Performance criteria for prosthetic sockets and orthotic devices include strength sufficient to withstand maximum expected stresses, durability, minimal weight, and minimal fabrication cost. Unfortunately, designing for these goals is difficult because of a lack of comprehensive data on component material properties. One study1 measured tensile strengths of laminations of carbon, glass, and nyglass and found relative strength ratios of approximately 100:50:1, respectively. Another study compared the strengths of a carbon and fiberglass hybrid with either fiber alone and found, contrary to expectations, that the hybrid was weaker.2 Valuable data have been obtained on the properties of socket interface materials as they relate to the friction and shear forces on tissue; however, they provide no information on mechanical strengths.3–5 Specific lay-up compositions are recommended by some manufacturers but without quantitative mechanical rationale.6 Similarly, no mechanical data exist to compare to the new socket compositions on the horizon that might be derived from polymer deposition,7 selective laser sintering,8 or other novel means of manufacturing prosthetic sockets.
This article reports material properties of typical prosthetic laminations that would allow prosthetists to fabricate lighter and more cost-efficient prosthetic sockets without sacrificing safety.
MECHANICAL BEHAVIOR OF LAMINATES
Laminated prosthetic sockets, fabricated by applying a resin (matrix) to lay-ups of one or more layers of materials (fibers) on a mold, are composites, or fiber reinforced plastics (FRPs). By adding a small amount of strong fiber to a weak resin, FRPs establish a strong architectural structure, resulting in a strong but inexpensive material. Strength of a laminate is determined by the material properties of both fiber and resin and the degree of intimacy achieved, as well as the resulting structural architecture. Strength of FRPs, in both tensile and bending modes, is greatest along the axis of the fibers. Because of this property, the axis generally is altered from layer to layer to increase stiffness in all directions. If a lay-up were to contain only weaves that had ±45° orientation, there would be a definite overall axis of strength, and correspondingly a direction of weakness. Random fiber orientation or 0/90° orientation should be incorporated for an ideal lay-up.
Physical properties of individual layers in the lamination lay-up are important because they help determine the physical properties of the entire lamination. Important material properties are:
Stress = the amount of force applied over a given area.
Strain = the amount of deformation for a given length.
Young's modulus = Stress/Strain = the amount of force to obtain a certain amount of deformation. Young's modulus is synonymous with stiffness.
Ultimate tensile strength (UTS) = Maximum force applied before a fiber breaks.
Yield strength = Maximum force applied before permanent deformation occurs.
Bending stiffness = Modulus during bending.
Brittleness and ductility = Relative terms describing how much plastic deformation occurs before fracture. Brittle materials break with small deformation, whereas ductile materials break after high deformation.
Published values of material properties for typical fabrication materials are presented in Table 1.9,10 Note that the modulus of the SPT resin (SPT Technology, Inc., Minneapolis, MN) is slightly higher than the modulus of nylon. The nylon (fiber) is supposed to reinforce the resin (matrix), but with the modulus of nylon fibers lower than that of the resin, it does not serve to increase strength. The material properties would be completely dominated by the resin.
Both carbon and fiberglass are very strong, but they do have limitations as prosthetic socket materials because of their brittleness. They break, rather than stretch, during postfabrication modifications. Significant strength is lost, making sockets with carbon and fiberglass not well suited for postfabrication modification unless they are significantly overbuilt.
Characterizing properties of sockets themselves is difficult because of their complex shapes and the dynamic, irregular nature of their stress patterns. Consequently, it has been difficult to establish expected stresses or loads on prosthetic and orthotic devices. Establishing expected values is the next step in optimizing socket design.
