As the adoption rate of additive manufacturing (AM) in orthopedics grows, OEMs call for clarity on validation, particularly for materials and processes. Robust procedures ensure that AM is employed and advanced in a way that ultimately won’t harm patients, and the industry as a whole.
With this in mind, we posed validation-related questions to additive experts from four key players. Topics included validation of parts made on different locations on a build plate, process challenge parameters and safety hazards of additive materials. Finally, we asked our experts to provide actionable advice for OEMs.
Note that all experts interviewed recommend FDA’s guidance document, “Technical Considerations for Additive Manufactured Medical Devices.” Finalized in December 2017, the guidance contains FDA’s thoughts on phases of design development, production process, process validation, semi-finished and final finished device testing.
3D Systems | Kim Torluemke, Vice President, Quality & Regulatory, Healthcare; Joseph Ruppert, Director, Quality Assurance Operations
Carpenter Technology | Richard Grylls, Ph.D., Director, Applications Engineering, Carpenter Additive
FMI Instrumed | Jelle ten Kate, Project Engineer, Additive Manufacturing
TRUMPF | Lars Neumann, Ph.D., Industry Manager, Medical
What is the best strategy for validating parts made throughout different locations on the build plate?
Torluemke/Ruppert: There is no specific requirement to validate the entire extent of the build plate. The key is to follow a risk-based approach, beginning with a risk analysis to understand key inputs of variance that may need to be assessed, including ones that may be affected by build locations.
Grylls: There’s no replacement here for good data—understand your equipment, how it works, the key variables, how those key variables can vary, how to calibrate equipment and how often, how to check on gas flow speed, laser power, humidity and so on. Then evaluate the machine and run tests with the right challenge parameters to determine if all locations in the machine build equally well, or if some locations give marginal properties. Also, I advise something that’s never very popular, but actually works: If there is a corner of your machine where the properties are more variable, consider not building in that area. Stick to the areas of the plate that work.
ten Kate: It is hard to say if there is one best strategy that simply covers all. It depends on many variables and requires a lot of test builds. We’ve done it partly by experience, but mainly by our process qualification. It also depends upon the kind of products you intend to produce—orthopedic implants will be different than industrial parts, for example.
Neumann: The industry has significantly improved homogeneity across the build plate in recent years, for example by optimization of the shielding gas flow. Nevertheless, users must still evaluate and monitor the remaining inhomogeneity. Typically, build area and height of interest are selected first. Various specimen, e.g. for tensile and fatigue testing, density, etc., are placed in this build volume such that they fully characterize the entire volume. During production, a limited number of testing specimens are built together with the regular parts to detect any deviation from the specified process. Type, location and number of testing specimens depend upon acceptable levels of deviation.
What is the best method to determine process challenge parameters for Operational Qualification runs when there is the potential to mix different parts on the same build plate during production runs?
Torluemke/Ruppert: Again, we start with a risk-based approach to define a validation plan. A series of experimental studies can characterize the process if interactions and worst case of process parameters are unknown. Within pre-validation work, statistical techniques such as factorial designs and analysis of variance may be used to better understand the influence of various process parameters and the potential limits of the process that may be challenged within the operational qualifications.
Grylls: If you have different parts on the same build plate, you have to verify that for any reasonable collection of those, building one part does not affect the mechanical properties of the other parts. If you’re making two different products, you’d often just make one on one build plate and a second on a separate run. Or, if you did mix parts on the same build plate, you wouldn’t allow arbitrary numbers of each. For example, you would qualify building a specific mix of parts but not any arbitrary mix—so processes are frozen, and changes in the mix of parts can’t be made during production. Aside from the concern about mixing parts, metal AM has a lot of key process variables—too many! Over time, machine tool builders and process experts have become better and better at identifying all the key process variables and controlling them. For example, even three years ago, we didn’t really understand that the humidity inside of the printer room is a key process variable. Too much humidity can get inside the printer, causing the powder to not spread as well as it should. This affects whether you can even build a part, let alone determine its actual properties. A good machine tool builder will have a humidity spec for the room where the printer is kept. A good practitioner will have a humidity meter on the wall and will stop printing if it goes out of spec. The more you can take control of your process variables, the less you have to challenge.
ten Kate: Although we try to mix as few different parts as possible and sometimes not at all, it is a benefit of AM to produce multiple different parts in one build. We always validate this during our operational qualification. Through our years of experience in AM, we have a solid understanding of which products and process challenge parameters have an influence on each other.
