2D characterisation and evaluation of multi-material structures towards 3D hybrid printing

ABSTRACT Multi-material manufacturing through the hybridisation of printed electronics and additive manufacturing has gained great interest recently. However, such hybridisation attempts are not trivial due to the need for functional material development and compatible process identification, as well as further performance understanding, comprehensive characterisation and long-term reliability evaluation of multi-material parts. While some multi-material structures from functional materials such as silver inks have been demonstrated via the integration of direct writing systems into stereolithography or material extrusion platforms, the performance assessment and characterisation of parts manufactured using such integrated systems is still required. Therefore, this research presents a comprehensive assessment of multi-material structures manufactured using syringe deposition and material extrusion platforms. Test specimens were subjected to various characterisation activities such as thickness measurement, resistance measurement, roughness tests, wettability measurement, adhesion tests, and morphological analysis. Results and statistical analyses suggested that the dry thickness and conductivity of deposited films were dependent on the substrate material. Adhesion between the conductive film and substrate was affected by both substrate material and ink deposition angle. Also, the interaction of conductive films with polycarbonate substrate was found to be noticeably better among all substrates due to low resistivity and enhanced adhesion at low thicknesses.


Introduction
Printed electronics (PE) is an emerging research area due to its potential over traditional electronics manufacturing (Khan et al. 2020;Kamyshny and Magdassi 2014;Perelaer et al. 2010;Gamota et al. 2004).Such manufacturing advantages include, but are not limited to; the ability to manufacture cost-efficient, thin, lightweight, flexible, and environmentally sensitive electronics in a non-subtractive way (Perelaer et al. 2010;Gamota et al. 2004;Subramanian et al. 2008).One common method for PE manufacturing is direct writing (DW) which is defined as a group of processes enabling the precise and additive deposition of functional materials onto flat or conformal substrates across pre-defined locations without the use of subsequent etching processes.Also, DW could be split into two printing approaches: droplet-based (noncontact) such as inkjet printing (IP) and aerosol jetting (AJ), and flow-based (contact) such as dispensing printing (Hon, Li, and Hutchings 2008;Lewis and Gratson 2004;Zhang, Liu, and Whalley 2009).
As the substrate is an integral part of the final product in the field of PE (Hon, Li, and Hutchings 2008), there has been recent interest in multi-material hybrid manufacturing, which employs the direct writing of functional materials into substrates manufactured via additive manufacturing (AM) for end-use applications such as smart additive manufactured textiles, protective structures with health monitoring capabilities, etc.With such applications, the polymer-based substrate can be fabricated using AM technology such as stereolithography (SLA), material extrusion (ME), material jetting (MJ), and powder bed fusion (PBF), while the aforementioned DW methods can be used to precisely deposit functional materials to realise tracks, interconnects, or sensors housed within the substrate (Macdonald et al. 2014)therefore resulting in a fully additive manufactured multi-material structure with embedded electronics.
Initial proposals for hybrid DW-AM systems to manufacture multi-material products date back to the 1990s (Weiss et al. 1997;Prinz and Weiss 1998;Beck et al. 1992;Cham et al. 1999;Kataria and Rosen 2001).These proposals demonstrated the manufacture of such structures through the integration of DW apparatus into SLA systems, enabling the direct writing of functional inks into polymer structures manufactured via SLA (Macdonald et al. 2014;Church, Fore, and Feeley 2000;Periard, Malone, and Lipson 2007;Robinson et al. 2006;Palmer et al. 2006;Navarrete et al. 2007;Castillo et al. 2009;DeNava et al. 2008;Lopes, MacDonald, and Wicker 2012;Jang et al. 2015;Maalderink et al. 2018;Li et al. 2016;Wasley et al. 2016;Malone and Lipson 2007).Additionally, the use of ME for substrate manufacture is also gaining interest due to its ease of hybridisation to a DW system (Espalin et al. 2014a(Espalin et al. , 2014b;;Vogeler et al. 2013;Goh et al. 2018;Carranza et al. 2019;Lee et al. 2019;Wałpuski and Słoma 2021;Esfahani et al. 2018).However, one concern about using substrates manufactured via ME is the rougher surface finish of the substrates when compared to traditional PE substrates such as glass, polyethylene naphthalate, and polyimide (Vogeler et al. 2013;Sridhar, van Dijk, and Akkerman 2009;Hoerber et al. 2014).Post-processing activities such as sanding can be applied to such substrates for a smooth surface finish so that a suitable surface roughness for the PE application is met.However, this also results in an extra manufacturing step and time.Even so, it was reported that the reliability of such demonstrations was not significantly different from that of traditionally manufactured electronics (Wałpuski and Słoma 2022).In addition to the utilisation of substrates manufactured via SLA and ME, other AM manufacturing methods have also been employed for substrate manufacture for PE, for example, the deposition of pastes directly onto substrates manufactured via MJ (Goh et al. 2018;Wałpuski and Słoma 2021;Perez and Williams 2014;Chang et al. 2015;Paulsen et al. 2012) and PBF (Hoerber et al. 2014;Folgar, Folgar, and Cormier 2013).
As presented, the multi-material hybrid manufacture of PE integrated into polymer-based AM structures has been developed to some extent.However, to our knowledge, there is limited literature exploring the interaction and performance characterisation of such multi-material structures, especially those manufactured using DW and ME.This research, therefore, focuses on the mechanical and electrical performance, and the characterisation of multi-material specimens manufactured using DW and ME at the 2D stage in order to characterise properties at a fundamental level so that findings will inform progressing to the more complex surfaces and 3D geometries.

