1. Introduction
Additive Manufacturing (AM), commonly known as 3D-printing, enables innovative product designs with enhanced functional integration, particularly through Multi-Material Material Extrusion (MM-MEX). Electrically conductive polymer composites (CPCs) can be produced by incorporating fillers such as carbon black (CB), graphene or carbon nanotubes (CNTs), enabling them to be used as conductors in electrical circuits. However, most applications of electrically conductive composite require an interface with a conventional electrical system. A significant challenge associated with these composites is connecting them to the conventional conductors, such as copper. (Reference Flowers, Reyes, Ye, Kim and WileyFlowers et al., 2017)
Due to their higher resistivity and smaller conductive surface, contact resistance of CPCs is significantly increased compared to conventional electrical conductors, causing power losses, localised heating and reduced performance. (Reference Watschke, Hilbig and VietorWatschke et al., 2019) The contact resistance depends to the real contact area (AR) which describes the sum of the contact asperities between the contact partners within the apparently contact area (AA), as shown in Figure 1. (Reference Mroczkowski, Jugy, Gerfer, Robik and RobokMroczkowski et al., 2016) The contact system consists of a conductive polymer composite (contact partner A) and a conventional electrical conductor (contact partner B). A bonding agent could partly fill the airgap to reduce contact resistance in the contact system.
Real (AR) and apparently (AA) contact area of an electrical contact system
A typical conductive polymer in MEX consists of a non-conductive thermoplastic matrix material, for example polylactic acid (PLA), thermoplastic polyurethane (TPU) or polycaprolactone (PCL) and conductive additives such as CNTs, CB, graphene or metallic additives. Bonding agents such as electroplated, sputtered, or paste based layers are used in post-processing to enhance conductivity. Conventional contact partners are often composed of Tin, gold, silver or copper. (Reference Daniel, Gleadall and RadadiaDaniel et al., 2021)
Various contacting strategies have been reported. Conventional electronic components can be integrated directly during the manufacturing process through adhesive bonding, embedding, or printing on (Reference Lazarus, Tyler, Cardenas, Hanrahan, Tsang and BedairLazarus et al., 2022). Examples include embedded copper tape (Reference Dijkshoorn, Ravi, Neuvel, Stramigioli and KrijnenDijkshoorn et al., 2022), elastic or screw-type connections (Reference Nowka, Ruge, Schulze, Hilbig and VietorNowka et al., 2024; Reference Watschke, Hilbig and VietorWatschke et al., 2019), and printing onto silver-coated textile substrates (Reference Grimmelsmann, Martens, Schäl, Meissner and EhrmannGrimmelsmann et al., 2016).
In many cases, the bonding agent itself serves as the dominant current path due to its higher conductivity and larger real contact area (Reference Kim and LeeKim & Lee, 2020). Silver-based conductive inks are most widely used, while vapor-phase deposition methods remain uncommon.
Only a few studies have investigated contact resistance (Reference Nassar and DahiyaNassar & Dahiya, 2021; Reference Watschke, Hilbig and VietorWatschke et al., 2019), and a systematic design methodology for electrical interfaces in conductive AM structures remains largely unexplored. This study therefore quantitatively compares various contacting strategies with respect to contact resistance and proposes a framework for the application-specific selection of optimal contact solutions.
2. Methods for design of electrical bondings
The subsequent section presents the types of contacts that are considered suitable for the MEX process in this study. Furthermore, the section will describe the methods and materials that are utilised to determine contact resistance and mechanical strength.
2.1. Contact principles
Nine contact approaches between electrically conductive polymers and conventional conductors are investigated in this study and classified into three categories: force-fit, firmly bonded, and positive-fit connections, as illustrated in Figure 2.
Classification of investigated contact approaches with exemplary contacted samples
Force-fit contacts are further divided into elastic (e.g., screw terminals and spring-loaded contacts) and friction-fit variants (e.g., red cube and bullet connector). Firmly bonded contacts include welding, soldering, and gluing, while positive-fit connections are realized via extrusion onto copper mesh and melt-in inserts. The figure additionally shows exemplary photographs of the investigated variants. Flat contact types can be combined with a bonding agent. In this study, silver paste, sputtering and electroplating are considered primarily, but only silver paste was applied to limit the experimental scope.
