When a plastic is exposed to a chemical environment, the material may undergo numerous changes. These can include weight gain if the plastic absorbs the chemical, weight loss if the plastic is degraded by the chemical or if the chemical extracts low-molecular-weight components of the plastic, dissolution if the chemical is a good solvent, or other changes such as variations in opacity or color. If the plastic is under stress, it may also experience environmental stress cracking, which can be defined as the crazing and cracking that may occur when a plastic under a tensile stress is exposed to aggressive chemicals. The potential for environmental stress cracking is of paramount concern when plastics are used in medical device components such as luers and stopcocks. In these applications, chemicals such as isopropanol and lipid solutions can initiate crazes—microcracks bridged by polymer fibrils—in the plastic and seriously compromise its mechanical integrity.

Products like medical luers can be susceptible to crazing when under stress and exposed to aggressive chemicals. Photo: Eastman Chemical Co.

The proper selection of a medical plastic requires a thorough analysis and interpretation of the phenomenon of environmental stress cracking. The goal of this article is to provide a framework for evaluating the suitability of a plastic for medical uses in terms of its stress-crack behavior. Our strategy will be to examine, in some detail, the roles of the three critical components of environmental stress cracking: the chemical environment, the plastic, and the tensile stress.

THE CHEMICAL ENVIRONMENT

Chemicals that cause environmental stress cracking can be divided into those that swell or wet the polymer and those that chemically react with the polymer. An example of the latter would be the caustic or aqueous sodium hydroxide that can hydrolyze poly(ethylene terephthalate) (PET).¹ The reduction in polymer molecular weight from the hydrolysis can lead to crazing and eventual catastrophic failure—a mechanism that has been identified as the probable cause for stress-crack failures in one-piece carbonated soft-drink containers.²

This article, however, will emphasize chemicals that cause stress cracking simply by swelling or wetting the polymer. (Lipid solutions and isopropanol fall into this category.) It is the general consensus in the literature that the majority of stress-crack failures experienced by end-users results from this category of chemical.^3–5 Numerous studies have linked the ability of a solvent to swell a plastic with its ability to craze the plastic. Perhaps the best-known work of this kind is that of Kambour et al., who demonstrated—in studies on polycarbonate,⁶ poly(phenylene oxide),⁷ polysulfone,⁸ and polystyrene⁹—that the absorption of the solvent and concomitant reduction in the polymer's glass-transition temperature can be correlated with a propensity for stress cracking. They also showed that absorption of the liquid by the polymer tends to be correlated by the solubility parameters of the liquid and polymer. The solubility parameter, , is defined as

where CED is cohesive energy density, ΔE_vap is the heat of vaporization, and V_m is molar volume. Hansen proposed that the solubility parameter was composed of contributions from the three major types of cohesive forces, namely dispersive, polar, and hydrogen bonding, so that

where _d, _p, and _h are the dispersive, polar, and hydrogen bonding components of the total solubility parameter.¹⁰ Values of these parameters for a few representative chemicals are shown in Table I. If the solubility parameter of the solvent is close to the solubility parameter of the polymer, the polymer will probably show some solubility in the solvent or undergo solvent-induced crystallization. Experience has shown that absorption of a liquid by a polymer may be better correlated by using the partial solubility parameters.

Because the solubility parameter of a polymer cannot be calculated directly from the heat of vaporization, indirect methods such as solvent swelling and group-contribution approaches are used. Solvents that swell or dissolve the polymer most effectively will have solubility parameters close to the solubility parameter of the polymer, in keeping with the adage that "like dissolves like." In the group-contribution approach, the solubility parameter is determined by using the equation

where F_i is the molar attraction constant and V_i is the molar volume for each subsegment of the polymer repeating unit (as demonstrated for PET in Table II).

Table II. Estimation of solubility parameter for polyethylene terephthalate (PET) using the group-contribution approach.¹² PET molecular structure shown at top.