Test specimen design and material
Test specimens incorporated two features: a material extruded substrate and a conductive film that was deposited onto the substrate.Substrates measured 50.00 × 50.00 mm in length and width with a thickness of 0.50 mm.Four tabs were also added to the central area of the substrate to facilitate clamping during testing.On the other hand, the square conductive film pattern was aligned as so at the centre on the top surface of the substrate with a measurement of 20.00 mm on each edge and a wet thickness set at 50.00 µm.Multi-material specimens were designed using SolidWorks® (Dassault Systemes, Waltham, MA, U.S.A.) as illustrated in Figure 1.
For the manufacture of the substrates, four lowcost filament materials (Ultimaker's polycarbonate (PC), acrylonitrile butadiene styrene (ABS), polylactic acid (PLA), and tough PLA (TPLA)) were used since they are a selection of the most common and fundamental materials used in ME (Wickramasinghe, Do, and Tran 2020).Additionally, a commercially available silver paste (Sun Chemical C2180423D2) was chosen as the conductive film material due to desirable material properties such as low resistivity.Sun Chemical's C2180423D2 is mainly formulated to be used with screen printing applications, however, based on the information gained from the manufacturer and the literature, it could also be suitably employed within a DW dispensing system that can accommodate the same ink formulations and properties as screen printing (Hassan et al. 2020).The main material properties of the silver paste utilised are listed in Table 1.

Test specimen manufacture
The manufacture of the multi-material test specimens was completed across two steps using two separate processes.While ME via an Ultimaker 3 3D printer (Ultimaker BV, Utrecht, The Netherlands) was used for the manufacture of the substrates, the fabrication of the conductive films was performed at three different deposition angles on a Hyrel System 30M platform (Hyrel3D, Norcross, Atlanta, U.S.A.).
For characterisation and performance tests, five replicas per deposition angle were fabricated, equating to 15 specimens per substrate material, and an additional specimen per deposition angle and substrate material was also manufactured for morphological analysis purposes, totalling 72 specimens.It should also be noted that the specimen manufacture was randomised to minimise any uncontrollable variables that might occur during fabrication.Fabrication of the multi-material test specimens is outlined in terms of both substrate and conductive film manufacture, as follows.

