There are 2 parameters you need to have good control over when printing any filament:
- Melting temperature
Of these, the melting temperature is directly correlated to the chemical composition of the polymer blend in the filament while diameter control is part of the manufacturing process.
And for 3D printing, we need to take a look at the usability of the material itself. For example, pure PET is not easy to print at all and as used in bottles might be unprintable. PETG (a glycol modified PET) on the other hand is much easier to print - and most filaments sold under PET actually are PETG or PETT.
Troubles of recycling
The melting point of a blend of polymers is often difficult to gauge before doing experiments and in case of recycled material, there are problems with recreating the exact same blend when using small batches unless you use exactly one material as the base for your manufacturing. This brings us to the big problem: errors in the base material. These come in several types:
Let's address these piece by piece:
Misidentified base material
Misidentification is when you chuck material into the wrong bin and then process it as if it was the stuff the bin was for. For example, if you'd chuck a chunk of ABS into the PLA bin, your blend will not come out as PLA but as some kind of higher melting composite of the two. The exact details of the result depend on the mixture and how well you mix the processed raw material, but in effect, you just made some kind of PLA+.
This can be overcome by testing and good training as well as knowing the base material well. For example, there is an Austrian company that takes back ski-shoes. Only the hard shells of a particular manufacturer (which used a red ABS) are shredded, pelletized, mixed with some virgin ABS and color for stability and uniformity, then turned into filament, and then printed into flutes.
Another ski-shoe manufacturer takes back their own shoes and recycles the shells back into the current manufacturing, but is silent on what their shells are made from but that they are a long-chain polymer.
When trying to differentiate between PET and PETG, you can not do that unless you do a chemical analysis of every bottle - which leads to a huge problem in reprocessing: PETG melts well before PET and clumps it up, acting as a contaminant (see here for more details)!
Contaminants are a problem that comes with a bad base material. in general, there are two types of contaminants: Chemical and Physical.
Physical contaminants can be avoided by removing them before and after shredding. In the case of Skishoes (that's why I chose that example) is, you'd remove the soft shells and the metal latches, disposing of them in separate ways. Then the plastic shells are roughly sized up, cleaned, and dried before further processing. Most physical contaminants can result in partial clogging during filament production, resulting in an uneven filament. Uneven filament or such containing non-melting particles can result in print failure, for example from being stuck in the extruder or clogging of the nozzle.
Chemical contamination is arguably worse. PET bottles for example: what if the user before used to store chemicals in them that can't be separated from the polymer easily? In the best case, the contaminating raw material is removed, in the worst, it ends up in the stream. This introduced contaminated plastic ends up melting somewhat evenly into a larger portion of the recycled plastic, altering the properties in hard to predict ways. As a countermeasure in industrial PET recycling, the batches are huge and get well mixed before the new plastic product is made. By diluting the chemical contaminants on a vast batch, the effects of the contaminant are vastly reduced and evened out. This is also why even in the case of the recycled ABS-shoes-into-flutes, they mix in some degree of virgin ABS pellets - to buffer against chemical contamination.
Not all polymers are suitable for recycling and some of them alter their properties depending on their surroundings. What actually happens depends on the material in question, but let's look at PLA as one example.
While PLA doesn't exactly break down in nature unless put into a high-temperature environment, prolonged UV exposure can bleach out the contained coloration and some blends do become more brittle, others do not encounter this. Angus/Makers Muse had several prints exposed to the harsh Australian sun for up to several years and concluded the worst enemy of PLA over time is the UV light.
A different type of degradation can happen from the environment. The one side of this is cold embrittlement, which means parts become more brittle in cold. This had some experiments done on by Stefan/CNC Kitchen. The other side of this is softening, for example by sitting in a hot car. Usually, this type of degradation is not lasting but could result in embedding contaminants into the mix by embedding them in the plastic, so see there.
Is it a good idea?
