaded materials? To me, composites are a significant step beyond ordinary plastics.

I think it can be done. We're just starting to see materials like polycarbonate, nylon, and PEEK. Currently, there's not that big of a market for retail sale of specialty printing materials. When you run filament, it's a lot easier to stick with a single material or you've got to purge the extruder before you load a different resin. Then just running 5#'s or so for one customer isn't exactly economical. Unless you have deep pockets, or have your own extruder, you'll probrably not find fiber filled filament.

Your maximum particulate size is limited by the nozzle orifice. Typically this is 0.35 or 0.50mm. Add a factor of safety on there and now maybe your max particulate size is 0.25mm. Any larger than this and you run the risk of jamming the nozzle.

I am not knowledgeable on how this limitation stacks up against normal injection molded FRPs, and how much strength one would expect to gain from using particulates this small.

An additional concern would be if FRPs would cause accelerated nozzle wear. We typically use brass or aluminum for our nozzles which is much softer than the materials used for injection molding machines and injection molding tools.

Finally, the major weak point of FFF parts is bonds between the layers. I don't see how FRPs would improve this, and in fact might make it worse.

Maybe you could drop a layer of 'shavings' between layers?

At Dupont, one of their secrets was to coat the fibers with Teflon, to allow them to slide through the nozzle more easily. So it could be possible to pre-treat the fibers with some kind of agent to get them through the nozzle more easily, and also to help them bond to the surrounding polymer matrix more effectively.

When the size of the particulate is physically too large to fit through the nozzle orifice, you can coat it with anything you want, it's still going to jam the nozzle.

So then don't make the fiber size larger than the nozzle. What's the problem? Many little fibers add up to a whole lot of strength.

Quote

So then don't make the fiber size larger than the nozzle. What's the problem? Many little fibers add up to a whole lot of strength.

Because short fibers add less strength than large/long fibers. My understanding of the mechanics of how FRPs work is many little fibers do not work as well as fewer larger fibers. To some degree this is why woven mat fiberglass is much stronger than chopped strand, and why chopped strand fiberglass is stronger than epoxy filled with, say, microballoons.

There's an upper limit on how much stuff you can add to the plastic before you start to adversely impact its material properties. The answer to using short fibers is not to add more of them - this will make the plastic extremely brittle.

I wonder if you could find a filler that would cut costs and maintain structural rigidity?

I wonder if fibers could be added by a separate route, other than the plastic extrusion nozzle? After all, this is additive fabrication and not injection molding. If you're adding the plastic a little bit at a time, then why couldn't you add the fiber a little bit at a time through a different additive mechanism or printhead?

What sort of method could be used to add just the fibers themselves, independently from the plastic extrusion nozzle?

And just think, if you could do this, then it might allow you more control over the arrangement of the fibers themselves. With current injection molding technology, your fibers are aligning themselves with the flow of the melt front. In the case of additive fabrication of the part, this won't have to be the case.

If you see the additive approach as a new opportunity to better align the fibers, by adding them through a separate means, then it might result in benefits that aren't possible through classic injection molding of fiber-loaded plastic.

I don't see any way to add fiber to the extrusion after it has left the nozzle. That kind of setup would be overly complicated. It would be much easier to use a filament pre-loaded with fiber.

I don't think you could use a .35mm nozzle for this, but a +.5mm nozzle shouldn't have a problem passing the fiber. Pretreating the fibers is interesting, but also overly complicated.

A carbon fiber could work well. I've heard that Goodyear uses carbon nanotubes in their tires to add strength. They are also small enough to flow through any nozzle commercially available and I don't think they would interfere with adhesion. That would be your best bet to add strength and maintain print quality.

crispy1 Wrote:
-------------------------------------------------------
> Your maximum particulate size is limited by the
> nozzle orifice. Typically this is 0.35 or 0.50mm.
> Add a factor of safety on there and now maybe
> your max particulate size is 0.25mm. Any larger
> than this and you run the risk of jamming the
> nozzle.

To touch on this point, as someone with some expertise in granular materials, I would say that a particulate size of 0.25mm in a 0.50mm nozzle would almost guarantee jamming, immediately and often. Frankly, people making hourglasses find that the constriction needs to be at least 5 particle diameters wide to eliminate jamming. This is without any plastic matrix around it, of course, but the particle density in the matrix may be much smaller than the sand in an hourglass. Worse, if the fibers are elongated, then it will jam very quickly even at low densities unless the fibers are well smaller than 0.1 x nozzle diameter. I don't have any hard evidence to make these numbers exact, but I do have a Ph.D. in physics, and granular flow was my thesis topic, so I have a lot of academic expertise and knowledge of best practices in industry.

