David Williams and Dr. William Pitt, Department of Chemical Engineering
Abstract
The purpose of this research project was to determine the optimal method to solvent bond polyvinyl chloride (PVC) pieces together. Understanding how to create a strong solvent bond in PVC is important to a related prosthetic leg project that has been performed at BYU. This was completed by bonding flat samples of PVC that were prepared with varying methods.
A half-factorial experiment with six variables and 32 samples was designed for this experiment. The samples varied based on solvent amount, solvent type, surface roughness, PVC temperature, clamp time, and clamp pressure. After all of the samples were created, the strength of the solvent bond for each of the methods was determined by performing a standard ASTM lap shear test (ASTM, 2008). These results were statistically analyzed and the influence of each variable on the strength of the bond was determined.
Introduction
An inexpensive prosthetic leg has been developed at BYU that is made out of PVC. This inexpensive prosthetic leg has been designed for use in developing nations, where it is greatly needed. The PVC sole plate component of the 2ft Prosthetics inexpensive PVC prosthetic leg is bonded to the PVC foot component using common PVC solvent glue. While developing this leg, the research group involved noticed that the quality of the solvent bond varied greatly, despite the efforts that were made to create a consistent bond. This research will benefit the prosthetic leg project by determining how to best solvent bond the sole plate to the foot component, thus improving the durability of the prosthetic leg. This knowledge in turn has the potential to benefit thousands of amputees who are unable to afford the more expensive prosthetic legs.
Our research has focused on determining the ideal method for solvent bonding PVC. Solvent bonding of PVC does not glue the pieces of PVC together; instead it temporarily dissolves the PVC, causing it to weld together into one piece. A solvent bond can prove to be very strong when formed properly, similar to a metal weld. Those solvents with solubility parameters similar to PVC (21.5 MPa^.5) should be used, including methlyene chloride (20.2 MPa^.5), acetone (19.7 MPa^.5), cyclohexanone (21.3 MPa^.5), 1,4 dioxane (20.7 MPa^.5), and methyl ethyl ketone (19.3 MPa^.5) (Rodriguez 2003). However to maintain the simplicity of the experiment and to make it easily replicable in developing nations where the 2ft Prosthetics inexpensive PVC prosthetic leg is used, readily-available solvent glues were used. Due to the current use of PVC piping, namely in household plumbing and sprinkler systems, research on PVC bonding is lacking. Therefore this experiment was created to determine what methods should be used to create a strong PVC solvent bond between two flat pieces of PVC.
Materials and Methods
To make the PVC segments, thick-walled PVC pipe was first cut into 2″ pieces using a band saw. These were then heated in an oven to approximately 300 degrees Fahrenheit and cut axially to form two equal halves. While hot, eight PVC segments were placed under a board with 50 lbs on top and flattened. After becoming sufficiently hard (after about three minutes), the PVC segments were placed under
water to be completely cooled and cleaned. The final PVC segments were 2 inches long and 1.5 inches wide. Samples were prepared by
overlapping two PVC segments by 0.75 inches, creating a bonding area of 0.75 inches by 1.5 inches (Figure 1). The remaining lengths were used to pull the PVC segments apart during testing. The samples were numbered one through thirty-two.
The variables tested were solvent amount, solvent type, surface roughness, PVC temperature, clamp time, and clamp pressure. The amount of solvent had two settings, “1” and “2”, meaning one or two coats of the solvent. Two solvent types were used – Oatey Heavy Duty PVC cement and Christy’s Red Hot Blue PVC cement. Samples with a surface roughness of “0” were left alone, and those of “50” had the bonding surface sanded with 50-grit sandpaper. The PVC temperature was either room temperature (72 degrees Fahrenheit) or heated to 125 degrees Fahrenheit. Heating of the PVC was done by placing the samples on a thin heating strip that was connected to a voltage generator and set to 90-95 volts. The samples were weighed down by a 10 lb weight to keep them flat, and periodically measured using a thermocouple until they reached 125 degrees (about three to five minutes). The samples were clamped for either 30 minutes or 24 hours using either 5 lb or 25lb weights. When clamping the samples together, an extra piece of flat PVC was used on each side to help make balancing the weight easier and to keep the PVC flat (see figure 1).
Once all the samples were prepared and left to dry for at least 24 hours, a standard ASTM lap shear test was performed. Each sample was loaded into the test fixture vertically, using the extra PVC that was not bonded to clamp and pull. The sample was then pulled until the bond broke, and the resulting time, loading, and displacement were recorded electronically. Sample number 19 had to be remade because of a faulty joint that did not bond correctly because the PVC curled and cooled too quickly.
Results
During the experimentation of the different variables, immediate variances and complications were observed with the samples that had been raised to 125°F. The heated PVC strips caused a difficulty in ensuring an even gluing surface as the PVC would tend to curl towards its originally pipe shape. In addition, the heated strips would cause the glue to dry faster, and the glue would boil in between the strips when they were applied and clamped under the same heated condition. The statistical analysis supported these observations. An F test comparing the max amount of pounds needed to separate the glued PVC (max load) showed that PVC temperature had the most significant affect on bond strength (F Value: 21.35; Pr > F: 0.001). The only other significant variable in the max load F test was when the amount of solvent was compared with the clamping pressure (F Value: 7.22; Pr > F: .0228). However, we also looked at the top ten pulls that had the highest bond strength. From those top ten: eight had Red Hot solvent glue; eight were at 72°F; seven were sanded; six were clamped for 24 hours; and six had 25 pounds of clamping.
Conclusion
This research will provide 2ft Prosthetics with valuable PVC bonding information. As 2ft Prosthetics makes their PVC leg by heating the PVC pipe, it will be valuable to know that the higher temperatures have a direct effect on lowering the PVC bonding strength. The analysis also showed that the clamping time and pressure are important for ensuring a proper bond. The statistical analysis of the F test did not isolate many of the variables besides PVC temperature, amount of solvent, and PVC clamping; however, through an observation of the top ten samples we see that there is more that needs to be researched. The next step to make this project a success will be to throw out the obvious factor of high PVC temperature. Once we analyze the variables under the same room temperature, we expect to isolate additional conditions like those shown in the top ten of our results: Red Hot solvent glue; sanding the surface; clamp time; and clamp pressure. This experiment that has been completed and the following analysis that is to come will improve PVC bond strength for all those who use PVC.