Originally published May1997
TECHNICAL PAPER SERIES
For medical device manufacturers, ensuring the reliability and durability of critical health-care products is perhaps their most important concern. Since many medical devices are made of plastics, in-use failures can and do occur. One of the major causes of polymer device failure is cracking of the plastic, either during use or as a result of transportation and storage. To prevent catastrophic failure of plastic medical devices, it is very important to understand the limitations of each polymeric material. It is not easy to find a supertough polymer at a reasonable price that can meet all performance requirements. Some materials, such as specialty nylons and toughened nylons, offer high toughness and good chemical resistance, but provide less than ideal dimensional stability, clarity, and sensitivity to moisture. Many polyolefins present similar examples of property trade-offs.
One class of materials that provide a practical balance among a variety of properties are rigid thermoplastic polyurethanes (RTPUs), which demonstrate a useful combination of toughness, clarity, fatigue and wear resistance, chemical resistance, and dimensional stability. Another unique feature of RTPU materials is that they are "live" polymers: during conventional injection molding, the high-molecular-weight RTPU macromolecules controllably depolymerize in the molten state and repolymerize during cooling to rebuild molecular weight. This behavior may have an impact on the integrity of injection-molded medical devices made of RTPU.
This article examines the way in which environmental-stress-crack resistance (ESCR) and processing-condition sensitivity relate to various mechanical properties of medical devices made from glassy RTPU resin. Two different grades of RTPU were chosen for the study. The effects of chemical exposure and heat-aging history on product functional performance and potential failure modes were also investigated.
Materials. Two rigid polyurethane materials--designated RTPU-A and RTPU-B--were provided by a supplier. Table I lists selected physical properties of the two resins as published in the technical literature. According to these data, both grades are high-molecular-weight, predominantly hard-segment polyurethanes. They are formed via an addition polymerization process from a difunctional isocyanate (methylene diphenyl diisocyanate) and a difunctional glycol (1,6-hexanediol). RTPU-B contains cyclohexanedimethanol, which offers higher stiffness and heat resistance (see Table I). In addition, RTPU-B has a very small amount of polyether (polytetramethylene glycol), whereas RTPU-A does not--a minor difference in chemical composition that is not detectable by FTIR-ATR spectral analysis.1 Because little information is available about the sequential structure of the connections between difunctional isocyanate rigid segment and difunctional hydroxy chain extender, the morphological effects of RTPU composition on physical properties could not easily be predicted. One objective of the current study was to determine whether the two RTPU materials employed would show significant differences in ESCR behavior against different chemicals.
Table I. Physical properties of RTPU materials.
Sample Preparation. One group of ASTM (D 638) tensile bars of RTPU-A and RTPU-B were supplied by the resin manufacturer. A second group of bars were injection molded in-house according to the manufacturer's processing guide, while a third and final group were molded under different mold temperatures in order to study the effect of processing conditions on mechanical properties of the two RTPU materials. Injection molding parameters for this third group included resin drying at 110°C in a hopper drier for at least 12 hours; melt-temperature settings of 249°, 232°, 227°, and 221°C; and mold temperatures of 13°, 24°, 41°, and 60°C. None of the tensile bar specimens used in the study underwent heat annealing.
Mechanical Testing. Tensile tests were performed on a Model 8501 universal testing machine (Instron Corp., Canton, MA) at a crosshead speed of 2 in./min, a grip distance of 4 in., and a gauge length of 2 in. A sample size of at least five was used for all tensile tests, unless otherwise noted. A Series IX automatic materials-testing system was used to calculate elongation at break and other mechanical properties. Tests were conducted in an environmentally controlled laboratory at a temperature of 22°C and a relative humidity of 50%. Test specimens were kept at ambient conditions for at least 1 week prior to testing.
ESCR Study--Time to Craze. An ESCR study was performed on metal fixtures with seven different strain levels: 0.312, 0.631, 0.946, 1.26, 1.57, 1.89, and 2.24%. Tensile bars were used for this study. The modulus of 1900 MPa was used to calculate stress level for RTPU-A, and 2100 MPa was used for RTPU-B. Isopropyl alcohol (IPA) was selected as the stress-cracking agent. Time to craze was recorded as visual observation of the start time of crazing.