The fiber materials tested were: Perlon Stockinette (Otto Bock Healthcare, Minneapolis, MN), Nyglass Stockinette (IPOS, Bauerfeind Prosthetics USA, Kennesaw, GA), Spectralon (Comfort Products, Kingsley Manufacturing Co., Costa Mesa, CA), nylon (Cascade Orthopedic Supply, Chico, CA), cotton (Cascade Orthopedic Supply), fiberglass matte (IPOS), fiberglass cloth and carbon cloth (Foresee Orthopedic Products, Oakdale, CA), and carbon braid (Foresee). Three lamination resins were tested: Laminhartz 80-20 (Otto Bock), Epoxyacryl (Foresee), and Lamination Resin 80-20 (IPOS). Thus, there were 24 combinations of samples for tensile testing. Tension values were chosen because they are an important component of bending and fatigue stresses, primary modes of socket failure. Compression testing was not done because the thickness of the samples would allow too much bending to get an accurate measure. To simulate vertical loading, all samples were tested across the axis that would normally be vertical in a socket lamination. For perlon, nyglass, nylon, cotton, and spectralon the fibers were at ±45°. Fiberglass cloth and carbon cloth are 0/90°. Fiberglass matte has randomly oriented fibers. Testing was done in accordance with Standard ASTM D 3039, with the exceptions that at least three specimens per sample, instead of five, were tested and the axis of loading is as stated above. This exception was made because of the large number of samples and to save time and expense for preparation, evaluation, and analysis.
Samples for bending point testing were made with two layers of carbon and six layers of nyglass so all samples had eight layers. An Instron four-point bending jig (Instron Corporation, Canton, MA) was used. Bending strength is related to thickness of the sample, so care was taken to maintain uniform thickness for all samples by creating all samples in one lamination to mitigate effects of the fabrication process. The first carbon layer was separated from the second carbon layer by zero, two, four, and six layers of nyglass, respectively. Half the samples were tested with one side up and the other half with the opposite side up. As a sample starts to bend, the surface facing up becomes convex and the surface facing down becomes concave. To increase sample size, each sample was then tested in the opposite orientation.
All laminations were performed under vacuum, within an inner and outer polyvinyl alcohol (PVA) bag at approximately 30 mmHg at room temperature. The laminations were done on one positive mold, rectangular cuboid in shape, with each face approximately 5″ × 5″ × 9″. One material was placed on each face with spray adhesive. Top and bottom faces were not used. Tensile samples were cut with a JPC Precision Sample Cutter (Thawns Albert Instrument Group, Philadelphia, PA) to 1″ × 8″. The first 1.5″ of each end was used as the tab, leaving a testing gauge length of 5″. Thickness and weight measurements were taken for each sample using a vernier caliper. Samples were tested in tension to failure using an Instron Series IX system (Instron). Bending point samples were cut to 1.5″ × 8″ on a diamond tile cutter (MSC Industrial Supply Co., Inc., Melville, NY). Samples were tested with Instron Series IX system in compression with a four-point bending jig. It was not possible to test bending samples to failure because of dimensional limits of the apparatus, so samples were bent to the limit of space in the bending jig, about 1/2 inch of compression.
Data are reported as mean of three or more replicates ± standard deviation. ANOVA was used to determine differences among samples. Statistical significance was set at p < 0.05.
SAMPLE PHYSICAL FEATURES
Figure 1 shows the average thickness and weight for the sample set. There were no significant differences across all fibers for any resin. Some resins seemed to bond better to specific materials than others, based on thickness comparison.
Figure 2 shows typical stress-strain curves obtained for laminates using fiberglass cloth and Spectralon, respectively. Note that the UTS is the highest point achieved during the curve and is used to indicate the amount of stress a material can undergo before fracture. This point may be close the fracture point (brittle), or far be away from the fracture point (ductile). Note the differences between fiberglass cloth and Spectralon in both UTS and brittleness. Failure breaks were generally in the middle of samples, but occasionally breaks were noted near grips. No differences were noted because of mode of failure.
Figure 3 and Table 2 show the UTS for combinations of resins and reinforcement materials. Statistically significant results, as tested by ANOVA with p < 0.05 were found for perlon, nylon, cotton, nyglass, spectralon, fiberglass matte, and fiberglass cloth.
The two fiberglass samples and carbon cloth sample were much stronger than the other materials, with fiberglass being about four times stronger and carbon about eight times stronger than the other materials. The data set can be divided into two basic categories (Table 3), structural materials (fiberglass and carbon) and filler materials (nylon, cotton, perlon, Spectralon, and nyglass).
