Lower-Cost Metal 3D Printing

Metal 3D printing conjures up an image of lasers depositing energy into explosive powders, and systems that cost millions of dollars. However, there are viable options that will allow you to print metal parts on a consumer-level filament 3D printer. The investments start at a couple thousand dollars and go up a couple orders of magnitude from there. Let’s find out what it takes to create metal 3D-printed parts with garage-level maker tech.

Like MIM — But With Filament

The roots of this technology lie in metal-injection molding (MIM). MIM parts are created from a feedstock made of metal powder and a binder. The feedstock is typically 80% or higher metal by weight, but, since the metal is so much denser than the binder (maybe a factor of 7 or 8), the split by volume is closer to equal.

The mix is injection-molded to create a green part. Next is debinding: the binder is baked out or removed chemically to create what is called a brown part. Finally, the brown part is sintered at high temperature to form a metal part. After this process, the part will shrink 15% or so, typically more in the vertical direction due to gravity, and a bit less in the horizontal plane.

If we have an imaginary cube of the metal-binder mix that is 60% metal by volume (Figure A), the relative volumes of metal (gray) left after sintering and binder (blue) would be as shown. The ratio of the sides of these three cubes is 1.00: 0.84: 0.74.

Now imagine that instead of injection-molding the part, we made this metal-binder mixture into 3D printer filament (Figure B). The parallels between injection-molding a plastic part versus 3D printing a plastic part are pretty much the same for MIM versus extrusion 3D printing with high-metal-content filament. That is, the filament system avoids the need for creating molds, and makes it cost effective to make a single part. (Note that these high-metal filaments are different from filaments with less metal powder that create parts that only look like metal, but have properties closer to plastic.)

At a minimum, a system to 3D print high-metal-content filament consists of:

• a 3D printer
• a debinding method to remove the plastic binder — some companies rely on a chemical debinder, while others use heat
• a furnace to sinter the metal particles together.

MIM is a mature technology that has been around for a while, and those techniques (and materials standards) gave a good jumpstart to companies trying to create consumer-level high-metal-content filament and metal extrusion printing workflows. What you need to try this at home varies, depending on which company’s systems you decide to try out. Since binders and debinding and sintering processes are designed to work together, you’ll need to buy into one end-to-end process or another.

High-metal-content filament is abrasive. Special abrasion-resistant nozzles made with materials like hardened steel, ruby, or sintered diamond are needed to print with these materials. A standard bronze nozzle will just be abraded away in very short order. That cost should be included in your budget for metal 3D printing. Some 3D printer manufacturers have add-on kits specifically for printing metal, with appropriate nozzle and sometimes bed modifications.

The Virtual Foundry

The Virtual Foundry focuses on users wanting to print metal parts at a minimum cost. Their filaments (branded Filamet) are designed for heat debinding, so a controllable furnace can do both steps. In addition to a hardened steel nozzle, the company also recommends a device to warm the filament (a Filawarmer) as it comes off the spool. This makes the filament more pliable and less likely to break.

Once you’ve set up your printer, you’ll need to print the part bigger in all three dimensions, by scaling it up in your slicing software or other means. Each material and process will require different scaling factors. The resulting green part (Figure C, left) can be sanded or otherwise cleaned up before the debinding process. A microscope image of a copper part surface at this green stage, showing how inhomogeneous the material is, can be seen in Figure D.

This green part is packed in a crucible with aluminum oxide Al2O3 (Figure E, below), which will support the print as the binder is baked out. The aluminum oxide is like beach sand and adds some stiffness to support the part. The crucible and kiln shown in Figure F are the ones listed in the price list later in this article. This kiln and can run on regular 120V wall power.

The result is the brown part (Figure C, center). Brown parts have had much of the binder removed, and can be fairly fragile. Brown parts in copper and bronze are about the consistency of a slightly stale store-bought cookie.

Finally the part is placed in a kiln, packed in talc to support it, heated to sinter together the metal, and then polished if desired (Figure C, right). Most metals also require carbon added to the talc to avoid oxidation, and other processes.

The finished part on the right was tumbled in a standard tumbler with stainless steel media, then polished with a brass wire wheel. The shrinkage (mostly during sintering) is clear in the series of images.

Including a typical filament-based printer, the costs of a first part and second part are shown in Table 1. The amount of filament (in grams and meters), and resulting costs for the Benchy print shown in this article are: bronze 76g, 7.37m, $16.36; copper 83g, 7.37m, $12.99. The small crucible listed here (and shown in Figure F) can hold a part of diameter 43mm or less and height 58mm or less. Depending on when you read this, of course, prices can vary. You can wait for everything to cool to room temperature before touching it, but having additional safety equipment like forge gloves is never a bad idea.

Copper and bronze are the easiest metals to start with, since they have fewer oxidation issues during sintering. Steels require sintering carbon added to the sintering mix to absorb oxygen. Other materials sold by The Virtual Foundry are not really intended for desktop users, since they are reactive and need special handling, or hotter furnaces than the introductory one mentioned here. In some cases, users stop with a green part, just to make a part that is very heavy. In those cases, the part is just printed and used as-is, binder and all.

