Electromagnetic and mechanical properties of conductive polyethylene composite antennas for utility location

High density polyethylene (HDPE) pellet was purchased from Premier Plastic Resins, a commercial supplier, and used as received. Conductive carbon black, with an average particle size of 30 nm, was purchased from Asbury Carbons, a commercial supplier. Aluminum flake, acquired from Transmet Corporation, has an average area of 1.61 ± 0.51 mm² and an average thickness of 0.038 ± 0.007 mm. To create conductive composite polyethylene, additives of varying percentages by volume were compounded into polyethylene in a Banbury mixer and mixed for fifteen minutes at a temperature of 160 ^°C until the fillers were distributed through the matrix.

Resistivity was measured for conductive polyethylene specimens with increasing amounts of fillers. Specimens containing up to 25% of carbon black and 20% aluminum by volume were produced. Four-point resistance of these specimens was measured according to the procedure presented in ASTM Standard D 4496 [29]. Specimens were measured in the plane of the material and through the thickness of the material to determine if the compression molding process resulted in anisotropy in the resistivity material.

Dogbone specimens were manufactured from unmodified polyethylene, polyethylene compounded with 15% carbon black by volume, and polyethylene compounded with 15% carbon flake and 10% aluminum flake by volume. A conductive polyethylene antenna molded to a polyethylene utility pipe was simulated with bilayer specimens consisting of a conductive polyethylene layer molded to an unmodified polyethylene substrate. A heated hydraulic press was used to mold plates of conductive polyethylene to plates of unmodified polyethylene, as shown in figure 1(a). The resulting bilayer polyethylene plate was molded to the desired thickness using aluminum spacers on the same hot press. Dogbone specimens of single materials and the bilayer polyethylene were cut to dimensions according to ASTM Standard D 638 [30]. Tensile testing was performed with a uniform speed of 0.005 mm s⁻¹. For the bilayer polyethylene, wires were molded into the specimens to serve as electrodes for measuring the change in electrical properties due to increasing strains on the specimens. A schematic of the bilayer resistance test can be seen in figure 1(b).

Zoom In Zoom Out Reset image size

A test fixture was constructed to measure the strains that a typical polyethylene pipe would experience in service. A 1" polyethylene pipe with dimension ratio 11, representing a standard distribution line, was bent over a mandrel to industry-standard minimum bend radii for both short and long term bending. On a pipe in bending, the location of the highest strain will exist on the outermost point on the pipe, which is the farthest away from the neutral axis and therefore undergoes the most deflection. Strain at the outermost point of a pipe during bending to radii of 32", 23.5", and 17" was measured with digital image correlation [31].

To test the performance of antennas made from the conductive polyethylene, cross-polarized bowtie antennas were cut from compression molded sheets of conductive polyethylenes using a 50 W laser cutter. The bowtie shape was chosen for ease of manufacturing and a wider bandwidth than similarly sized dipoles. Because the lobes of the antenna are cross-polarized, the response of the antenna to GPR is less dependent on the orientation of the antenna within the pipe. One manufactured antenna can be seen in figure 1(c). After manufacturing, the antennas were molded to sheets of polyethylene to simulate an antenna molded to a polyethylene utility pipe.

The performance characteristics of an antenna are dependent upon the properties of its surroundings [32]. Antennas intended for utility location should therefore be tested in conditions similar to those that would be expected in service. An underground testbed was constructed by burying a 8" PVC pipe to a standard coverage depth of 18". During testing, a nylon line running through a conduit was used to pull an antenna into the underground pipe. To conduct a radar survey of the area, the GPR transmitter (GSSI-SIR-3000 with a 200 MHz antenna) was moved in a straight line perpendicular to the direction of the pipe and a B-scan of the area was collected. RMS amplitude of the radar response was analyzed to determine the percentage of the radar signal above a noise threshold. The radar signal of an antenna with 5% carbon black, with a lower conductivity, and an antenna containing 15% carbon black and 10% aluminum, with a higher conductivity, were compared to the signal of an empty pipe.