2.2. Contact resistance measurement method
The present study employs a measurement approach that utilises the length-dependent resistance of the specimens to compensate for the line resistance, with the objective of determining the contact resistance. The approach has already been successfully applied to determine the contact resistance of AM components (Reference Grimmelsmann, Martens, Schäl, Meissner and EhrmannGrimmelsmann et al., 2016; Reference Lazarus and TsangLazarus & Tsang, 2020). For the purpose of this study, the contacts are manufactured with a variety of lengths (L = 10, 30 and 50 mm) between the contacts on/in the CPC contact partner. The electrical resistance is then measured. The contact resistance RC can be extrapolated from the data independently of the line resistance RS (L). The three measuring points ensure a high degree of accuracy in the extrapolated functions. In order to ensure comparability between the different contacting approaches, the available design space is kept as consistent as possible for all variants. The design space is represented by an imaginary cylinder with a diameter of Ø 5.5 mm. The flat contact approaches thus have a round contact area with a diameter of Ø 5.5 mm (see Figure 3(a)), while the penetrating contact variants have a theoretical contact area that approximates the surface area of a cylinder with the approximately same diameter (d) (see Figure 3(b)). The equivalent circuit diagrams of the measurement setups are illustrated in Figure 3.
Specimen geometry and measurement setup for flat (a) and penetrating (b) contacts
The electrical resistance was determined by utilising a Keithley 2460 source meter (Keithley Instruments, Solon, OH, USA) in a four-wire configuration as an ohmmeter. The four-wire measurement compensates the lead resistances of the force leads (RFL) and sense leads (RSL) of the measurement setup. The sense and force leads were both connected directly to the conventional contact partner. It can thus be concluded that the measured resistance (RM) is the sum of the two contact resistances (RC) and the length-dependent line resistance (RS (L)) of the sample (see Equation 1).
The line resistance (RS(L)) between the contact points exhibits a linear increase in proportion to the distance between the contacted points (L). The line resistance is also contingent upon the cross-section (A), and the resistivity (ρ). In instances where the contact resistances exhibit significant variability, the ratio of resistivity to cross-section (ρ/A) can also be determined for each individual sample. This is achieved by measuring the resistivity for each sample with four-wire measurement with the measuring-leads directly connected to the CNC. The value of the linear function at L = 0 corresponds to twice the contact resistance. The contact resistance of a single contact is therefore calculated according to the following Equation 2:
In order to reduce the relative influence of the contact point on the length-dependent measured resistance, the sample body was made longer for the penetrating contact variants. In the case of flat contacts, the contacts were manufactured and measured from the outside inwards in order to rule out any influence. The presence of two contacts per length resulted in the averaging of the contact resistance with current flow in both directions. Consequently, this ensures that any potential current direction dependence was already accounted for in the measurement. All samples were manufactured five times in each of the 0° and 90° infill orientations, with the current flow direction as the orienting axis. In addition, 100% infill was employed in all cases. For the penetrating test specimens, a single-wall shell was utilised in order to maintain the geometric tolerance.
For measuring the contact resistance using the method described, it is necessary to aim for low dispersion within a sample and the lowest possible total resistance of one sample. Therefore, a preliminary test was conducted in order to ascertain the number of layers required for the test specimens shown in Figure 3. The objective of this experimental trial was to achieve low resistance with low dispersion. The results are presented in Figure 4.