The final metal properties, such as how dense the final parts are, are acutely dependent on how carefully the debinding and sintering processes are managed. Density depends on the length of time the parts are sintered, among other factors.

The Virtual Foundry has extensive directions on their site for particular materials, and a Discord community for users to share experimental data and tips. If debinding and sintering feels like more than you want to deal with, their partner Sapphire3D provides those steps as a service, with prices based on material and size of the part.

Zetamix

French company Nanoe produces a filament line called Zetamix, two versions of which are 316L and H13 steel. Their materials are more aimed at a small industrial user who wants to make tooling in-house with fast turnaround, like these parts printed in 316-L stainless steel (Figure G).

Nanoe separates their debinding and sintering steps. Chemical debinding occurs in an acetone bath. Sintering is done in a furnace that uses gas like argon with a small amount of hydrogen, a common welding gas mixture. Their sintering furnace is designed to run at high temperatures with the part in this gas. As a result, this is a pricier system, with entry-level costs around $10,000.

In this process, brown parts retain some of their binder. Nanoe’s COE, Guillaume de Calan, likes to make the analogy of building sand castles. Removing all of the binder would give you a pile of sand, but if some of the binder remains, the castle would still be damp and hold together. Sintering takes place in a gas environment, so supports may be needed through that stage.

Prices for 3D-printed parts are always very dependent on the details. Having said that, Nanoe gave us the example of a small lubrication nozzle that has to spray a very high-pressure fluid on machined parts. The part was just not possible to machine, due to its internal channel. It is approximately 5cm high and weighs 50g.

Cost for this print for a European customer would be around €25 ($27) of material, and around €120 per cycle run (€70 of electricity, argon gas, and consumables, and about €50 of furnace depreciation). You can put a few parts in a furnace run, so depending on the number of parts in the run, total cost per part would be lower — in the range of €55 for four parts per run, for example. Outsourced sintering of the same part was estimated at about €85. Sintering partners are listed on the Zetamix resellers page.

BASF Ultrafuse

If you don’t want to wade into debinding and sintering yourself, chemical company BASF has created Ultrafuse 316L and 17-4 PH stainless steel filaments. The chemical debinding and sintering they require are industrial processes that are not feasible in a home environment. Because the sintering process takes place in a gas environment, supports would need to stay on through sintering for this option, and be removed with metalworking tools at the end of the process. (The provider just says parts cannot have overhangs for this reason.)

As with the other options in this article, at a minimum, you’ll need a hardened steel nozzle for your printer. Once these modifications are made, though, at around $148 for a kilogram of filament plus $50 to sinter parts made from it, this might be the cheapest option. However, your control over the process is limited.

To debind and sinter the parts, users buy coupons (currently $50 for up to 1kg of parts) and ship them out to an industrial partner who runs a batch every couple of weeks. The maximum part size is 100mm on a side; other design constraints can be found on the Ultrafuse product pages at forward-am.com.

Other Considerations

The bottom line is: What system makes the most sense for you depends on how often you want to make a print; materials you are interested in; and your patience with tuning a process with a lot of variables. If you have a very short-term project, then outsourcing the debinding and sintering might make sense. If you have more challenging needs (and perhaps experience dealing with welding gases) you might want to invest in a more sophisticated furnace.

The filament is heavy: three kilograms of bronze filament from The Virtual Foundry is about the same length of filament as one kilogram of PLA. A 3kg spool might snap off your spool holder if it’s flimsy.

Supports needed for printing are typically removed at the green part stage, when metalworking tools are not required. However, some thought might be required about the best orientation for sintering, which might not be the same as the orientation for printing. Be sure you understand the details of how the part is supported during the sintering process for the particular workflow you choose. Parts may warp during sintering if this is not taken into account. If high dimensional accuracy is essential, you may need to step up to a more pricey system.

As with all 3D printing, slicing settings and preparation of the print bed are critical for good results, and you should consult the documentation for any manufacturer’s product (and the rationale, which can be a little counterintuitive for those of us used to plastic).

Next Steps

The next step up in sophistication and price is one of the larger systems that either uses some sort of feedstock with a binder mixed with powder, or sprays a binder on a powder. For example, Desktop Metal’s Studio System (Figure H) uses metal feedstock rods instead of filament; their Shop System uses a jetted binder. Markforged uses bound powder filament in their MetalX system. (In the interest of full disclosure, we did a curriculum-development job for Desktop Metal last year.)

These integrated systems come with software tools to make it easier and more predictable to create metal prints, but of course at higher cost than the more-DIY versions above. (See SJ Jones’ starter guide.)

Alternatives

A final question is whether your project actually needs to be printed in metal. There is a wide range of materials between PLA and stainless steel, and, depending on the real requirements, you might have better options.

If the part needs to be strong and stiff, a composite material (like continuous or chopped carbon fiber) might fit the bill. A composite part is likely to be a lot lighter than a metal one as well. If heat resistance is an issue, there are ceramic 3D-printable materials that are sintered similar to the process for metal. However ceramic debinding and sintering can be an easier process than the metal equivalents.

If you’re interested in learning more, we have written a course on LinkedIn Learning. Check out “Additive Manufacturing: Metal 3D Printing” 

This article appeared in Make: Volume 86.