Although great effort was taken to reduce variability of samples, it was still present. Thickness of individual samples varied, so multiple measurements were taken and the thinnest measurement was used. It is worth noting that the average measurement is very close to the thinnest measurement; variation was generally in increasing thickness, very likely because of excess resin. The average standard deviation during tensile testing for all samples was 13 percent, 9 percent, and 12 percent of the total value for Epoxyacryl (Foresee), Lamination Resin 80-20 (IPOS), and Laminhartz 80-20 (Otto Bock), respectively. It is not clear how these results extrapolate to thicker samples. Single layer laminates have a maximum surface-to-interior ratio, so any surface defects would have a greater impact on these samples. In addition, it may be that variations in individual layers could average out, reducing the overall variability in thicker samples. It is also possible with the large variations in equipment and technique used that the variation is equal to or greater than that found here.
Results from the bending tests are shown in Figure 4. All samples were of equal thickness, so the difference is solely attributed to carbon placement within the sample. With the carbon originating on the concave side, a linear increase in force required to bend a specific distance was noted with R2 = 0.868, which is acceptable for such a small sample size. The distance was the maximal distance allowed by the four-point bending jig, determined to be just before contact of the sample with the shaft of the jig—approximately 1/2 inch. Results showed a clear trend of increasing stiffness as the distance between carbon fibers increased. Although the sample is flat, after it is bent, one side becomes concave and the other convex. The results showed that carbon placed on the convex side of the bend increased the bending strength relative to the same placement on the concave side of the bend (Figure 5). No differences caused by starting orientation were noted. Samples with separation of six layers are symmetrical, and there is no difference. An example for a transtibial prosthesis would be at the mediolateral wings; for a transfemoral lateral wall, this would correspond to the inside of the prosthesis. This difference is attributable to the fact that when carbon fiber is on the convex surface, the carbon fiber is in tension, and when carbon fibers are on the concave surface, the nyglass is in tension. Carbon fiber has a much higher tensile modulus than nyglass, for which on the opposite side compression is influenced mostly by the resin (matrix) properties.
LIMITATIONS OF THE STUDY
The material properties presented here on bending and UTS do not indicate fatigue resistance; this important property was not tested with repeated loadings. In addition, raw materials were obtained from only a few manufacturers, which may bias the results; a more thorough test of all manufactured materials was beyond the scope of this study. However, we do not expect substantial variations in property among the different vendors of the major components used in the field.
Measuring weight and thickness of a single layer is not optimal because there were variations in sample thickness both between samples of the same material and within a single sample. Although thickness variations were not major throughout the sample, they would cause slight errors in the stress calculations. It would be expected that multiple layers would balance out to provide a more uniform measurement because small variations in thickness of individual layers would tend to cancel out with other layers. In addition, more fiber layers might have a better fiber-volume ratio, increasing strength.
There are difficulties laminating small samples of single layers, which may not make it the ideal method for obtaining experimental data. First, the material does not encircle the mold, increasing the likelihood of wrinkling and requiring some sort of adhesive to hold the sample in place. There have been reports of adhesives having an adverse affect on material properties.1 Second, any effects from the inner and outer edge would be magnified in proportion to a thicker lamination, leading to a decreased volume fraction and more potential for voids. However, thicker laminations would require testing with higher forces than available from the Instrom.
This work provides preliminary data on material properties of laminated prosthetic sockets. Although there were quantitative differences among different resins, this did not yield a clearly superior resin. Each resin seemed to perform better with certain materials. The Epoxyacryl seemed to be easier to use because of fewer problems with air bubbles and better saturation. It can be seen from Table 3 that mechanical properties fall into three ranges: at the low range, perlon, nylon, cotton, nyglass and Spectralon contribute approximately equivalent strengths; the midrange includes either fiberglass cloth or mat; and carbon fiber provides the highest strength.