Length-normalized resistance of ALFAOHM for varying numbers of deposited layers
The plot demonstrates the resistance of a sample divided by the measurement length (L) against the number of layers. The measurement of each of the three samples was conducted for each number of layers at length-intervals of 10, 30 and 50 mm. Two trends can be deduced from the evidence presented. Firstly, the measured resistance relative to length decreases exponentially as the number of layers increases. Secondly, the dispersion between the measured lengths and thus within a sample becomes smaller. It has been determined that from 15 layers (layer hight 0.2 mm), there is no significant change in either resistance or dispersion. Therefore, a layer count of 15 is used for fabrication of the test specimens. The first layers of an AM component frequently exhibit divergent properties in comparison to subsequent layers (Reference Dijkshoorn, Neuvel, Stramigioli and KrijnenDijkshoorn et al., 2020), thereby resulting in the initial layers exerting a substantially diminished influence on the measurement outcome as the number of layers increases.
2.3. Materials
The test specimens were fabricated from the PLA/CNT/CB/graphite composite ALFAOHM (Filoalfa, Turin, Italy), which exhibits excellent electrical conductivity. The extrusion temperature for all tests was 220° C, and the extrusion speed was 50 mm/s with a layer height of 0.2 mm. The manufacturing of all samples was undertaken on a tool changer (E3D, Oxfordshire, United Kingdom) with Hemera extruders that were equipped with 400 µm hardened Nozzle-X (E3D, Oxfordshire, United Kingdom).
The conventional contacting materials employed in this study included M3 × 3 brass inserts (ruthex GmbH, Wiefelstede, Germany), Red-Cubes with tin-plated surfaces as press-fit material (part number: 7461057, Würth-electronics, Waldenburg, Germany), spring-loaded contacts with gold-plated surfaces (Feinmetall GmbH, Herrenberg, Germany), banana plugs/tower plugs with tin-plated surfaces (SKS Kontakttechnik GmbH, Niederdorf, Germany) and two copper blocks. PCL Electrifi (Multi3d, Middlesex, USA) filled with copper were utilised for welding, as it is by far the most conductive commercially available polymer composite.
The adhesive used in this study was a silver-filled two-component epoxy (8331D-14G, MG Chemicals Ltd., Burlington, ON, Canada). The determination of contact resistance between an electrical conductor and the composite extruded onto it was achieved through the utilisation of a printed circuit board (PCB) with exposed contact points. In addition to the microscopic form fit with the smooth PCB surface, the investigation extends to encompass a form fit at the mesoscopic level. For the purpose of this procedure, the composite was extruded onto a copper mesh of 100 mesh size. The manufactured contact points are illustrated in the lower section of Figure 2.
The present study additionally investigates the use of conductive silver paste (CSP) EMS #12640 (Electron Microscopy Sciences, Hatfield, PA, USA), sputtered silver layers (experimental setup) and electroplating in a silver bath (cyanide-free silver bath 31g Ag/L, 50 mA/mm2, 3-6 V, 5-10 min deposition time, no preparation) as bonding agents.
2.4. Mechanical strength measurement method
A preliminary categorisation of the mechanical strength of the contact types was derived by testing one specimen from each connection type. Three samples were tested in a quasistatic test (loading rate: 0,2mm/s) to failure on an Instron 5900 10 kN tensile testing machine (Instron, Darmstadt, Germany), recording the maximum tensile force. Flat contact variants were subjected to shear stress, whereas pull-out forces were determined for penetrating variants. The specimen geometry was 20 × 70 × 3 mm; the shear contact area was 10 × 20 mm, while the pull-out test geometry was 20 × 20 mm with the contact area defined by the respective contacting approach.
3. Results
This chapter presents the results of the tests carried out. First, the choice of bonding agent is considered before the contact resistance and influence of contact pressure is discussed.
3.1. Bonding agent
The characterisation of bonding agents was achieved through the measurement of contact resistance, utilising the same method previously described. The contact surfaces have been modified to a size of 20 × 10 mm, ensuring optimal precision during application. The number of samples series were reduced to three to minimize manufacturing time and costs. Spring loaded contacts were used for contact between the measurement and force leads and the bonding agent. The measured contact resistance was significantly reduced for all types of bonding agent in comparison to the absence of a bonding agent. The contact resistance for sputtering was measured at 2.3 ± 0.3 Ω, for electroplating at 0.7 ± 0.2 Ω and for CSP at 0.96 ± 0.42 Ω. Figure 5 presents the SEM images of the boundary layer of the bonding agents.
