Material Considerations for Flexible Joint Design

Medical Plastics and Biomaterials Magazine
MPB Article Index

Originally published July 1998

In the medical industry, the innovative use of plastics is essential to the continued development of new surgical procedures. For many applications in which metals have been traditionally used, engineering thermoplastics can now offer sufficient strength, while also enabling product designers to exercise more freedom in their approach to a project. Plastic components can be fabricated in ways that are impossible for metal parts, and the properties of the resins themselves allow for the design of components not feasible using metal. This article will discuss the creation of one component of an innovative medical device—a flexible joint, the design of which was made possible through the use of plastics.

The joint design was critical to the overall performance of the device, the ETS-Flex Endoscopic Linear Cutter (Ethicon Endo-Surgery, Cincinnati). It was necessary for the joint to incorporate the required functional properties in order for the device to work, but it also had to follow good plastic design principles and be easy to manufacture. This called for an innovative concept that presented a particular challenge to the designer.

The joint needed to carry substantial loads during use of the device while maintaining the ability to move and flex—a difficult combination. The overall dimensions of the part were 20 mm in length and 11.5 mm in diameter (see photo on page 33). The design featured sections as thin as 0.64 mm (0.025 in.), and even thinner ribs or fins—down to 0.51 mm (0.020 in.). There were also hollowed-out sections that resulted in wall thicknesses of 0.51 mm (0.020 in). These design features were necessary to allow the part to flex, but made it problematic to find a material that could easily fill the part.

Design criteria were developed to quantify the type of material needed to perform successfully in this application. Because flexural fatigue of the material was the most critical design parameter and also the hardest to determine from standard ASTM protocols, a practical test was developed to measure this property and to gauge weld-line strength. The strain induced in the part during flexing was also considered. Flexural modulus was measured, since the joint needed to have a fairly low torque required to flex. The notched Izod impact test was used to judge a candidate material's notch sensitivity, and compressive, bending, and tensile tests were employed to determine column load capacity. The ability to fill the extremely thin wall sections of the design was estimated by testing melt-flow rate.

The material requirements for the flexible joint design are outlined below:

As can be seen, the joint would be subjected to significant compressive loads in addition to large strains due to flexure. In fact, this design was shown to be more than a simple flexural fatigue application. Compressive stress and strain emerged as critical variables.

The critical material properties required for this design included excellent cyclic flexural fatigue (which was dependent on good weld-line strength and low notch sensitivity), flexural modulus, and load capacity. The material also had to fill long, thin sections and survive gamma sterilization with no reduction in properties.

Five materials were initially considered: engineering thermoplastic polyurethane (ETPU), amorphous polyamide (PA), polyetherimide (PEI), polypropylene (PP), and high-density polyethylene (HDPE). Table I details physical property data for each of the materials, and Table II lists the material trade names and grades.

Soon after the design criteria were established, the PP and HDPE were eliminated from consideration. HDPE could not meet the required mechanical properties, even though it could flex. PP could provide the necessary flex, but was insufficiently strong and unable to withstand the gamma sterilization. The remaining three materials that were selected for further testing—ETPU, PA, and PEI—are all unfilled, amorphous engineering thermoplastic resins. They were chosen because of their high strength, good ductility, and ease of processing.

The first test performed was to measure the flexing properties of each material. This was the most important test, since flexing induced the highest combination of flexural stresses and induced strains.

Preliminary flexural tests were conducted on PEI and PA joints, evaluating the difference in fatigue life for both single-direction, 30° flexures and ±45° reversing flexures. The number of cycles to failure was measured, with failure defined as the material yielding (i.e., reaching the point of permanent deformation), exhibiting cracks, or actually breaking. A minimum number of cycles was defined to give a pass/fail point that factored in a safety margin of three times the number of cycles in actual use. For the sake of this discussion, this number will be normalized to 100%. The PA was also tested after being moisture conditioned, since it was thought that such conditioning would improve the material's flexing properties.

Following the testing described above, the prototype mold was redesigned to add triangular ribs on either side of the center blade. The purpose of the ribs was to cause the center blade to be a close fit in the tool, so that it would not be able to deflect during mold fill. This redesigned configuration resulted in two different outcomes, depending on resin type and processing parameters. In some instances, the center blade stayed in position and the flexing area walls stayed uniform, providing a fully filled part with equal strength on either side. Alternatively, the shape of the triangular ribs resulted in the walls of the flexing section tapering to a feather edge near the center of the part, which was harder to fill than the nominal 0.51–0.64 mm (0.020–0.025 in.) walls in the flexing region.

Once the prototype mold was redesigned (as explained above), further testing was completed on all three materials. This round of testing used a part that was closer to the actual commercialized component, and is most indicative of final results.

Table III lists detailed flexural fatigue test results, as described in the following section.

Preliminary Testing. ETPU was not considered in the preliminary testing. Initially, the compressive loading was expected to be 25 MPa (3450 psi) and the off-axis bending stress 69 MPa (9677 psi), which would have combined to exceed the mechanical properties of ETPU. Also, the induced strain rate—approaching 7.5% from flexure alone—exceeded the ETPU's 6% published value, especially when strains due to compressive and off-axis loading were taken into account. Based on this analysis, materials with higher flexural strengths and elongations were tested first, and ETPU was initially considered to be a "low-end" material.

In the preliminary flexural testing, only three samples each of PEI and PA were tested under each condition. Under single-direction loading, both materials flexed 30° without fatiguing, but under ±45° degree flexure, some PA and some PEI components did not reach the minimum number of required cycles.

