Porosity is often thought to be a major concern in metal additive manufacturing, but current levels of porosity are not like they were fifteen years ago. Most additive parts we typically see today are at least 99.5% dense, and in many cases, 0.5% porosity does not have an influence on the properties of the part. 

We caught up with Dan Frydryk, an advanced lead engineer on the materials and processes team at GE Additive, to see what role porosity plays in additive parts, how we can characterize it, and what the protocol is, if a part suffers from adverse porosity issues.

What role does porosity play in metal additive manufacturing?

At a high level, porosity in additive manufacturing can be defined as any space where there isn’t metal when there should be. We typically categorize porosity as either lack of fusion porosity or gas porosity, including sub-category gas porosities such as process-induced keyhole porosity.

Once we realize there are discontinuities in the material, then we can really start to look at how the shape and distribution of the pores could affect the properties of the additive part. 

Depending on the customer’s part requirements, we can characterize these defects using various methods to make an informed decision about the level of risk they impose and provide strategies on how to abate that risk.

How does porosity vary across beam-based additive technologies?

Both laser (L-PBF) and electron beam (EB-PBF) methods have similar working mechanisms, and we can see similar types of porosity such as keyholing, lack of fusion and gas porosity. Both modalities may show porosity which are generated by the parameter set or incoming feedstock. 

The two beam-based methods differ slightly in the ability to remove the pores:

In EB-PBF, the pore is empty – literally a vacuum – and the spherical pores can be closed nicely by a hot isostatic press (HIP) post-processing step, using temperature and pressure. 

On the other hand, laser additive methods have gas-filled pores. You can close the gas pores in laser operations as well with a HIP cycle, but this requires forcing the gas that was previously in the pore into the bulk of the metal.

And what about binder jet?

Binder jet has a completely different architecture and an additional media to consider. By adding a binder into the mix, we enable the consolidation to take place outside of the machine in the form of a sinter step. So, any residual porosity in binder jet methods comes from that sinter step. 

For example, if you sinter incompletely, inadequate amounts of sinter necking and diffusion takes place and you don’t get a totally solid material. On the other hand, if you over-sinter the part, pores filled with vapor or other gases can appear and expand. And, if you sinter for too long, or at too high temperatures, then the sizes of the pores can become exaggerated. 

There’s a happy medium with binder jet, in that, you sinter for long enough to completely solidify the part but not for so long that you get large pores.

What are some of the common misconceptions you often encounter when porosity comes up in conversation?

The first thing we sometimes get asked is, "have you fixed the porosity problem?", to which we typically answer, "what porosity problem?" 

15 years ago, the porosity of PBF parts was typically around 5-8%, but now we won’t release a parameter set or theme for our machines unless it’s over 99% dense, and over 99.5% dense in many cases. So, now we’re only seeing 0.5% porosity levels, which is on par with forged materials. It turns out that many customers’ estimates on porosity are often off by an order of magnitude.

Secondly, some additive users feel any porosity is significantly detrimental and must be eliminated through HIP post-processing. However, our experience is that it is much more important to characterize the mechanical properties as a function of the entire additive manufacturing process, including porosity, and work towards accurately understanding the property requirements for the application. 

There is no need to HIP the components if the properties are acceptable for the performance of the part - as piecewise cost increases with additional post-processing steps. Moreover, if we’re leaving an as-printed surface on the part, that as-printed surface is going to drive material behavior much more than 0.1% porosity might.

What are the main ways you can detect porosity?

The most common and simplest method to measure porosity is through metallographic examination. A sample of the material is cut, mounted, ground, and polished to a mirror-like finish. 

Typically, using a light optical microscope, the polished surface is then viewed in bright-field where porosity is revealed as dark regions and dense metal will appear light and reflective. Archimedes density is also sometimes used to measure the density or relative density of solid parts with complex geometries and provides a useful non-destructive method. This entails measuring the ‘dry’ weight (in air) and the ‘wet’ weight (submerged in water) to determine the true density. 

Because any sort of trapped air (inside the part or bubbles in crevices on the surface) can lead to significant measurement errors, this is best used when porosity levels are expected to be rather high (> 1-2%) and not always applicable for AM due the low as-built porosity levels. 