Interfaces of ALFAOHM: (a) with CSP; (b) sputtered Ag; (c) electroplated Ag
Figure 5 Long description
Panel A: A cross-sectional view of ALFAOHM with silver paste. The image shows the interface between the silver paste and ALFAOHM. The silver paste layer is on top, with a distinct boundary marked by a dashed line. A magnified inset highlights the texture and structure of the silver paste. Panel B: A cross-sectional view of ALFAOHM with sputtered silver. The sputtered silver layer is on top, with a clear interface marked by a dashed line. The texture of the sputtered silver is visible, showing a granular structure. Panel C: A cross-sectional view of ALFAOHM with electroplated silver. The electroplated silver layer is on top, with a distinct interface marked by a dashed line. The texture of the electroplated silver is rough and porous.
It can be seen, that all bonding agents achieved an even coating over the CPC. It is evident from Figure 5a that the boundary layer for the silver paste is considerably thicker (20–25 µm) in comparison to the sputtered (Figure 5b) and galvanised (Figure 5c) samples (both 0.05–0.1 µm). The boundary layer between the composite and the CSP (Figure 5a) is not clearly visible at the highest magnification. The sputtered layer (Figure 5c) is thinner due to the low deposition rates and, as a result of the line-of-sight process, it is only coated on the area directly in line of sight with the source (target). The SEM micrograph of the galvanised sample (Figure 5c) demonstrates the growth of the silver layer over the entire surface without affecting the surface structure of the sample.
3.2. Contact resistance
Compared to conventional components, AM parts show higher surface roughness and ductility, making contact pressure a critical factor for real contact area and contact resistance. For the screw-terminal contact, the effect of different surface pressures was investigated. The results are shown in Figure 6.
Contact resistance according to contact pressure and time dependence of measurement
The contact force was set on a force transducer 85041 (Burster Präzisionsmesstechnik GmbH, Gemsbach, Germany) in six increments (1, 10, 20, 50, 100 and 200 N). Three cylindrical copper profiles with different contact areas (50, 66 and 78.5 mm2) were used as conventional contact partners. The contact area was reduced by central holes in the copper block. In addition, a temporal drift in the measured resistances was noticed during measurement, which also was investigated. The time dependence of the measured values is demonstrated in the upper right diagram in Figure 6, employing a contact resistance that has been normalized to the lowest contact resistance over time. Conversely, an increase in contact pressure has been shown to result in a decrease in drift. For all contact pressures, the drift was found to be negligible after approximately 60 seconds. Thus, a wait time of 60 seconds was utilised for each following measurement.
The curve depicting the relationship between applied contact pressure and resultant deformation exhibits an exponential decrease. It is evident that, within the range of 0 to 1.5 MPa, a discernible correlation exists between the apparently contact area of the copper blocks and the contact resistance. However, this effect decreases with increasing contact pressure. Consequently, at high contact pressures, there is nearly no dependency of apparently contact area, and the contact resistance for all contact areas is approximately 5 Ω.
The utilisation of CSP (application area of 78,5 mm2) resulted in a contact resistance of approximately 8 Ω being achieved, even at very low contact pressures. This contact resistance was then no longer influenced by the contact pressure. The intersection of these curves occurs at a contact pressure ranging from approximately 0.5 to 1.25 MPa, depending the apparently contact area.
The results of the contact resistance measurements for the examined types of contact are summarised in Figure 7. For the purpose of comparison with other types of contact, the contact resistance for the screw terminal variants is utilised at a medium contact pressure of 1.28 MPa.