Polyethylene containing only aluminum flake as a filler was not conductive at any measured concentration level because the large flakes were unable to form a continuous conductive path through the polyethylene matrix. Measured resistivities of conductive polyethylene containing only carbon black, shown in figure 2(a), indicate that percolation for this carbon black in polyethylene occurs at around 15% by volume, and the inclusion of additional carbon black in the composite does not result in further reduction of resistivity. Resistivity for composite polyethylenes with only carbon black as a filler material was isotropic when measured in the plane of the material and through the thickness of the material.

Zoom In Zoom Out Reset image size

Measured resistivities of specimens with both carbon black and aluminum are shown in figure 2(b). For the specimens without aluminum flake, no significant difference exists between the resistivity measured through the length of the material and the resistivity measured through the thickness, as shown by the schematic in figure 2(c). For composites with aluminum flake, average resistivity measured through the length of the material is often an order of magnitude lower than resistivity measured through the thickness of the material. Previous studies have reported similar differences in the resistivity of the material when measured in plane or through thickness due to alignment of fillers [21]. In the plane of the material, percolation of the aluminum flake within the conductive polyethylene matrix occurs at 5% loading by volume. However, when resistance is measured through the thickness of the material, the aluminum flake does not reach percolation until 10% by volume and has higher scatter. Examination of cross sections of the specimens reveals that aluminum flakes in the material became aligned with the plane of the material during compression molding, as seen in figure 2(d), resulting in the anisotropy in material resistivity.

Representative stress-strain curves and typical failure behavior for the individual materials and bilayer specimens is shown in figure 3. Neat polyethylene can draw to several times its own length under tension and exhibits an extremely ductile fracture surface, shown in figure 3(c). The high concentrations of filler materials in the conductive polyethylene, which increase the stiffness of the material, restrict its ability to draw, resulting in brittle failure behavior at much lower strains. The brittle fracture surface of a conductive polyethylene specimens can be seen in figure 3(b). In the bilayer specimens, composed of conductive polyethylene molded to unmodified polyethylene, brittle fracture which begins in the doped polyethylene layer can propagate quickly into the neat polyethylene layer, resulting in brittle failure through the cross sectional area of the specimen. The sharp decrease in the bilayer curve in the stress in figure 3(a) represents the point at which brittle fracture occurs and propagates through the cross sectional area of the specimen. Some tendrils of neat polyethylene remain, however, and drawing of these tendrils is shown in the gradual decrease in the stress after the initial brittle fracture, though these tendrils are small and will only draw for short displacements before fracture. Evidence of this behavior can be seen in the fracture surfaces of the bilayer specimens, as in figures 3(d) and (e).

Zoom In Zoom Out Reset image size

Experimentally measured elastic modulus for all specimens is shown in figure 4(a). The presence of carbon black and aluminum, two stiff materials, in the conductive polyethylene resulted in an increase in stiffness. As the amount of aluminum flake in the composite increased, the stiffness of the material did not increase, but the amount of scatter within each dataset did, indicating that the aluminum flake within the material can act as void content rather than reinforcement due to the materials not effectively bonding. The bilayer specimens had a lower effective modulus than any of the single material specimens.

Zoom In Zoom Out Reset image size

Failure strains for all tested specimens are shown in figure 4(b). Failure strain data is critical for industry use, because field practices for installation of polyethylene pipe rely on physical dimensions that are safe for the pipes, but load measurements are not taken during this time. Typically, the failure strain for any material is defined at fracture, but because pipes used in industry are limited to strains below drawing, failure strain for this study was taken at the onset of drawing within the polyethylene. As expected from the failure behavior, adding such high quantities of carbon black to the composite results in a brittle material with much lower strain-to-failure than the neat polyethylene.

For each bilayer specimen, the resistance of the specimen between the grip sections was measured during tensile testing. Stress and normalized change in resistance were both plotted with strain. A representative graph of this data be seen in figure 4(c). The sharp decrease in the magnitude of the stress occurs when the crack propagates through the doped layer and most of the neat polyethylene layer. The following gradual decrease in magnitude shows drawing in the remaining polyethylene tendrils after initial separation. The initial change in resistance occurs at the sharp decrease in stress, which corresponds with the initial specimen separation. The sharp increase in resistance represents the point at which the two ends of the specimen became completely separated.