Substrate manufacture
All substrates were manufactured using process parameters previously established by Cicek, Southee, and Johnson (2022) with the following exceptions.Because the fabrication of the conductive structures directly onto material extruded parts having higher roughness due to the nature of their untreated surfaces was not practical, attempts were made with the intention of providing a smoother surface finish with fewer imperfections with a view to enhance the conductive film's mechanical reliability and electrical performance, which included (Bellacicca, Santaniello, and Milani 2018;Buj-Corral, Bagheri, and Sivatte-Adroer 2021;Wankhede et al. 2020;Yang et al. 2020): . Setting the layer thickness as 0.05 mm since the surface profile of material extruded parts depends mainly on layer thickness .Reducing the substrate print speed to 20.00 mm/s and layer width to 0.04 mm so a better surface for the conductive film in terms of wettability and adhesion is provided.
In addition to those mentioned, no further treats or post-processing was applied to the substrate surfaces.The final print parameters of substrates depending on the material are presented in Table 2.

Conductive film manufacture
Slic3r open-source slicing software was utilised to finalise slicing settings for the dispensing of the conductive paste, as summarised in Table 3.It should also be noted that preliminary trials were conducted to ascertain optimal ink deposition parameters such as amount, speed, and z-height.
Apart from those settings summarised in Table 3, all other pre-process parameters were set to default values as outlined by the Hyrel manufacturer (Reservoir Heads, n.d.).
Substrates cleaned using isopropanol were then clamped on the Hyrel printer using a piece of masking tape at the centre of the build platform, and a small amount of ink was then allowed to be dispensed to ensure that ink flow had reached a steady state before dispensing onto the substrate.Conductive paste deposition was then completed at three different deposition angles of 0, 45, and 90°relative to the layer direction of the top surface of the substrate as demonstrated in Figure 2.
Following the deposition, specimens were placed inside a thermal incubator (GenLab INC/50/DIG, Wolflabs, York, UK) for two hours to facilitate curing.The curing time was kept as same for all substrates to minimise the effect of the curing time on film conductivity as curing time was not evaluated within the scope of this experimental study.While Greer and Street (2007) showed that silver inks are to be cured at 150 °C, in this study, specified curing temperatures were based on the substrate material's glass transition temperature (T g ) compatibility according to the manufacturer datasheets.Curing temperatures for the deposited films were 100 °C for PC and ABS, 55 °C for PLA, and  60 °C for TPLA.Following the curing of conductive films, specimens were left to cool to room temperature.

Dry thickness
While the pre-print thickness of the conductive films, i.e. wet thickness, within this study was set at 50.00 µm, it is known that ink solvents disappear by evaporation during curing causing a reduction in the wet thickness set.Additionally, different material extruded substrates absorb different amounts of wet ink due to their different degrees of surface properties such as roughness, porosity, and surface wetting, which has also been shown to lead to variability in the wet ink thickness of such conductive films (Kattumenu et al. 2009;Merilampi, Laine-Ma, and Ruuskanen 2009).That is why the determination of film thickness after drying was essential for this study both to understand how different ME materials could be used as substrates to accommodate PE, but also in order to assess the quality of the conductive films.
For thickness measurements, a digital thickness gauge (Mitutoyo Thickness Gauge 547-320S Absolute) having 0.01 mm accuracy was used.Since the dimensional accuracy of parts manufactured via ME could also vary depending on factors such as material choice, process parameters, etc., (Boschetto and Bottini 2014), substrate thickness was first measured across four different points and averaged.The total thickness of specimens (combined thickness of the substrate and conductive film) was then measured from five different points and averaged.The conductive film thickness for each specimen was then calculated by subtracting the mean substrate thickness from the mean total thickness.The thickness measurement points are illustrated in Figure 3.

Sheet resistance
The electrical performance of the printed silver films was evaluated by calculating the sheet resistance using a four-point resistance measurement device (Jandel HM21, Jandel Engineering Limited, Leighton Buzzard, Bedfordshire, UK).The supplied current to the conductive films was set to 10 mA and the resultant voltage difference between probes was then measured.Measurements were taken at the same five locations as the attained thickness measurements presented in Figure 3.The measured voltages across the five tested locations were averaged to ensure accuracy and used to calculate the sheet resistance (R s ) according to Equation 1 (Smits 1958): where; c f : the correction factor; V: the mean voltage difference measured between probes (mV); I: the current supplied (mA).
The correction factor was found to be 3.73 according to the information provided by Smits (1958).

