solarbird | Wood-filled PHA filament: impressions, test prints

A few weeks ago someone posted on Reddit about a more-compostable PHA filament that claims higher temperature resistance than PLA, so I thought I'd get some. And while I was at it I went all-in with the 20% wood variety.

What's most interesting to me is the claim of a much higher temperature glass point. They also make claims about being higher strength. That's also somewhat interesting, if less so.

The 20% wood infill variety requires a 0.6mm nozzle or bigger, according to the manufacturer. My experiences validate that, and I got the kind of filament flow in tests that I'd expect from a PLA with that nozzle size.

So I ran a couple of temperature towers, one with the parts fan ON (left side), one with the parts fan OFF (right side). The filament maker recommends the parts fan be OFF. All prints are sliced by Cura at 0.20mm layering for 0.6mm nozzle.

This first pair got me my first important data point: this stuff does not bridge well, but the parts fan does help:

Front view, left side parts fan ON, right side parts fan OFF

To be fair to the maker, the high temperatures are outside manufacturer recommendation, so... maybe you shouldn't. But I did. And there are circumstances where that appears to pay off; bridging is clearly best at 220°C with fan ON, as long as you don't mind doing a bunch of fine webbing cleanup afterwards. If you do mind, 180°C is clearly best.

And you can see the structural issues showing up on the right side with parts fan off. The left side at least kept structural integrity.

The backs of the towers are also interesting, and highlight the different characteristics of the stringing between parts fan on and off:

Back view, left side parts fan ON, right side parts fan OFF

I was, I'm pleased to say, able to get a reasonably good Benchy at 180°C:

you can't see it, but there is a little failure-to-bridge inside on the ceiling

This side had a little more issues with the arches; I don't know why

I also printed a 220°C version, it's definitely worse overall, and the spiders really got to it xD - contrast minimised for better visibility of layer issues late:

Webchey

So you can get a reasonable benchy out of it, meaning you can get some reasonable prints handled correctly and without much spanning. And it really does rather have a woodlike feel, moreso than the one wood PLA I've used - and I prefer how the PHA looks, too.

Which brings me to the next important test: is it actually more heat-resistant than PLA, with a higher glass point?

Spoiler: YES. It definitely is, and it's not a small difference.

I printed three very small open rectangles using the same .gcode file and weighted them on a flat surface in a reasonably-well-temperature controlled toaster oven that I have slightly modified to make more temperature-consistent. The materials are the most basic plain white/semi-translucent PLA I have (SunLu), the 20% wood PLA I have (JAYO low-temp), and the 20% wood PHA (filaments.ca REGEN). Here's my setup, at room temperature:

Plain PLA, 20% Wood PLA, 20% Wood PHA, left to right

After taking that photo, I started heating. It is a slow heat climb, due to a variety of reasons including the extra thermal ballast and insulation I added to the oven (in the case, not visible) to make it more temperature-consistent.

At 50°C, all three still maintained good geometry, as you'd expect:

50°C

But at just over 55°C, you began to see the predictable slumping of the PLA-based materials:

56-57°C, slumping in both PLA-based test objects

By 100°C, the PLA structures have rather given up. But as per manufacturer, the PHA is continuing to maintain geometry! I was surprised, particularly given the 20% wood infill, but, well, here you are:

100°C - left to right PLA, 20% wood PLA, 20% wood PHA

By 110°C, it was definitely glassing. (Really, you could see it at 105.) But it was subtle and held shape surprisingly well through 130°C when I stopped. Still: it had definitely lost structural integrity before then:

130°C, PLA, 20% wood PLA, 20% wood PHA

So the temperature claims really are legit, even in wood infill. I was, honestly, surprised. If you can get a good print with this stuff, which you should definitely be able to do if it doesn't have much spanning, it'll have much higher temperature tolerance.

As for the claims of better strength: unfortunately, I'm not really as able to test that. I did try to set up kind of a duffer's pendulum test:

this is not a good test setup, it's just what i had

My intent was to test against cylinders printed with layers perpendicular to and parallel to the length of the cylinders by holding up 1-2-3 blocks and letting them swing down into the material samples; how high I had to hold the blocks to get a break would determine relative resistance to impact.

However, due to the crudeness of my bullshit, I wasn't really able to get what I consider good comparisons. The only possible exception was in comparing the parallel-length-layers strength of 20% wood PLA vs. 20% wood PHA. In that case, the PLA clearly took less force to break than the PHA. It was the only clear result.

But it was not more than the strength needed to break the plain PLA, and it's all very inaccurate anyway. So take it all with several grains of salt.

Anyway, to sum: it's a difficult filament, and very bad at bridging, but as long as you have few bridging requirements you can get reasonable prints out of it, and these prints will be much more resistant to high temperatures than standard PLA.

And I wasn't able to validate or dispute the manufacturer's strength claims. But I was able to validate some of the other claims, which is good.

As for composting - I'm going to take my three little rectangular test objects and bury them in a flower pot outside for the winter. That's not exactly composting! We have city composting so don't maintain a compost pile of our own. But if there's any difference next summer, that would be interesting. My hypothesis/expectation is no change in the PLA, some possible de-integration of the 20% wood PLA, and as for the PHA - well, we'll see.