Overview contact resistance of various contact approaches for CPC
The contact resistance of all contact variants examined, with the exception of extrusion onto a fine copper mesh and Red-Cubes, is in a similar range between 8 Ω and 18 Ω per contact point. Extrusion onto a smooth surface shows the greatest variation, with outliers of up to 110 Ω. It should be noted that in some cases, there was no longer any electrical contact due to the (partial) detachment of the sample extruded onto the PCB. However, with an intact connection, contact resistances below 15 Ω could also be achieved with this variant. When extruding onto the copper mesh (100 mesh), the mesoscopic form fit made the interface more mechanically resilient and the contact resistance and dispersion are lower than between the composite and the smooth surface of the PCB. The melted insert has a slightly lower contact resistance of around 8 Ω and similar dispersion. It is therefore in the range of good force-fit connections and has slightly better values than the firmly bonded variants of gluing and welding, although welding tends to deliver slightly lower resistances. The solder did not really adhere during soldering and could not be applied reproducibly. Individual measurements yielded values between 15 Ω and 30 Ω.
Among the force-fit connections, bullet connectors show the greatest variation with a similar average contact resistance compared to spring contact pins. By using a CSP, spring contact pins could achieve a resistance level that corresponds to that of screw terminals with a contact pressure of 1.28 MPa. Press-fit of Red-Cubes shows higher variation, but with an average of 7 Ω, they have the lowest average contact resistance of all the variants examined.
3.3. Tensile strength
In addition to electrical contact resistance, the mechanical load capacity of the interface is particularly important for applications. Tensile tests have shown that the mechanical strength of force-fit connections with optimal force transmission can exceed the tensile strength of the polymer. Adhesion can also almost fully exploit the tensile strength of the material if the design parameters (e.g. contact area, load direction, gap) are correct. However, the situation is different with positive fit and penetrating force-fit connections. These reduce the strength of the test specimen due to the cross-sectional changes introduced. The pull-out strength was reduced to approximate 10-25 % of the tensile strength of the polymer. All shear loaded specimens failed as brittle fractures whereas the pull-out specimens failed without damaging the polymer. The results are shown in Table 1.
Results tensile strength of contact approaches for CPC
4. Discussion
The methodical characterisation of the contact variants has shown that a bonding agent can reduce the contact resistance. It is also possible to contact a larger area, thereby further reducing the contact resistance. In addition, the contact resistance is less force-dependent and the interface can be manufactured reproducibly. When selecting the bonding agent, particular consideration must be given to scalability, adhesion to the polymer, restrictions on design freedom, contact resistance and the stress on the polymer during application. The various bonding agents are compared in Table 2, where their properties have been evaluated comparatively.
Comparison table of different bonding agents
++ = excellent; + = good; ○ = adequate; - = bad
Scalability is ensured when no bonding agent and CSP are used. In contrast, masking is very complex in electroplating and sputtering. Adhesion is not a problem in all processes, although sufficient compatibility between the polymer and the conductive paste must be ensured when using CSP. Design freedom is unrestricted without a bonding agent and is limited by accessibility in electroplating, sputtering and CSP, although CSP is slightly easier to apply without masking. Performance is highest in electroplating, but closely followed by the other bonding agents. Without a bonding agent, the contact resistance is higher with a larger contact area (see preliminary test). The thermal load during application is slightly higher with sputtering and electroplating than with CSP due to the risk of thermal overload of the sample by the applied current. Depending on the polymer, the solvent in the CSP can dissolve the surface layer.
The contact resistances of the different contact types are generally similar. High dispersion for imprinting on smooth surfaces results from sensitivity to mechanical stress, which can be reduced by improved anchoring to porous copper mesh. Bullet connectors show higher dispersion and resistance due to lower contact pressure and greater sensitivity to manufacturing tolerances.
The higher contact resistance of spring-loaded contacts likely results from their lower spring rate and thus lower contact pressure. In force-fit connections, increasing contact pressure reduces resistance by enlarging the real contact area. This effect is more pronounced for conductive polymers due to their ductility and the need for sufficient conductive particles at the surface to contact the conventional conductor, requiring a larger real contact area than in purely metallic systems.
In such instances, the deployment of bonding agents proves advantageous, as they facilitate the establishment of a substantial contact area between the conductive polymer and the substrate. This enhancement in contact area is attributed to the efficient wetting of the uneven surface, while concurrently ensuring minimal contact resistance when compared to conventional contact partners. This eliminates the force dependency. It has been established that, above a certain contact pressure, the contact resistance with CSP is higher than in the absence of such paste, for the same apparent contact area. This is due to the fact that, with sufficient force, the same contact area to the conductive particles is present, but with CSP, a further current transition with contact resistance is necessary.