The materials tested can be ranked in order of cost from least to most expensive as follows: nylon, cotton, perlon, nyglass, and Spectralon. Spectralon had a lower UTS than did nylon and cotton, whereas nyglass was equivalent. This raises the question, if there is no advantage in performance, what is the advantage of using the more expensive materials? We found this to be a surprising result and investigated what others had found. Taylor et al.1 found that nyglass did not contribute to the structural integrity of a lamination and decided it should not be used as a structural support for a prosthetic socket. Spectralon is supposed to have a high tensile strength and be impact resistant (tough). The results did not support a high tensile strength for this weave, and the authors are unsure of how toughness is increased when the fiber is typically buried in the middle of a lamination and unable to deform without delaminating from the more rigid fibers. At this time, there is no evidence that Spectralon or nyglass offer any improved properties when used as filler between carbon or fiberglass. Additional toughness tests could be done in the future to analyze this point further. Klasson2 proposed that it was because the matrix (resin) was stronger than the fibers from some materials, a property he called matrix dominated.8 Results from this study support that view, that the structural integrity is attributable more to the resin than to the perlon, nylon, nyglass, cotton, or Spectralon.
Bending tests show strength increased as the carbon fibers separated. This result, sometimes called the “I-beam effect,” is expected. This principle has been incorrectly interpreted to mean that separation of fibers (i.e., carbon) is desirable by alternating carbon and a filler for a layered lay-up. A sandwich type lay-up, with the strongest material on the outside and a less stiff filler material core on the inside, is a better design because it puts carbon as far away from the neutral axis as possible. It is not cost efficient to use carbon fibers near the neutral axis.
The ideal socket should be neither under- nor overdesigned: it should have optimal ratios of cost/benefit and strength/weight. The mechanical data presented here demonstrate that there are definite limits in these important ratios, and reaching optimal compositions can be achieved only in accord with them. Thus, a database of material properties is essential to good prosthetic practice and can lead to setting high standards for sockets similar to those applied to other essential components. Adherence to standards will reduce variation in sockets, liability in material failure events, and costs of materials.
Material properties are important, but what a prosthetist needs to know is what happens when a layer of material is added. Carbon fiber is stronger than fiberglass matte, but because of the difference in thickness, a layer of carbon and a layer of fiberglass matte take about the same amount of force to break. A layer strength could be approximated as the UTS × thickness. Even with exacting materials testing, there are still many questions to be answered. If one can evaluate sockets quantitatively, what are the criteria for the best socket? Three possibilities seem probable: strength-to-weight ratio, strength-to-cost ratio, or strength-to-cost-and-weight ratio. How should it be determined what socket strength is necessary? How should bending rigidity be determined; can it be too stiff? With continued testing, these and other important questions can be answered.
Thanks are extended to Won Jon Rhee for the use of the sample cutter; Dr. Noshir Langrana for the use of the Instron machine for tensile and bending tests; and Allied Orthotics and Prosthetics for use of lamination equipment.
1. Taylor AD, Gilbert JC, Lautenschlager EP. Testing of composite materials used in orthotics and prosthetics. J Assoc Child Prosthet Orthot Clin
2. Klasson BL. Carbon fibre and fibre lamination in prosthetics and orthotics: some basic theory and practical advice for the practitioner. Prosthet Orthot Int
3. Mak AFT, Zhang M, Boone DA. State-of-the-art research in lower limb prosthetic biomechanics: socket interface. J Rehabil Res Dev
4. Sanders JE, Daly CH. How does vacuum forming affect Pelite mechanical properties. Prosthet Orthot Int
5. Sanders JE, Greve JM, Mitchell SB, Zachariah SG. Material properties
of commonly-used interface materials and their static coefficients of friction with skin and socks. J Rehabil Res Dev
7. Goh JCH, Lee PVS, Ng P. Structural integrity of polypropylene prosthetic sockets manufactured using the polymer deposition technique. Proc Inst Mech Eng [H]
8. Rogers B, Stephens S, Gitter A et al. Double-wall, transtibial prosthetic socket fabricated using selective laser sintering: a case study. J Prosthet Orthot
9. Flinn RA, Trojan PK. Engineering Materials and Their Applications,
4th ed. Boston: Houghton Mifflin, 1990.
10. Product literature SPT Technology Inc. Minneapolis, Minnesota.
Keywords:© 2005 American Academy of Orthotists & Prosthetists
laminations; material properties; mechanical testing; socket design