ESCR Study--Elongation Retention. Samples for analyzing elongation retention were prepared as follows. Cylindrical metal fixtures with varying diameters were used to generate different strain levels of 0.3, 0.6, and 4.0%. Tensile bars were mounted on fixtures for 3 minutes, during which chemical solutions were applied on the surfaces of bended bars by either constant swabbing or immersion. The tensile bars were then washed with water and dried for a few hours before tensile tests. IPA, cyclohexanone (cyclo), methyl ethyl ketone (MEK), and a cyclo/MEK mixture were used as cracking agents. Elongation at break was generated by mechanical tests.
Figure 1. Schematic illustration of a typical joint geometry for a medical device. The joint is designed to undergo bending during use.
Preparation of Joint Assembly. Figure 1 shows the joint geometry of typical medical devices used in this work. The male parts were injection molded from the RTPU-A resin, and the female parts were made of RTPU-A or another clear engineering plastic such as polycarbonate. The joint area is designed to undergo bending during use of the device. The joints were manually bonded together by different cyclohexanone solutions under a controlled assembling force. Assembled parts were dried for a few hours and then heat-aged at 57°C for 1 week before being submitted to bending tests. Bending tests were performed with an Instron 8501 with holding fixtures. Joint integrity was characterized by bending torque.
RESULTS AND DISCUSSION
Processing Sensitivity. As discussed previously, these "living" RTPU materials depolymerize at high temperature (molten state) and repolymerize during the cooling process. Therefore, the molecular weight and thus the physical properties of the final products may be sensitive to the processing conditions. It is well known that the toughness of polymeric material depends greatly on molecular weight. The transition between a ductile and a brittle RTPU occurs at a molecular weight of about 100,000. In addition to influencing wall thickness and the level of impurities, the mold temperature might play an important role in the repolymerization process since the degree of polymerization is dependent on cooling rate.
Figure 2. Elongation at break of RTPU-B injection-molded samples at different mold temperatures. Great sensitivity of bulk mechanical properties (elongation at break) to molding conditions was observed.
In this study, the two RTPU materials were injection molded into ASTM tensile-bar form (3.175 µm thick) at four different mold temperatures. As shown in Figure 2, the elongation at break of RTPU-B is very sensitive to the molding condition. All samples except the supplier-provided ones were injection molded at mold temperatures lower than the supplier-suggested temperature. The scattering data of elongation at break of the in-house-molded samples might be a strong indication of unfavorable mold temperature, given that the supplier-molded samples show much better results. However, the elongation data of RTPU-A (see Figure 3) show that the elongation at break of RTPU-A is not as sensitive to mold temperature as is RTPU-B. In fact, within the mold-temperature range of 13° 60°C, no significant effect was noticed.
Figure 3. Elongation at break of RTPU-A injection-molded samples at different mold temperatures. No significant effects of molding condition on bulk mechanical properties (elongation at break) were noted.
As noted, there is only a slight difference in chemical composition between the two RTPU materials. RTPU-A has no polyether segment and is derived from 1,6-hexanediol, whereas RTPU-B has a very small amount of polyether and is made of 1,6-hexanediol and cyclohexanedimethanol. The striking difference in processing-condition sensitivity between the two materials may be partially caused by the minor chemical composition variations and potential morphological changes responding to the cooling processes. The inconsistency of elongation at break from specimen to specimen of RTPU-B may be attributed to variations arising during the repolymerization process because of a higher polyether or cyclic-diol content compared with RTPU-A, which shows much less sensitivity to mold temperature. As explained in following sections, these slight differences in chemical composition may also significantly affect the chemical resistance of the RTPU materials.
ESCR Performance. We have seen that the sensitivity of bulk mechanical properties to mold temperature showed an apparent discrepancy between RTPU-A and RTPU-B. In the following discussion, we will investigate the effects of injection molding conditions, chemical agents, and time of exposure to those agents on ESCR performance of the two RTPU materials.
Environmental stress cracking is one of the most common causes of failure in plastic devices. Key factors that may influence the environmental-stress-cracking behavior of plastic medical devices include (1) the complexity of the stress, (2) the presence of surface-active agents, (3) surface morphology and orientation, (4) exposure time, (5) temperature, (6) the material's morphological structure and chemical nature, (7) part design and geometry, and (8) processing conditions.