With gluing the contact resistance is slightly higher, presumably because the higher viscosity of the adhesive means that not the entire surface of the polymer is wetted and the conductivity of the adhesive is lower than that of silver conductive paste. Chemically adhesive contacting variants and contact mediators must be compatible with the polymer, as in the case of non-polar polymers (e.g. polypropylene), sufficient adhesion cannot be achieved.
The low contact resistance of the inserts can be explained by the large contact area resulting from the mechanical anchoring lamellae and the good wetting of the contact surface by the polymer melt during fusion. However, the lowest contact resistance was found in the Red-Cubes. During joining, these cut through the edge layer of the polymer, which usually contains slightly less filler Reference Nowka, Ruge, Schulze, Hilbig and Vietor(Nowka et al., 2024). Shearing during the joining process causes more conductive fillers to come into contact with the contact partner at the contact surface. However, the joining process is associated with high mechanical stress, which can damage the sample, and the susceptibility of the fit quality to manufacturing parameters (flow, material diameter, etc.) leads to a large degree of variation.
4.1. Design rules
The results were used to derive the design rules listed in Table 3. These rules should be taken into account when designing electrical contacts on compounds manufactured using MEX. It is conceivable that this could be transferred to alternative materials, such as more flexible polymers. However, it can be hypothesised that the mechanical influence on the contact resistances will be significantly stronger.
Design rules for reduced contact resistance for AM-parts out of CPCs
4.2. Design catalogue
In addition, the results can be used to compile a design catalogue for contact variants of CPCs in the MEX process. The outline section first classifies the connection category, separability and contact type (penetrating or flat). The access section shows the operating principle and working scheme, as well as examples with the respective reference contact resistances from the test results. In addition to the variants examined, wire wrap, clip, wire melting and riveting are also listed. These were not tested because they are either very similar to one of the other contact variants or could not be implemented in the considered design space. The design catalogue is shown in Table 4. To make an informed decision, separability and contact area must first be considered. The specified contact resistance can then be used to select the contact approach.
Design catalogue for electrical bonding methods for AM-parts out of CPCs
5. Conclusion
This study systematically examined nine bonding methods for additively manufactured conductive polymer composites to conventional conductors. Based on experimental results and contact procedures, nine design guidelines and a design catalogue were developed leading to the following key findings:
-
• Type of contact: The type of contact must be adapted to the design. Not every contact approach is suited for every application, due to geometrical, mechanical or electrical limitations.
-
• High contact pressure: The real contact area rises with contact pressure. This has a greater effect than just increasing the apparently contact area. It is limited by the mechanical stability.
-
• Penetration of boundary layer: Penetrating the CPC boundary layer results in direct contact of the conductive particles with the conventional conductor, leading to low contact resistance.
-
• Bonding agent: A bonding agent greatly increases the real contact area, resulting in independence of contact pressure.
-
• Mechanical stress: Mechanical stress can affect contact resistance and must be taken into account during design.
This study shows that all contact types have potential for further optimisation. Future work should focus on welding different CPCs and integrating them into MM-MEX parts with functional structures such as heaters, sensors, and conductor tracks. A more detailed analysis of the relationship between apparent and real contact area using a bonding agent is needed, as well as further investigation of the influence of mechanical loads on contact resistance. In addition, thermal imaging could be used to visualise and mitigate contact resistance effects.
Acknowledgement
This work was supported by the German Research Foundation (DFG) under grant numbers 452679573 and 452009430, and by the German Federal Ministry for Economic Affairs and Climate Action (BMWK) under grant numbers 16KN112736 and KK5325702TA2. We thank Christopher Gassen for acquiring the images of the contact types. The funding sources had no influence on the study. The data, models and figures are available in Figshare: 10.6084/m9.figshare.31242406.
Open access