In terms of composite materials in 3D printing, one approach would be to directly print the composite by using two different heads, and allowing the printing process or a subsequent heat or irradiation treatment to fuse them into a unified structure. Fibers could only conveniently be laid in the x and y directions using current methods, but z isn't out of the question with changes to the software. I suspect that printing a strong internal structure of one material would be quicker, cheaper, and more productive than a composite material when created by a 3D printer

Regards,
aeronaut

I think another problem is there is no reason for these fibres to link across layers. The limit in strength is the layer to layer bonding and I don't think this would offer any improvement :s

@aeronaut: Interesting, it's good to hear from someone with actual experience in this stuff chiming in :-)

Quote
konwiddak
I think another problem is there is no reason for these fibres to link across layers. The limit in strength is the layer to layer bonding and I don't think this would offer any improvement :s

Quote
crispy1
Finally, the major weak point of FFF parts is bonds between the layers. I don't see how FRPs would improve this, and in fact might make it worse.

-_-

Derp....... I was just emphasising........

There are different kinds of enhancements - like tensile strength, compressive strength, hardness, etc.

You can add a rubber for compressive resiliency, you can add carbon black for thermal conductivity, and even some electrical conductivity. I'm sure that nanotubes could significantly alter some of these properties.

Quote

In terms of composite materials in 3D printing, one approach would be to directly print the composite by using two different heads, and allowing the printing process or a subsequent heat or irradiation treatment to fuse them into a unified structure.

I don't recall seeing this before, seems like someone would have looked at it, but this seems promising? I am thinking of resins that harden with heat treatment. I guess there be a problem with gases given off, you might not want to do it in your kitchen oven.

Some metal printers mix resin and metal particles, I think the temperature required for curing must be quite high.

It seems the only way to settle this is for someone to try it and see if it works. But the point is moot, because 3mm or 1.75mm FFF's don't exist, to my knowledge. And a magic dual extrusion fiber fusing gizmo thing doesn't exist either.

Unless you count spiders and silkworms

Edited 1 time(s). Last edit at 02/20/2013 11:07PM by ohioplastics.

well I'm new here but have a cnc machine I could make some parts for to make an extrusion system that I could convert to use a 3d printer extrusion head and heated bed no problem. I'm interested mostly in more advanced materials like a carbon fiber mix. I wonder if you could make a system that makes a combination of carbon fiber filament and a coating of plastic material to make your own filament. Manually you might be able to do it by chopping up a fine fiber and brushing it on after the printer head (though time consuming) would be interesting to test just the different of strength of a few printed layers

Aren't the various types of 'wood' style plastic (here, for example) basically pla loaded with wood fibers/particles? If wood doesn't interfere with or clog the nozzle then much smoother glass or carbon fibre shouldn't be a problem, as long as it's kept under a certain grain size (research needed?)

CNT-infused nylon or polycarbonate or somesuch could yield a crazy-strong polymer matrix. I'd really like to experiment with some of that... Let's just hope the guys developing bulk production of said nanotubes gets the process sorted out.

We tried 30% and 40% long fiber glass filled nylon and PU.

Its a disaster. Don't do it, or you will clog any nozzle and back it up for miles. Its like hair drying to go down a drain, just ugly, messy. Even if it extrudes, it comes out looking like a strange spider web of fibers.

Its best to think of the fibers and being shard of rigid material in a nearly-liquid medium. So if you overheat it even slightly, you end up separating it into two phases.

Damn strong stuff though, if you could get it working. I think it was meant to be held under pressure and under tightly controlled conditions.

No, it is not PLA. Because the wood burns above 200C, they had to extrude in a polymer base that can work with lower temperature (or get a black brittle product). So the carrier could be something like PP, or PE. It is certainly made of some kind of "wood flour". They may even be using the same masterbatch from the companies using it for making plastic-wood decking and such. Those would be 10-30 micron particles-ish after some kind of filtering process I guess. They aren't going to be long fibers. So its a different game, and no clog is possible.

I love how they are distant from the resellers and vague about what the product is. it is 40% wood, (not 60% polymer) therefore it is "wood" and not plastic. Makes sense......right? So It i make 35% metal filled plastic does that make it metal? Great product, but I am suspicions about their carrier since they aren't revealing that it is PLA or not.