Surface roughness
The electrical performance and mechanical reliability of deposited conductive films are dependent on substrate surface topography.For example, the rougher substrate surfaces increase the sheet resistance of conductive films while improving adhesion (Chou, Chen, and Lai 2016;Mikkonen and Mäntysalo 2018;Öhlund et al. 2012).In addition, the surface roughness of material extruded substrates has been shown to be greater than traditional PE substrates.Therefore, defining the surface textures of the selected material extruded substrates was important to understand their surface characteristic's effect on the conductive film performance and integrity.
Non-contact (optical) surface roughness measurement of specimens was therefore performed using a 3D optical profiling system (Talysurf CLI 2000, Taylor Hobson, Leicester, UK).Roughness scans were completed at 1.00 mm/s speed with a scanning frequency of 500.00Hz and a measuring range of 3.00 mm.During the optical surface roughness measurements, the roughness data of both substrates and conductive films were collected in both horizontal and vertical directions for a total sampling length of 40.00 mm-20.00mm from the central axis in each direction, as shown in Figure 4. Analysis of the collected data was completed using the TalyProfile Silver software package (version 8.2, Taylor Hobson, Leicester, UK) and two roughness values (R a ) were obtained by taking the mean of two scanning directions for each of the substrates and the conductive films.

Contact angle
An optical tensiometer (Theta Lite, Biolin Scientific, Manchester, UK) was used for the surface wettability measurement of the substrates with an accuracy of ±0.10°.Type I pure water as defined in ASTM D1193-06 (2018) was selected as the dispensing liquid for the contact angle measurements.PURELAB® Elga water purification system (ELGA LabWater, High Wycombe, UK) was used to purify the dispensing water.Analysis of the contact angle was conducted in situ using OneAttension software (Version 1.8, Biolin Scientific,  Manchester, UK) which captured images every 0.01 s through 10 s of resting time while the water droplet spread over the substrate.Since the contact angle value may vary from one point to another because of the material extruded substrate's surface profile, four measurements of the substrates were averaged to eliminate measurement errors.Contact angle measurement points were the same locations as the thickness measurement locations as indicated by the blue circles in Figure 3.

Mechanical reliability
Evaluation of the mechanical reliability of PE, i.e. adhesion, is crucial in reliable electrical design and one convenient method for such assessment is via a crosscut tape test.In this experiment, test method B of ASTM D3359-17 (2017) was used to determine the adhesion level of conductive films to material extruded substrates.
Eleven cross-cuts in both horizontal and vertical directions were performed on each conductive film.For peel purposes, a piece of pressure-sensitive tape (3M 8981) having 7.10 N/cm peel force was used.A visual check of the removed tape and lattices using a magnifier and illumination device was performed.According to the pattern left on the substrate, the adhesion grade was assigned by referring to the standard classification, in which 5B and 0B represent the highest level of adhesion and the worst adhesion grade, respectively.It should be also noted that all tests were conducted at room temperature and 50% relative humidity since extremes in temperatures or relative humidity may have affected the adhesion of the tape or conductive films.

Morphological analysis
Scanning electron microscopy (SEM) images were taken for the microstructural evaluation of the multi-material structures on a JSM-7800F model Schottky field emission scanning electron microscope (JEOL, Tokyo, Japan).As cross-sectional monitoring is a key parameter determining the electrical and mechanical performance of fabricated structures (Happonen et al. 2016), specimens were cross-sectioned as so 1.00 × 1.00 mm from their back face using a surgical scalpel with the intention of not damaging the silver films that might have otherwise affected the visual image's accuracy.The cross-sectioned specimens were coated by sputtering a thin layer of gold (80% Au) and palladium (20% Pd) blend in an argon environment for 90 s using a rotary pumped coater (Q150R S, Quorum, East Sussex, UK).Then, the Au/Pd coated specimens were placed in the microscope and SEM was performed by taking magnified images at an accelerating voltage of 5.00 kV.