Neumann: Since we manufacture all components on a single build plate in parallel and not sequentially, the object of qualification of the AM process is the build job, not the individual component within a build job. This differs from other production technologies. As long as we choose and fix a particular build job, it is irrelevant for the qualification process whether it consists of multiple identical components or a mix of different components, e.g. an acetabular cup and a hip stem. We always manufacture the same build job, again and again. In this scenario, we qualify this build job and remain well within standard regulatory procedures. A different scenario is mass customization. Personalized implants fall into a different regulatory regime than standardized implants, and standard qualification processes are not applicable anymore. From a manufacturing point of view, grouping similar components with respect to relevant manufacturing dimensions is a good starting point to yield a stable manufacturing process, especially with respect to volume, height and minimal feature sizes.
Fine particles of titanium and other alloys present certain health and safety hazards. What are the considerations when installing AM equipment and handling the powder?
Torluemke/Ruppert: The first thing to do is conduct a thorough risk analysis of materials used, paying close attention to particle size. Site preparation instructions are provided by most printer manufacturers and include elements such as environmental condition requirements, fire suppression systems, oxygen monitoring (if using shielding gas) and exposure limits for operators.
Grylls: I categorize this into three areas: 1. Immediate danger: Could the titanium powder combust? 2. Medium-term danger: These are things that could happen over time while the AM equipment is operating. For example, when using powder bed fusion, the machine creates soot that accumulates behind walls and in pipes. Over time—after several months—this can cause a problem. 3. Long-term danger: For example, this considers the danger of breathing in fine particles over many years. The bottom line for this is that the use of proper personal protection equipment is critical to avoid breathing fine titanium powder. Every manufacturer should develop and follow safety policies and procedures for handling all materials, including the feedstock and all the byproducts. Note that UL has published a nice guide that details how to use, install and run AM equipment safely and how to consider these hazards.
ten Kate: First, make sure the AM equipment is placed in the right environment, in our case a dedicated department apart from the other machining equipment. The environment has to fulfil the requirements regarding temperature and humidity. In case something goes wrong, you have to ensure that you have the suitable material to stop the hazard. For this, you need the right fire extinguisher and eyewash station. To protect our employees, we make sure they always wear protection shoes, an overall, suitable gloves and the correct gas mask. We also do scheduled, regular checks on these items.
Neumann: Safety data sheets should be the first source of information to evaluate health and safety risks for a particular powder. While AM powders seem novel, metallic powder is well-known as a by-product of many metal processing technologies. Regulatory bodies have long classified health and safety hazards of metallic dust. In general, the best strategy to address these hazards is to process metal powders within an enclosed environment, minimizing interaction among powder, operator and outside environment. As with any emerging technology, AM machines were initially designed for use in public and corporate R&D facilities. Here, their primary, revolutionary purpose was to turn previously impossible designs into a reality, rather than providing a work environment in accordance with the requirements of a production plant. As a consequence, the design of such systems made frequent, direct contact among powder, operator and outside environment necessary. As AM transitions into a production technology, machine designs evolve to address health and safety hazards. The latest machine generations typically allow to refill, store and recycle powder as well as separating build and used powder within a fully enclosed system.
What three recommendations would you give OEMs as they’re considering proper material validation?
Torluemke/Ruppert: 1. Rely on industry knowledge. 2. Understand—and seek expert advice on—the FDA guidance document. 3. Become aware of new additive standards via ISO, ASTM and other standardization bodies.
Grylls: 1. Get help. 2. Get help. 3. Get help! The first places to get help are the FDA guidance, ISO 13485, ASTM, UL and other industry guidelines and standards. The next is from the machine tool builder: They can help you understand key process variables and ensure safe operation. And finally, from materials and processing experts, like us. Carpenter can help manufacturers choose the best materials, develop process parameters, validate and test properties, all the way to full production implementation. This can save OEMs much cost, resource time, and pain—allowing them to focus their efforts on the highest value-added activity for them: designing better products for their customers.
ten Kate: 1. Although we all use the same powder, we can still make parts with different material properties. OEMs have to consider that every company has developed its own strategy and together you have to see if this strategy suits both parties. 2. Have a good understanding of what exactly is changing by reusing powder several times or learn how to cope with different materials. For example, think about whether the material conditions are still within specifications. 3. Make sure that you use AM as an extra production technique; do not make AM a goal. Rather, see it as another tool in the toolbox to produce your part. Designing for AM provides so many advantages versus just copying a design.
Neumann: 1. Progress in quality, cost structure and process capability will make your next AM device easier to validate. 2. Expanding regulatory framework and standards provide more and more guidance and clarity. 3. Process monitoring provides new data sources to detect variations.
Image courtesy of 3D Systems