Well, from an ecological standpoint, it certainly is a good idea to recycle plastic. But with all the troubles to get any good filament, will it be viable under all viewpoints? You certainly can't sell filament which is of very varying quality unless you make it dirt cheap. Also, all this machinery takes a lot of power and initial investment before you can produce your first spool - which means it might not be economical or profitable.
So, let's go back to the main question:
[Is there] a machine that can turn a plastic bottle into usable filament? [...] [Is it] currently on the market?
Yes, you can certainly extrude plastic from bottles into filament shape, and the tools are out there - for a price. However, not all bottles might be useable due to the chemical composition and you will need to make larger batches to reduce chemical contamination.
On an industrial scale, the process consists of several steps: sorting, cleaning, shredding, pelletizing, mixing, extruding, and finally spooling the filament.
Of these, the steps of shredding, pelletizing, and the combo of extruding & spooling need dedicated machinery. Even if hobby projects exist that manage to do this with well-known polymer blends, e.g. recycling 3D prints, such is usually heavy industrial machinery. In hobby-grade machinery, quality control is often problematic, as filament diameter control is the crux, and the price tag to get even filament without readjusting the machine every few minutes is high.
The Shredder might be the cheapest part, only costing several thousand euros professionally and a couple of hundred in hobby grade. A proper pelletizing machine that turns the shreds into pellets for the filament extruder has a price tag of about 10 000 € and I have not yet found a hobbyist kit. A basic inquiry on the absolute minimum investment into a professional filament manufacturing stream without pelletizer came up with about 14 000 GBP (ca. 16 800 € / 19 000 USD) plus shipping, while hobbyist kits for only one of the two seem to come up with price tags between 500 and 3000 €.
This brings the minimum investment using hobby-grade machinery to roundabout 3000 € but without a pelletizer, while an industrial setup comes out starting at about 25 000 GBP (ca. 30 000 € / 34 000 USD).
There are machines out there that turn PET bottles directly into filament by cutting them up directly before entering a filament formation system. This setup is called Pulltrusion, and it turns a plastic strip into an almost-cylindrical, folded-over filament.
While no industrial size machine of this is available, Stefan/CNC Kitchen just released a video investigating the device to manufacture such filament and then tested the print properties of such a filament. Joshua/JRT3D operated the machine in question and manufactured the samples. The base machine is the PetBot engineered by Roman Naskashev, which is commercially available for about 400 € assembled plus shipping and import taxes from Russia. Joshua also managed to re-engineer a similar machine using the same method from a 3D printer, so the price point for a self-made machine might be lower.
Each bottle weighs about 20 grams, but neither the mouth nor the bottom can be used, resulting in not 100% useability. The process also means, you can't get any deposit for the bottle back. Assuming an useable portion of about 50%, this would in Germany result in a price of 25 cents per 10 grams, so about 25 € per kilogram - which for PET filament would be quite competitive. Some bottles have larger useable portions than others, and others might not require a deposit, making these a very good price, to maybe even free filament.
Do note that the manufacturing path creates a filament that is not solid but contains a void, which is accounted for by increasing the flow multiplier, and it does require a higher temperature than PETG: with settings of 265 °C for the nozzle, 80 °C for the bed and a 130 % flow rate, 30 mm/s extrusion rate, Stefan could use an otherwise PETG profile to gain good results.
However, the higher base temperature requires an All-Metal hotend, which is part of why PET is hard to print with many machines. Other problems are the PET's crystallizing properties, which makes the melting properties at times hard to predict and can induce clogging. Also, Layer adhesion can be problematic.
The biggest problem is the tiny production size of each spool: even if one would manage 15 grams per bottle in filament, this means that one needs to change the spool 66 to 100 times more often, making larger prints nearly impossible unless one comes up with a good solution for splicing the short pieces.
While the tools are available, the price tag for a full recycling chain of raw material into filament, either as a hobby or industrially, can be kind of high. This means it might not be economical unless you can manufacture large batches and beat the price point of fresh filament.
However, with small batches and the proper tooling, it might be somewhat viable depending on bottle size and deposit system.