Statistical analysis
The gathered dry thickness, sheet resistance, and adhesion data were analysed using GraphPad Prism (version 9.4.1,GraphPad Software Inc., San Diego, CA, US), with P-values less than 0.05 considered as the significance threshold for analysed data.Data presented as means with standard deviation were tested for normality using the Shapiro-Wilk test due to the small sample size (n < 50).Normally distributed data were analysed in terms of statistical significance using twoway analysis of variance (ANOVA) with substrate material and film deposition angle as the two independent factors (3 × 4; substrate material x deposition angle).Tukey's multiple comparison test was also performed to test differences among all specimen groups.

Silver film: dry thickness
The thickness of the conductive films was measured after drying to ascertain whether there are any effects on the print quality and accuracy as a result of the different substrates useddue to variations in their material and process propertiesand silver film deposition angles.The mean dry thickness of specimens depending on the ink deposition angle is presented in Figure 5.
Two-way ANOVA test revealed that conductive films displayed statistically significant differences in their dry thickness depending on the substrate material (p < 0.0001).However, the ink deposition angle had no significant impact on dry thickness (p = 0.3718).Further multiple comparisons among substrate materials with the Tukey test demonstrated that there was no significant difference (p > 0.05) between PC and ABS (p = 0.9895) substrates and between TPLA and PLA (p = 0.8660) substrates in terms of dry film thickness.
Further assessment of Figure 5 in light with the statistical analyses also showed that the PC substrate accommodated the thinnest silver films at all print orientations with mean results of 19.85 µm at 0°, 25.11 µm at 45°, and 22.84 µm at 90°by a reduction of more than 50% of the 50.00 µm ink wet thickness among all substrates.Similarly, silver films on ABS substrate were slightly thicker (as previously shown statistically insignificant) than their PC counterparts -20.82 µm dry thickness at 0°, 25.27 µm at 45°, and 24.77 µm at 90°.However, compared to PC and ABS, deposited films on PLA and TPLA substrates were relatively thicker at around 35 and 40 µm, respectively.
Such variation in the thicknesses of deposited films across the different substrate materials may be due to several factors as shown in previous works such as: . The varying curing temperatures of the silver suspensions depending on their substrate's T g compatibility.This may have an effect on the evaporation rate and speed of the binder, dispersants, and solvent, which could increase with a rise in curing temperature (Bidoki et al. 2007), therefore resulting in a thinner dry thickness. .The severity of the surface roughness and porosity across substrates, therefore, allowing the wet ink to be absorbed into porosities at different rates (Kattumenu et al. 2009;Xie et al. 2012).
Thickness measurement results revealed that the dry thickness of the deposited films significantly varies depending on substrate material choice rather than the ink deposition angle.