Having worked with and helped a student with his diploma thesis involving extrusion onto a rotary axis, I tend to think that the first experiments with fiber-reinforced FFF parts might be best carried out on a machine similar to his ("additive lathe"; already done before by others). If one were to add a second print-head that just guides a carbon fiber strand over the cylinder being printed in order to wrap it around, one can overextrude the next layer (or also partially re-melt the previous) to embed the fiber into the plastic.
Curiously, people already built tube lamination machines working on the same principle, but with epoxy instead of thermoplasts.

I already wrote a simple G-code output software (you can't call it a slicer yet) for the additive lathe. Adding code for a carbon fiber applicator to that shouldn't be too difficult.

About fibers aligning to Z: How about heating the fiber to a point where it can melt the extruded plastic and just ram it into the previous some layers? Sure it won't be a long strand that can be injected into the part like this, but even if it connects 3 or 4 layers, I am sure that gains in part strength should be noticeable.

I love this idea save for the magic step of a carbon fiber extruder. How will this work?

What exactly are you referring to, Simba?
If you mean the additive lathe, carbon fiber can just be wound around the workpiece and fixed in place by extruding over it.
The carbon fiber doesn't have to be extruded since the cylindrical workpiece acts like a spool. You just need a guide.

As for injecting short pieces of carbon fiber into a plastic part, I guess a pinch wheel similar to our direct drive extruders and a guiding pipe pressed against the plastic should be enough. After a certain length, the fiber can be cut off with shears.

Yes. How would we push carbon fiber?

The idea of fiber reinforcing plastic as you print takes real unique advantage of reprap has to offer. 40% fiber filled is an idea for injection moulding. With reprep, we may be able to do 60%, in a layer by layer approach.

In fact, i frequently place hot-nozzles that i test on glass or carbon fiber as insulation, and find the plastic makes it clear through several layers. But, to take advantage of fibers, they need to be stretch a little and we need to lay long long lengths. you can't push a rope, so how would be extrude it?

You all have overlooked one key factor, running glass through your nozzle will shred it, its basically like running fine sanding grit.

@Simba: Similar to how we push filament, I guess? As long as the pinch wheel is close enough to a guiding system and the carbon fibers go through a pipe butted against the plastic like in a bowden cable, the way with the lowest friction should be straight into the plastic part as long as the carbon is hot enough. Of course, after a short length, the carbon might want to buckle in the plastic, but the injected length might already be enough by then. That's the setup I think is most logical for reinforcing in Z direction.
For all other directions, the start of the fiber might be fixed onto the printed part by a little bit of extruded plastic. After that, not even a pinch wheel is necessary, just the movement of the carriage alone can deposit some fiber, again fixed with extruded plastic at the end of the track.

@aduy1: Why are you using the name of a registered forum member with a "1" added?
And as others pointed out above, the nozzles seem to clog anyway. So a dual-deposition system might work better (also yields longer fiber strands) if fiber deposition is is fact as easy as I imagine it could be made...

Apparently there is an SLS material called Windform XT which is a carbon fiber-loaded nylon powder. Maybe someone can put some of it through a Lyman extruder?

uGen, the additive lathe technique for using a fiber loaded polymer is quite clever - I can see a lot of applications for piping and other pressure vessel work (like the Ultem airduct that Solid Concepts made: [www.solidconcepts.com])

I got to chat a little bit with Scott Crump of Stratasys the other day about materials and asked him about fiber reinforced polymers. Nozzle erosion is a minor issue (the big industrial machines have their nozzles replaced after every 3 spools of material anyhow), but the primary problem is just what people have hit on - your inter-layer adhesion (which is already the weak point with a homogenous material) is even more of an issue, as the fibers do not cross between layers.

So, here is what I propose - the previously deposited layer needs to be somehow deformed or re-melted in order for the fibers to cross-link the layers together. I think using the concept of friction stir welding (look it up on youtube for videos of the process) is applicable, so here's the sort of extrusion nozzle I have in mind:



The green shell is a standard nozzle, the blue piece is simply a liner, while the purple part is a needle that rotates at high speed. The needle would plunge about 50% to 75% of the way through the previous layer. Since the needle tip's ambient temperature would be held at the material extrusion temperature, spinning it would would provide friction and enough additional heat to partially re-melt the previous layer, and the spinning would thoroughly combine the two layers, allowing the fibers to actually lock them together.

The reduction in cross-sectional nozzle area, however, is a detraction especially since it further reduces the workable length of the reinforcing fibers. On the other hand, this technique would be just as applicable to non-filled materials and could quite possibly allow printed parts to have truly isotropic mechanical properties.

---

[haveblue.org]