The ability of the solubility parameter approach to correlate the absorption behavior of plastics has been demonstrated by the authors using PET, PCTG (a copolyester), and polycarbonate. Pieces of 3-mil-thick amorphous, unoriented, extruded film were suspended in sealed jars above a few milliliters of liquid. The films were removed every two weeks for weighing until they reached an equilibrium weight. The results, listed in Table III, indicate that all three plastics appear to show a broad peak with a maximum in liquid absorption at a solubility parameter of approximately 20 MPa^½. The solubility parameters for PET, PCTG, and polycarbonate, as determined by the group-contribution approach, are 23.5 MPa^½, 22 MPa^½, and 21.9 MPa^½, respectively, which fall into the range of the broad maximum in liquid absorption. (The alcohols are exceptions to this trend, as seen in Table I, presumably because of their strong hydrogen bonding characteristics.) These results highlight the difficulty in using the solvent-swelling technique for determining the solubility parameter of a polymer and in using the total solubility parameter to correlate a polymer's absorption behavior.

THE PLASTIC

All plastics can be classified as either amorphous or crystalline materials. In amorphous plastics such as polystyrene and poly(methyl methacrylate), the polymer chains are randomly configured, displaying no significant order. In crystalline plastics such as polyethylene and nylon 6/6, the polymer chains are aligned or ordered into crystallites.

Crystalline plastics, however, are never completely crystalline, but rather contain regions of amorphous material. A few plastics, among them PET and polycarbonate, can be entirely amorphous or semicrystalline, depending on processing conditions. At room temperature, the thermodynamically favored state for these plastics is the crystalline form; however, if they are cooled rapidly enough from the melt to below their glass-transition temperatures (T_g), they will remain in their amorphous forms. Under normal injection molding conditions, parts made from such plastics are clear, indicating the absence of crystallinity. If the finished parts are heated to above T_g or are exposed to strong solvents, they will crystallize. The latter phenomenon is often called solvent-induced crystallization, and was observed during the absorption studies discussed above.

The nominally amorphous PET, PCTG, and polycarbonate films turned opaque upon exposure to certain solvents (see italicized data in Table III), indicating that crystallization had occurred. Absorption of these liquids decreased the T_g's of the plastics to at least ambient conditions, giving the polymer chains sufficient mobility to align and crystallize. It was noted that crystallinity developed much more quickly in PET and PCTG than in polycarbonate, primarily because the copolyesters have much lower T_g's (approximately 80°C) than does polycarbonate (approximately 150°C), and therefore needed to absorb less liquid before T_g was depressed to ambient temperature. Solvent-induced crystallization may have a pronounced effect on stress-cracking behavior, as will be discussed later.

Although both amorphous and crystalline plastics are susceptible to environmental stress cracking, it is generally recognized that amorphous plastics tend to be more at risk.^3–5 The closely packed crystalline domains in crystalline plastics act as barriers to fluid penetration and therefore tend to resist environmental stress cracking.

THE TENSILE STRESS

Plastics will exhibit environmental stress cracking when exposed to an aggressive chemical environment if and only if a tensile stress is present. The tensile stress may be applied externally or may simply be a consequence of residual, or molded-in, stresses. Residual stresses can be minimized through the use of proper design guidelines and the control of critical variables in the injection molding process.¹³ Externally applied stresses can result from subassembly processes, shipping and storage conditions, or improper packing. An externally applied tensile stress may also be part of the intended end-use of the device. A female luer, for example, may be subjected to extremely high hoop stresses upon insertion of the male luer.¹⁴

Obviously, the most reliable method for evaluating the stress-crack resistance of a plastic in a given application is to analyze its performance under simulated end-use conditions. Alternatively, stress-crack resistance can be determined by some type of standard testing procedure whose results can be related to the stress and strain levels observed in end-use conditions. A few of the numerous tests that have been developed to evaluate environmental stress-crack resistance are listed in the box on page 41. The tests differ primarily in the way the external stress is applied.

ASTM D 1693 describes a test for evaluating the stress-crack resistance of ethylene plastics in environments such as soaps, wetting agents, oils, or detergents. Strips of a plastic, each containing a controlled defect, are placed in a bending rig and exposed to a stress-cracking agent. The number of specimens that crack over a given time is recorded.

ISO 4600 details a ball or pin impression method for determining stress-crack resistance. In this procedure, a hole of specified diameter is drilled in the plastic. An oversized ball or pin is inserted in the hole and the plastic is exposed to a stress-cracking agent. After exposure, tensile or flexural tests may be performed on the specimen.