Silver film: sheet resistance
The sheet resistance of the silver films was calculated and the mean results with standard deviation are presented in Figure 6.
Regarding the impact of the substrate material type and the ink deposition angle on the sheet resistance of conductive films, two-way ANOVA showed that substrate material type (p < 0.0001) had a significant impact on the sheet resistance of conductive films while ink deposition angle (p = 0.2117) had no significant effect.Further, Tukey's multiple comparisons demonstrated that no significant differences (p > 0.05) between conductive films on PC and TPLA (p = 0.2237) substrates in terms of sheet resistance was found.However, multiple comparisons among other possibilities illustrated that there were significant differences between other comparison pairs (p < 0.0001 for all possible comparisons).
It is apparent from Figure 6 that the mean sheet resistances of the printed films on PC were 20.87, 18.86, and 20.61 mΩ/sq for printings at 0, 45, and 90°d egrees respectively, while 26.97, 25.99, and 26.16 mΩ/sq were recorded for silver films on TPLA substrate.However, films on ABS and PLA had more sheet resistivity with their calculated sheet resistance of 70.17, 64.30, and 66.18 mΩ/sq, and 45.83, 36.36, and 44.12 mΩ/sq for 0, 45, and 90°, respectively.
It was also observed that 0°of ink printing on PLA substrates demonstrated greater deviation due to the third replica of conductive films on PLA specimen almost doubling the mean value, while it is believed the rest showed an acceptable and negligible degree of deviation to some extent.This was thought to be due to the amount of ink on this substrate could have been thinner than other films on PLA at 0°printing, and so affected the resistance of the ink layer.However, when the aforementioned specimen thickness was checked, no variation was seen, which was concluded that the reason for the higher resistance demonstrated by this specimen was not related to its dry thickness.In this case, there were other factors to consider affecting the resistance such as possible dust particles left on the substrate or non-homogeneity of the ink mixture since these were done by hand.
The assumption, as also mentioned by Tan, Tran, and Chua (2016), was that the thicker deposition of conductive pastes would lower sheet resistance.So, it was expected that the overall resistance of the silver patterns on PC among all other substrates would have the highest sheet resistance since they were the thinnest structures with a mean of 19.85, 25.11, and 22.84 µm for 0, 45, and 90°respectively, while silver coatings on TPLA could have the least resistance as they were the thickest with a mean of 33.75, 38.13, and 37.92 µm for 0, 45, and 90°respectively, as previously depicted in Figure 5.However, it is clear from Figure 6 that the silver films printed on  PC substrates demonstrated the lowest overall sheet resistance and therefore the highest conductivity, followed by TPLA (no significant difference found compared to PC), PLA, and ABS, respectively.This supports the work by Merilampi, Laine-Ma, and Ruuskanen (2009) where the thickness of dispensed paste has no remarkable influence on the mean sheet resistance of conductive films.There were possible explanations for this conflict such as resistance to high curing temperatures, surface features, and wettability of such substrate materials.Further visual examination of the silver film-PC specimens under SEM is therefore presented in Figure 7.
The SEM image in Figure 7 illustrates that PC substrates had porous-like gaps between the top surface layers filled by the silver suspension, which decreased the overall film thickness measured from the substrate surface, while maintaining high conductivity and so support the presented low resistivity findings in films on PC specimens.Occurrences of the porous substrates absorbed the paste into branched paths beneath the substrate surface, therefore creating the conductive phase both underneath and on the substrate and consequently decreasing the overall thickness from the substrate surface while maintaining conductivity.This finding also aligns with previously reported works (Kattumenu et al. 2009;Xie et al. 2012;Kim, Kim, and Jung 2012;Barmpakos, Tsamis, and Kaltsas 2020;Southee et al. 2007;Hay et al. 2005).Therefore, it is important to set the wet thickness so that for the dried film it has enough conductive materials left on the surface to be conductive following the ink penetration into the impurities of the substrate.Otherwise, it could cause the printed patterns not to be fully conductive or create open circuits since there is not enough material left on the substrate surface to make conductive paths.Therefore, within this investigation, as the wet thickness was appropriate, it is clear that the porous structures and high curing temperature of the PC substratesrelative to other substrateshelped silver films to maintain their conductivity with thinner structures compared to other substrates.
On the other hand, films on ABS substrateswhich were also cured at 100 °C as same as films on PChad the highest sheet resistance indicating poor conductivity.As mentioned in the literature, higher curing temperatures could cause warping on other parts of PE such as substrates (Espera et al. 2019).As demonstrated in Figure 8, this issue was witnessed with ABS substrates which were warped from the corners upon curing.
Also, to further aid to understand the effect of substrate warping on the conductive film layer, SEM captures of the conductive film-ABS specimens were investigated as presented in Figure 9.
As seen in Figure 9, the thickness of the deposited films overall increased towards the centre of the substrate, which was a result of the warping on the ABS substrate that pushed the conductive ink towards the centre where the least warping occurred, and therefore may have caused those particles to be agglomerated around the substrate centre during the wet-to-dry transition phase.Hence, the silver particles could not be uniformly distributed because the point-to-point contact between them varied across the substrate which resulted in them being less conductive near to edges and displaying greater sheet resistance.This also highlights the limitation of polymer material choices to be employed as substrates in the field of PE due to their limited resistance to excessive temperatures.
Considering the low bulk resistivity of the silver paste used which would enable low resistance of printed structures, films on ABS and PLA substrates that had high resistivities were shown not to perform well as a material extruded substrate candidate which is essential for PE applications.However, PC and TPLA substrates demonstrated a promising performance as substrates, but further optimisation between substrate porosity and film wet thickness is still required.