The phenomenon of environmental stress cracking of plastic materials has long been recognized, with its mechanisms and test methods often investigated in the literature.29 ESCR is also a very well established topic in the field of inorganic materials such as glass and metals. More than 75 years ago, Griffith developed a classic theory to account for stress failure and crack propagation of metal materials.10 With certain modifications, his formula is still useful for explaining ESCR behavior in polymers:
where T = tensile strength, E = Young's modulus, * = surface fracture energy, * = Poisson's ratio, and c = length of surface defect. With a reduction in surface energy, the critical length to failure of a surface defect will decrease, as was confirmed by Rebinder et al.1113 However, Griffith's equation is limited to uniaxial applications and does not contain any term for time-dependence--a very important and unique factor for viscoelastic polymeric materials.
In the late 1940s and early 1950s, Howard started to work on the ESCR phenomena of semicrystalline polyolefin materials.2,14 Consideration of the unique properties and polyaxial stresses in polymeric materials gave significant insights into the ESCR phenomena of plastics. More recent ESCR studies have focused on the effects of specific polymeric materials and applications, especially processing conditions.1517
Figure 4. Relationship between time to craze and stress level of RTPU-A (mold temperature=13°C) and RTPU-B (molded by supplier). Pure IPA was used as a crack agent.
IPA and IPA-containing solutions are common agents used in hospitals to clean or sterilize plastic medical parts. The crack resistance of plastic materials against IPA is, therefore, very important. Figure 4 reveals that the time-to-craze behaviors of RTPU-A and RTPU-B in pure IPA are quite different, especially at the lower range of stresses. RTPU-A demonstrates better chemical resistance against IPA than does RTPU-B. This difference was also confirmed by ESCR elongation-retention study results (see Table II). RTPU-B lost more than 90% of its elongation after exposure to IPA for 3 minutes at a very low stress level (about 10 MPa), whereas RTPU-A retained most of its original elongation even at very high stress levels (more than 34 MPa).
Table II. Elongation retention of two RTPU materials after 3 minutes of solvent contact. The elongation at break (%) of the control samples was 324 for RTPU-A and 284 for RTPU-B.
Table II also presents the ESCR performance of RTPU-A and RTPU-B against cyclo, MEK, and a cyclo/MEK mixture--agents that are sometimes used to bond plastic devices together. Once again, RTPU-A shows superior ESCR performance compared with RTPU-B. As mentioned previously, the chemical compositions of RTPU-A and RTPU-B are similar, though too few details are available for calculation of solubility parameters. The drastic difference in ESCR responses of the two materials to different chemicals appears difficult to explain solely on the basis of solubility-parameter match theory. Morphological and sequential structures after repolymerization during injection molding may play greater roles in this situation.
Both RTPU-A and RTPU-B had better chemical resistance to Intralipid solution, a 20%-intravenous-fat emulsion that is a milder stress-cracking agent than IPA.18 Based on results from mechanical-property/molding-condition and ESCR screening studies, RTPU-A was selected to make joint devices (see Figure 1) to closely simulate end-use conditions such as polyaxial stresses under more complicated tension/compression modes.
Effects of Chemicals and Heat Aging on Devices. When cracking or crazing develops on a plastic device, four key parameters are determinative: stress, temperature, chemical environment, and exposure time. The effect of temperature can reflect a combination of ambient temperature and the glass-transition temperature of the plastic materials. In the present study, cyclohexanone-based solutions were used to bond the injection-molded devices. Since cyclohexanone is potentially a crack agent to RTPU-A under high stress levels--as shown by the ESCR screening study (see Table II)--caution was taken to prevent joint cracking failure. The as-assembled RTPU-A parts did not show any brittle failure during bending tests immediately after assembly.18 The bending torque was very high, even without tight control on assembling forces or interference.
Figure 5. Bending torque results of a device molded from RTPU-A. Parts were assembled without tight control on interference and were heat-aged at 57°C for 1 week. The horizontal line at approximately 5 in./lb represents the product functional requirement.
Because the durability of heat-aged polymeric devices is a major concern, an accelerated-heat-aging study was conducted on the assembled parts. After heat aging at 57°C for 1 week, some parts that had been assembled without tight interference control started to show brittle failure at a relatively low bending torque (see Figure 5). This decrease in bending torque might be caused by heat-accelerated crack propagation of those microcrazes generated during the solution bonding process. A tight control on assembling force was implemented to reduce applied stress, and the bending-test results showed significant improvement even after accelerated heat aging (see Figure 6).