A constant tensile-stress method is outlined in ISO 6252, in which a test specimen is exposed to a constant tensile force while immersed in a stress-cracking agent so as to determine time-to-rupture under a specified stress. Variations of this test include a tensile creep test that monitors strain, and a monotonic creep test that uses a constant stressing rate instead of a fixed stress.⁵

Another bent-strip method for evaluating stress-crack resistance is presented in ISO 4599. In this test, strips of a plastic are positioned in a fixed flexural strain and exposed to a stress-cracking agent for a predetermined period. After exposure, the strips are removed from the straining rig, examined visually for changes in appearance, and then tested for some indicative property such as tensile strength.

A variation of the ISO 4599 test was performed by the authors on medical-grade versions of PCTG, polycarbonate, and an acrylic resin. Injection-molded tensile bars were placed under fixed strains of 0%, 0.5%, 1.5%, and 2.7% and exposed to a lipid solution and isopropanol using a wet-patch technique. After a 72-hour exposure period, the specimens were removed from the strain rig, rinsed clean of chemical with distilled water, allowed to equilibrate at ambient conditions for 24 hours, and then tested for residual tensile properties according to ASTM D 638. The results for specimens exposed to isopropanol and to lipid solution are displayed in Tables IV and V, respectively. In isopropanol, polycarbonate crazed at strain levels of 1.5% and higher, resulting in dramatic losses in tensile properties. PCTG also crazed at strain levels of 1.5% and higher but maintained most of its tensile properties. The acrylic resin fractured on the strain rig shortly after the appearance of crazing, at a strain level as low as 0.5%.

In the lipid solution, polycarbonate crazed at a strain level as low as 1.5%, and fractured while still on the strain rig at a strain level of 2.7%. The acrylic resin crazed only at the 2.7% strain level, and demonstrated a dramatic loss of tensile properties for this one strain level. PCTG crazed at a strain level as low as 1.5% but retained most of its tensile properties. One can speculate that the ability of PCTG to retain most if its mechanical properties despite being heavily crazed is linked to its ability to undergo rapid solvent-induced crystallization.

The critical strain test attempts to determine the minimum strain required to initiate crazing in the presence of a stress-cracking agent. The test is most commonly performed using a Bergen elliptical strain rig: a strip of plastic is placed on the rig, which is patterned after a quarter of an ellipse, and exposed to a stress-cracking agent.¹⁵ The strain at any point along the elliptical rig, , is given by the equation

where a is the semimajor axis, b is the semiminor axis, t is specimen thickness, and X is the distance along the semimajor axis to the point of interest. With a 0.125-in.-thick specimen, an elliptical rig with a = 10 in. and b = 5 in. will give minimum and maximum strains of 0.31% and 2.50%, respectively.

Results of critical strain tests on PET, PCTG, and polycarbonate are listed in Table VI. Injection-molded flexural bars were strapped to a Bergen elliptical strain rig and exposed to the liquids using a wet-patch technique. The width of the patch was smaller than the width of the flexural bar to avoid exposing the edges of the flexural bar to the liquids. Polycarbonate displayed outstanding critical-strain values in alcohols and aliphatic hydrocarbons but very low critical-strain values (<0.31%) in most other liquids. In fact, polycarbonate fractured while still on the strain rig when exposed to acetone and dimethyl sulfoxide. PET and PCTG displayed moderate to high critical-strain values in most solvents. Finally, it is of interest to note a modest correlation between the critical-strain values reported in Table VI and the weight-gain values reported in Table III: liquids that tend to swell the polymers the most also induce the lowest critical-strain values.

Several laboratories have used the concepts of fracture mechanics to evaluate stress cracking in plastics.^16–19 The basic premise of fracture mechanics is that the strength of a material is determined by the presence of flaws. In fracture-mechanics testing, a well-defined flaw or crack is machined into a plastic. The specimen is stressed and the growth of the flaw in the presence of a stress-cracking agent is monitored until failure. Whereas fracture mechanics is ideal for studying the effect of a preexisting crack on the residual strength of a plastic, it provides no insight into the mechanism for initiation of a crack or craze upon exposure to a stress-cracking agent.