Surface characterisation
The surface characterisation of material extruded substrates and printed films was performed to define the specimen's surface roughness.The mean R a values are presented in Figure 10.
It is apparent from Figure 10 that PC substrates had the greatest surface roughness with a R a value of 9.52 µm followed by TPLA with 9.35 µm.ABS substrates presented the lowest R a with a mean of 2.73 µm, while PLA substrates demonstrated a mean roughness value of 7.21 µm.However, these roughness values of substrates are still relatively high for PE applications compared to traditional substrates.This makes the control of ink deposition difficult because the rougher substrate surfaces allow for the wet inks to seep through valleys between AM traces of the surface via the capillary effect, which could result in printed patterns not being functional due to open circuits or discontinuities if the pre-print thickness is not set thick required enough.This is because contacts between conductive particles varies from location to location depending on the substrate surface roughness profile, i.e. valleys have more conductive material while peaks have less and so a uniform conductive material distribution across the substrate is not possible.Hence, the silver particles could not make uniform paths which might result in having possible open circuits.
Furthermore, the printed films were evaluated in terms of their surface roughness to assess if they varied after drying.The presented results in Figure 10 could be concluded that the roughness of the printed films was not affected by material extruded substrate surface roughness since their mean R a values are very close to each other.On the contrary, it is believed to be more dependent on other factors such as the chemical composition of the ink, additives, nanoparticle size, etc.However, it could be suggested that films on rougher surfaces tend to be slightly rougher than films on less rough surfaces.For example, films on PC still were the roughest with 3.86 µm followed by patterns on TPLA, PLA, and ABS with 3.72, 3.65, and 3.42 µm, respectively.Even so, it is believed that these printed conductive patterns are too rough to be classified as of good enough quality for industrial application, and therefore process optimisation is needed.

Substrate: wettability
Significant variations in the wettability of material extruded substrates across four materials were witnessed.As demonstrated in Figure 11, PC substrates had the lowest contact angle indicating better wettability compared to other substrates, and the mean contact angle of PC reduced from 32.25°to 19.62°within 10 s of resting time.
ABS followed the PC with a slightly higher mean contact angle reduced from 46.92°to 31.39°,but the deviation between ABS substrates was the greatest.However, the contact angles of water drops on PLA and TPLA substrates were relatively higher with 63.87°t o 61.88°for PLA and 68.92°to 67.53°for TPLA.It seems that the resting time did not affect PLA and TPLA substrate's wettability much since their contact angle profiles reached equilibrium far quicker than PC and ABS.Contact angle captures for each substrate at 0, 5, and 10 s are shown in Figure 12.
The print settings used in this study provided an acceptable level of surface energies (all smaller than 90°) for substrates that helped substrates demonstrate hydrophilic behaviours, but it is believed that this could be improved further with modification of process parameters of substrates such as layer thickness, print temperature, and speed.