For this dataset, the average difference in displacement between the initial separation and the loss of electrical contact is less than the average length of an aluminum flake. It is likely that aluminum flakes bridge the two sides of the specimen for a short distance before being pulled out of one side of the doped polyethylene matrix.

Strain to failure for the bilayer specimens showed large variance due to the relative thickness of each of the layers. Figure 4(d) shows the modulus and failure strain for each bilayer specimen compared to the relative thickness of the conductive polyethylene layer in that specimen. For instance, a specimen with equal layers of conductive and neat polyethylene would have a thickness ratio of 0.5. As seen in figure 4(d), both elastic modulus and failure strain were dependent on the thickness ratio of the conductive polyethylene layer. As the thickness of the conductive polyethylene increases, the effective modulus of the composite increases, but the strain before fracture of the composite decreases because the conductive polyethylene layer limits the strains that the composite specimen can attain.

Fracture surfaces of the conductive polyethylene specimens, as shown in figure 5(a), were imaged with a scanning electron microscope. Figure 5(b) shows a high magnification image of the interface between the conductive and unmodified polyethylene layers. At high magnification, the unmodified polyethylene shows evidence of more ductile behavior in the fracture surface than does the conductive polyethylene, which appears more brittle, with shorter tendrils. Though the two materials exhibit visually different behavior when examined at high magnification, the two materials appear to be well bonded across the entire width of the specimen. The bond between the materials allows for fracture to propagate through the specimens.

Zoom In Zoom Out Reset image size

Figure 5(c) shows gaps between an aluminum flake and the surrounding polyethylene. Figure 5(d) shows valleys in the fracture surface where aluminum flakes were pulled out of the polyethylene matrix during tensile testing. The lack of bonding between the aluminum flakes and the polyethylene matrix contributes to the increased variance in material properties as the percentage of aluminum flake in the filler increases. These gaps are likely formed during the cooling process due to the differences in the coefficients of thermal expansion of the two materials. These gaps have previously been reported in studies of aluminum and polymer composites [33].

In industry, polyethylene pipes are classified by their dimension ratio (DR), the ratio of the outside diameter of the pipe divided by the wall thickness [34]. A smaller DR represents a higher wall thickness, and thus a higher pressure rating. A pipe with a higher DR can be bent to a tighter radius of curvature because the thicker wall reduces the tendency of the pipe to kink during bending. Because polyethylene has such high ductility, a pipe can often be bent to a tighter bend radius for short periods of time, such as during transportation or installation. A standard distribution line with DR 11 has a long term minimum bend radius of 32" but can be brought to a bending radius as small as 17" during short term usage, such as transportation or installation [31].

Experimental failure strains for bilayer polyethylene samples were compared to the strains measured on a standard polyethylene pipe during bending, as seen in figure 6. Strains measured on the pipe during bending are within the range of measured failure strains, suggesting that damage could occur to the conductive polyethylene antenna at strains seen during in service. However, the failure strain experienced by a specimen was dependent upon the relative thickness of the conductive polyethylene layer, and all specimens with a thickness ratio less than 0.3 had failure strains higher than the highest measured strain on the pipe in bending. Pipes with installed antennas that are thin relative to the pipe wall thickness would be likely to survive in service.

Zoom In Zoom Out Reset image size

Figure 7 shows a schematic of the GPR testbed and one of the tested antennas, a bowtie antenna made from conductive polyethylene with 5% carbon. As the GPR was moved across the survey area, each individual response was recorded and combined to produce the response of the whole area, called a B-scan. Figure 7(c) shows the B-scan of the GPR survey of the empty pipe with the analysis window highlighted. For this same scan, the RMS amplitude of the signal in the amplitude window and the signal threshold are shown in figure 7(d). The amplitude of the radar scans for the empty pipe, the pipe with an antenna made from conductive polyethylene with 5% carbon, and the pipe with an antenna made from conductive polyethylene with 15% carbon and 10% aluminum were compared with the signal threshold. Figure 7(e) shows the percentage of each of the three radar scans above the signal threshold. Both conductive polyethylene antennas tested were able to increase the radar response of the pipe to the GPR. The antenna made from polyethylene containing both carbon black and aluminum, which had a higher conductivity, had the strongest response during testing.

Zoom In Zoom Out Reset image size