Figure 6. Bending torque results of a device molded from RTPU-A. Parts were assembled with tight control on interference and were heat-aged at 57°C for 1 week. The horizontal line at approximately 5 in./lb represents the product functional requirement.
An attempt was also made to improve performance by increasing molecular weight through design changes (thickening the wall) and processing variations (raising the mold temperature). However, the bulk molecular-weight data from gel permeation chromatography did not correlate very well with bending performance.18 It appears that the skin and core morphological difference might contribute more to ESCR performance than do bulk properties. Fast cooling on the skin of a device may create a very thin but lower-molecular-weight layer that might be more vulnerable to stress cracking/crazing by chemical solutions. Once the crack initiated on the skin, the stress-concentrated cracking front would propagate more easily into the higher-molecular-weight bulk material (core) and cause the low-force failure. A final observation worth mentioning is that the stress distribution on the joint of the plastic device can be very complicated and different from ESCR predictions based on uniaxial stress considerations. This situation can make it very difficult to predict failure modes based on the ESCR information generated from controlled experiments.
Two rigid polyurethane materials, while differing only slightly in chemical composition, show very different ESCR behavior. Under the same testing conditions, RTPU-A, a simpler addition polymer of methylene diphenyl diisocyanate and 1,6-hexanediol, has much better chemical resistance than RTPU-B, which contains an additional chain-extending diol (cyclohexanedimethanol). It can be concluded that the use of these additional chain extenders in RTPU-B--added to enhance the stiffness and heat resistance of the rigid polyurethane--causes the reduced resistance to environmental stress cracking.
The mechanical properties of RTPU-B are also very sensitive to molding conditions compared with those of RTPU-A. The assembling process control is very critical, and may have significant impact on the integrity of the devices made of RTPU-A. Finally, the combined effects of chemical attack and heat aging must be taken into consideration in order to ensure the long-term ability of RTPU-A devices to prevent brittle failure.
1. Qin C, Ding S, Hong KZ, et al., "Environmental Stress Crack Resistance of Rigid Polyurethanes in Medical Devices," in Society of Plastics Engineers, Inc., Technical Papers, vol XLII (ANTEC 96), Brookfield, CT, Society of Plastics Engineers, pp 36903694, 1996.
2. Hopkins IL, Kaker WO, and Howard JB, J Appl Phys, 21:206, 1950.
3. Berry JP, J Appl Polym Sci, 6:617, 1962.
4. Howard JB, paper presented at the First Annual Conference of the Plastic Institute of America, December 1963; reprinted in SPE Transactions, 4:217, 1964.
5. Howard JB, Polym Eng Sci, July, p 125, 1965.
6. Kambour RP, J Polym Sci: Macromolec Rev, 7:1, 1973.
7. Kambour RP, Gruner CL, and Romagosa EE, Macromol, 7(2):248, 1974.
8. Henry LF, Polym Eng Sci, 14(3):167, 1974.
9. Titow WV, Plast Polym, June, p 98, 1975.
10. Griffith AA, Phil Trans Roy Soc London, A221:163, 1921.
11. Rebinder PA, Z Phisik, 72:191, 1931.
12. Rebinder PA, Izvestiya Akademi Nauk SSSR (Ctd Mat I Estestv, Nauk), p 639, 1936.
13. Rebinder PA, and Kalinowskaya N, Zhurnal Tekhnichkeskoi Fiziki, 2:726, 1932.
14. Howard JB, SPE J, 15:397, 1959.
15. Pham HT, Bosnyak CP, and Sehanobish K, Polym Eng Sci, 33(24):1643, 1993.
16. Liu X, Zhou Z, and Brown N, Polym Eng Sci, 34(2):109, 1994.
17. Sjoerdsma SD, and Boyens JPH, Polym Eng Sci, 34(2):86, 1995.
18. Internal reports of Baxter Healthcare Corp., Round Lake, IL.
Chuan Qin, PhD, is an engineering specialist at the Medical Materials Technology Center of Baxter International Inc. (Round Lake, IL). He specializes in biomedical material and process development for medical devices and drug-delivery systems.
Y. Samuel Ding, PhD, is a senior engineering specialist at the same facility, concentrating on biomedical polymer development.
Victor Zepchi is a principal engineer at Baxter Healthcare's I.V. Systems Division (Round Lake, IL), where Himansu Dhyani is a research associate
K. Z. Hong, PhD, is the division manager.