Other techniques can include computer tomography (CT) and X-Ray scanning as non-destructive methods, but the current resolution is limited to relatively large defects that are not as commonly observed in stable additive manufacturing processes.

Porosity in metal additive manufacturing 2
What does porosity do to the properties of a part?

It depends. For tensile properties, we see its presence reflected in elongation first, so much so that the troubleshooting process is almost formulaic. If elongation is low, immediately check for porosity. If the porosity is excessive, you may also see a debit in the strength measurements. 

Fatigue is more difficult because the tests are dynamic, and for printed surfaces, the presence of porosity might not be obvious from the fatigue results alone. For components with as-printed surfaces, fatigue life will be limited by the surface condition, not internal features (i.e. porosity). On the other hand, in components with fully machined surfaces, pores can act as crack initiation sites and accelerate crack growth. Thus, machined surface fatigue life is more likely to be negatively impacted by porosity than as-printed surface fatigue life.

You also need to consider the shape and location of the pores as well. Spherical pores created by gas are much less detrimental than a lack of fusion between layers - as it is much easier to separate layers that have not fused properly. 

For the location of pores, patterned porosity is the enemy. If the pores are stacked in a row, then a crack is going to unzip those pores and failure will happen much quicker than if the pores were distributed randomly.

Does form factor have an influence on porosity?

Absolutely. With L-PBF and EB-PBF, thin walls and ultra-thin walls require extra care and attention due heat dissipation and how that can change during a build as a function of geometry. 

It is really important to characterize any critical thin walls in the application because, firstly, lack of fusion porosity is most likely to occur in the thick-to-thin transition regions, and secondly, thin walls surrounded by loose powder really limit the ability for heat to be conducted efficiently out of the part and these poor thermal pathways can lead to heat build-up and result in overmelting and gas porosity.

Another shape-related factor to consider are overhanging surfaces, or "downskins." Any time a beam is melting powder underneath which there is unmelted powder, as compared to powder-over-part, we risk creating porosity. 

Therefore, any porosity measurements performed need to be representative of the part that’s being build. For example, there’s little point in measuring the porosity of a 1cm cubed cube of metal if we’re making a heat exchanger which has thin walls and no bulky sections.
 
There’s still work ongoing around standards and specifications on porosity in metal additive manufacturing.

Porosity in metal additive manufacturing 1

Where are we today, and what can we expect in the short to mid-term?

Today, there isn’t a specification for standardized porosity measurement in metal additive. A lot of customers are writing their own or are trying to use pre-existing standards which weren’t written for additive manufacturing. Companies are using steel cleanliness specifications and adapting them to additive where certain methodological portions are being extracted and applied. 

For example, inspection frequency and the specified fields of view have been taken from steel inclusion specifications and adapted for the presence of pores in additive. Some porosity standards from the casting world have also been adapted but these are less applicable because the pores in castings are orders of magnitude larger than in additive. 

As of today, it is up each additive manufacturer to characterize its own material for its application and determine its own limits, working alongside any regulatory bodies that may need to certify the part and process.

In the near term we can expect to see specs and standards related to "what" and "how much." That is, what does gas porosity look like, how much lack of fusion do we have, what shape do individual pores take, and how much contour porosity exists. 

What’s further off are the measures of acceptability in the form of severity tables or similar. The additive manufacturing community needs to agree on a method, first, before using that method to quantify any debit or detriment to part performance. 

Summary

Like conventionally manufactured materials, there will always be some amount of porosity or defects in the final part and additive manufacturing is no different in this regard. But much like with traditional materials, the objective is not to eliminate every pore, but to minimize overall porosity and to characterize the material’s behavior and response to that established level, while also developing a process to flag any unexpected defect levels. 

Developers, including GE Additive, produce themes and parameter sets which regularly achieve 99.5% density in parts, a far cry from what the technology was capable of 15 years ago. Levels of additive porosity are now far less than conventional casting techniques.

There are several post-processing methods available to close out internal pores and achieve near 100% dense additive manufacturing material, each of which drives the part price higher with potentially little benefit to mechanical properties. 

Engineers need to think critically about real part requirements and how in-depth characterization of their process and material can be used to establish the best balance of microstructure, properties, and cost for a particular application.


If you would like to find out more about how pores may or may not affect the performance of the metal additive parts you use (or intend to use), then get in touch with Dan.