Computation is a powerful tool for engineering problems. Utilization and implementation of this tool was the central message of the “Evaluating Computational Modeling in Ceramic Processing” panel, held on May 2 at the Ceramics Expo in Cleveland.
“We are trying to link empirical data—which can be from the manufacturing or processing side, or even the laboratory scale—with theoretical information and fundamental physics. How do we glue all of that together so that we can make our products better?” asked moderator Krishna Rajan, a professor in the Department of Materials Design and Innovation at the University at Buffalo. “This is a huge challenge across many industries. We have to keep in mind that our interpretation of how data is analyzed is important; it has to make physical sense and it has to be able to produce materials and designs that actually work.”
I attended this panel discussion mainly to learn from Ian Dunkley, Ph.D., a Senior R&D Engineer with Medtronic Spine & Biologics, who spoke about outlining end-user issues with ceramic powder characterization. His thought-provoking presentation also included a look into the future of ceramic processing in the spine space.
I also took significant interest in another speaker’s deep dive into data analytics; I think you’ll find it of interest as well.
Here are three observations from the panel:
In ceramic manufacturing, material characterization is critical to understanding the material behavior and final performance of the part.
There are scales of varying lengths at which material behavior can be modeled, from atom to microscale, by various different methodologies, all of which fall under the umbrella of computational modeling. The challenge is turning the data provided by computational modeling into real, tactile solutions that companies can use on their manufacturing floors.
“The diversity of info [that manufacturers of ceramic parts] have to pull together is a huge challenge,” Rajan explained in his introduction to the discussion. “There are issues of uncertainty. For example, how rapidly the data changes with time is one of several metrics that must be accounted for. We need to take this kind of information and relate it to very practical problems. We can now look at properties that were previously very difficult to measure, or calculate.”
Rajan also discussed how multiscale approaches to modeling of ceramic production processes offer solutions for more accurate analysis and prediction of material behavior and performance. This is done by combining the information from different scales, which is enabled by big data, which enables the integration of information from multiple characterization and modeling datasets.
Ceramics have a long history and a bright future in orthopaedics, and challenges to overcome.
Ian Dunkley has over 10 years of experience in the design, development and manufacture of calcium phosphate ceramic scaffolds for use as bone graft substitutes in spine applications, including the past three years with the biologics team at Medtronic. “We all want a consistent final product for clinical practice. We need to start with a well-characterized ceramic powder,” he said during the panel discussion.
Resorbable ceramics have been used for many years as cements, bone void fillers and to extend autologous bone in skeletal reconstruction. Dunkley mentioned that ceramics were used in skeletal reconstruction as far back as the Napoleonic Wars in the early 1800s.
These materials, typically calcium-based salts, may be formed into porous structures that provide a scaffold for new bone formation and that are, upon completion of healing, resorbed and replaced over time by the patient’s own bone. The chemistry, microstructure and surface features of these ceramic devices influence their clinical behavior, and are dependent upon the quality and uniformity of material input and processing parameters.
Dunkley pointed to moisture and changes in purity as two significant challenges faced by ceramic manufacturers. “Final chemistry is very important to how the product performs clinically and how it resorbs so that the surgeon can select the appropriate patients for that product,” he mentioned.
After his presentation, I asked Dunkley to provide a five- to ten-year look into the future of ceramics in the spine space. “Ceramics as resorbable bone void fillers continue to grow in the spine space,” he said. “The first-generation products that are out there are probably going to become more commoditized and available to more and more geographies outside the U.S., as they are the lower-priced option for these procedures. The second generation’s products have grown quite well; we could potentially see growth of these products in the high-single-digits in the next five to ten years.”
With today’s advanced tools, we are producing vast quantities of data that the human brain struggles to sort through. But this data can ultimately be used to make better devices.
This was one of the points made by Jeffrey M. Rickman, Ph.D., Senior Technical Expert with GrainBound. I had no knowledge of GrainBound before this panel, but after listening to Rickman’s presentation, which I found to be fascinating, I did some reading up.
GrainBound is an R&D company that “provides solutions for reliable processing and predictable performance of advanced ceramics, metals and composites by applying a new materials diagnostics approach,” according to its website. GrainBound comes from “grain boundary,” which refers to the interface between two grains, or crystallites, in a polycrystalline material. Grain boundaries are 2D defects in the crystal structure and tend to decrease the electrical and thermal conductivity of the material.
Interesting, yes, but how do grain boundaries factor into manufacturing? By eliminating the grain boundaries or reducing the grain sizes down to sub-micrometer or even nanometers, high-performance ceramic medical devices can be made with improved performance. Rickman explained the manner in which data analytics can be used to highlight important processing/property correlations in manufacturing. Knowledge of these correlations can inform subsequent processing and property optimization, thereby improving performance.
Thousands of grain boundaries exist in a single part, but only a few govern overall performance. What I learned is that GrainBound combines advanced characterization tools with grain boundary “informatics,” a statistical data analysis method, to identify key grain boundaries and critical processing variables that influence them. This method can help companies that may experience issues such as inconsistent batch-to-batch performance.
As use of data (and our understanding of it) continues to expand, it will be compelling to observe in years to come how orthopaedic companies leverage it in their device manufacturing processes, toward the ultimate goal of improved patient outcomes.