The moisture-conditioned PA joints survived only 20% of the required flexure cycles (±45°), compared with 144% for the unconditioned parts.

Testing of Parts from Redesigned Mold. Following these preliminary tests, ETPU was added to the testing based on its historical success in flexure fatigue applications and the poor performance of the other two materials. In addition, the original method used to attach the joint to the rest of the instrument called for a material with high compressive and shear strength. This attachment method was modified to reduce these requirements. Using the redesigned prototype mold, parts made from all three materials were molded and tested.

The PA parts demonstrated varying performance, depending on the mold temperature and the injection pressure (all temperatures and pressures used were within recommended processing conditions). Parts molded at lower temperatures and pressures tended to fracture during the first 4–8% of the required cycles, with the fracture initiating at the weld lines between the fins. These weld lines progressed radially inward from the heavier spine sections of the part and axially inward from the fin on either side. Parts that did not exhibit these early fractures tended to flex for approximately 12–16% of the required cycles, then fracture at the junction between the flexing section and the bulkhead.

Flexing joint made from ETPU.

Cross section of flexing region. Photos: Ethicon Endo-Surgery

Six PA joints made from the redesigned prototype tool, but molded at higher temperatures and pressures, survived an average of 112% of the required ±45° flexural cycles, with values ranging from 100 to 120%. PEI joints molded at the same time fractured within 4 to 8% of the required cycles. Six ETPU joints made from the redesigned prototype mold each cycled without failure—at 400% of the required flexural cycles—before the test was discontinued.

Even though ETPU successfully passed the flexure fatigue test, a final tool modification was made in which a radiused contour was added to the center-blade side ribs. This eliminated the feather edge from the flexing section walls. With this geometry, the ETPU material routinely exhibited flex-life values of more than 400% of required reversing cycles without noticeable signs of yielding, and more than 600% of required cycles without major fracture.

Additional Testing. ETPU was the only material that successfully passed the flexing tests, as both PA and PEI did not pass the flexure fatigue testing. Based on these results, the substantially superior flexural performance of ETPU, the low weld-line strength of PEI and PA, and the apparent sensitivity of PA to processing parameters, PEI and PA were dropped from consideration.

The ETPU resin was then taken through functional testing on assembled units to ensure that it met the other loading conditions. The ETPU material successfully passed all of the additional tests.

The ETPU material exceeded performance estimates regarding its ability to surpass the maximum 6% yield strain predicted from calculations using ASTM test data. This shows that published material property data are not always the best predictor of performance in a molded component. There are two theories as to why ETPU was able to outperform expectations. One explanation is that, because the strain was occurring in the thin section, the combination of strain and wall thickness resulted in a different stress/strain behavior than can be predicted with standard tensile bars used in ASTM tests. Another possibility is that some yielding occurs in the initial flex, and subsequent cycles occur in the yielding portion of the tensile/compression curves.

Both PA and PEI have higher yield strains than ETPU, but did not perform as well. A possible explanation is that the ribs could interrupt the geometry of the part and lead to stress concentrations. This could induce local areas of higher strain and in turn cause cracking and fracture in PA and PEI, instead of the yielding observed in ETPU.

At the ends of the flexing region, there are minimal radii in the corners because of the lateral ribs (0.51 mm (0.020 in.)) connecting to thicker sections (3.2 mm (0.125 in.)). If bigger radii were used, the flexing region would be reduced. However, in the flat section, better performance was expected from PA and PEI, based on their published physical properties.

PA flows well, but is more notch sensitive than ETPU. The performance of PA was close to the minimum required flexural cycles. At failure, it fractured at the interface of the thin flexing area and the thick nonflexing bulkhead, where stress concentrations would be expected. It was anticipated that the conditioned PA parts would perform better than the unconditioned parts, but such was not observed in this case study. Even though the average performance of the unconditioned parts was above the minimum requirement, there was too much variability to satisfy the demands of the product designer.

PEI is also notch sensitive, and required higher pressure to fill the part. This caused some molding problems, as the elevated pressures resulted in core shift. Because of these processing difficulties, few "good" PEI parts were made.

Joints molded from PEI—and some from PA—fractured at the weld lines that formed between each fin. These weld lines were more evident when there were problems processing the material. The ETPU offered higher weld-line strength—a result of its chemistry and behavior during molding.¹

Based on its flex fatigue properties, excellent weld-line strength, and relative notch insensitivity, ETPU was selected for this flexible joint design. ETPU also met the other material requirements, including low flexural modulus, high column-load capacity, good flow into the part, gamma radiation stability, biocompatibility, and excellent chemical resistance. The joint design has since been successfully commercialized using engineering thermoplastic polyurethane.

1. Moses PJ, Chen AT, and Ehrlich BS, et al., "A New Unique Family of 'Live' Engineering Thermoplastics," in Society of Plastics Engineers, Inc., Technical Papers, vol XXXV (ANTEC 1989), Brookfield, CT, SPE, pp 860–865, 1989.

Karen L. Winkler is market development leader for the Dow Plastics Medical Group (Schaumburg, IL), with more than nine years' experience providing consulting on materials selection, plastics design, and processing. She currently serves on the board of directors of SPE's Medical Plastics Division. Thomas W. Huitema is a senior research and development engineer at Ethicon Endo-Surgery, a Johnson & Johnson Co. (Cincinnati, OH). Over the past 20 years, he has engineered new products for three different industries, and has obtained 19 patents, with 3 pending.

Copyright ©1998 Medical Plastics and Biomaterials