Mechanical bonding
Mechanical bonding between deposited films and substrates was deduced visually that the greater amount  of silver remaining on the substrate after pull-off, the greater bonding of the film to the substrate.The mean results are presented in Table 4, whereas 5B and 0B show the highest and the worst adhesion grade, respectively.
The two-way ANOVA statistical analysis demonstrated that there was a statically significant interaction between structural integrity of specimens and both selected factors, substrate material type (p = 0.0056) and the print angle (p = 0.0099).
Regarding the substrate material, films on the PC substrate showed overall the best adhesion performance at all print angles with almost no ink removal (less than 5%) ranging between 4.2B and 4.6B among all specimens.This could be attributed to the rougher surface and the better wettability of the PC substrate, in which the better contact angle implied more contact area and ink penetrated the impurities of the PC substrate and made a strong bond resulting in films on PC having the strongest mechanical bonding, thus better adhesion.This finding also aligns with the literature in that there may be a positive relationship of surface roughness and wettability on ink bonding (de Gennes, Brochard-Wyart, and Quéré 2004;Caglar, Kaija, and Mansikkamaki 2009;Cheng, Dunn, and Brach 2002).
Following that, films on TPLA and PLA demonstrated overall an acceptable level of adhesion ranging from 3.8B to 4.4B for TPLA and 3.8B to 4.2B for PLA.Even if TPLA demonstrated a R a value as nearly as PC had, the adhesion performance of films on TPLA was slightly weaker than films on PC.The most likely explanation for the difference in adhesion between films on PC and TPLA is inherent differences in their surface wettability since films on TPLA substrate could not make a contact area with the substrate as films on PC as due to lower wetting.It was also found that patterns on ABS substrates were the poorest in terms of adhesion with around 5-15% ink removal even if ABS substrates had good wettability, but a smoother surface.
Regarding deposition angle, parallel ink printing relative to the substrate top surface provided the best  adhesive performance with all substrate materials with a minimum 4.0B adhesion demonstrated by films on ABS.
As seen from Table 4, 0°printing of the silver paste on the PC substrates demonstrated almost no ink removal with a mean of 4.6B adhesion showing an extremely strong mechanical bonding.While printing with 45°to the substrate resulted in the poorest adhesion, 90°printing of films to the substrates also demonstrated an acceptable level of adhesion which was slightly better than 45°deposition.
Results and analysis of the adhesion tests revealed that the structural integrity of such multi-material structures is significantly dependent on the substrate material and film deposition angle.While PC substrate was the best among all substrates in terms of adhesion, printing parallel relative to the substrate's top surface was found to be the best deposition orientation.
There are also a number of points to address regarding the structural integrity of such structures.Firstly, the stand-off distance between the deposition nozzle tip and the target substrate is key, as if it is not set properly adhesion/structural integrity would be a problem.If the stand-off distance is set too high, this would cause delamination of the printed film from the substrate as it does not have required constant pressure to support adhesion on the substrate during deposition.Secondly, the low-temperature resistance of such polymer structures restricts curing temperatures and accordingly the bonding force, as a higher curing temperature (around T g of the substrates) is required for better bondings.When using two separate platforms for multi-material printing, the transferring of the substrate from Ultimaker printer to Hyrel was manually performed due to the use of two separate platforms rather than a hybrid platform, as such may result in misalignment of the substrate and the deposited printed films.Additionally, during such transfer between platforms, an opportunity exists for dust particles from the local environment to be located on the substrate potentially influencing overall integrity.

Conclusion
In this study, multi-material test specimens were fabricated using a silver ink deposited on low-cost material extruded substrates manufactured across a range of materials via two separate AM processes and subjected  to a number of experimental tests in order to characterise the mechanical and electrical performances of such 2D specimensan important step towards functionally integrated 3D multi-material structures.The results and statistical analyses have identified a set of outcomes that could be used for the fabrication of such multi-material structures including: . Film dry thickness and conductivity were significantly affected by the substrate material because material extruded substrate features such as roughness, porosity, and wettability were found to be the most important parameters to influence ink transfer consequently the PE's conductive performance.Ink transfer was the greatest with substrates with high levels of roughness and low surface energy resulting in thinner films, but maintaining high conductivity. .Film deposition angle had no significant effect on film dry thickness and conductivity. .The structural integrity of multi-material structures was significantly affected by both substrate material and film deposition angle.Results identified that a rough and hydrophilic surface is desired in PE-AM integration so that it could provide a better adhesion performance.Also, parallel printing (0°) relative to the substrate top surface is recommended for a reliable integration due to improved adhesion performance compared to 45 and 90°of printing. .There was no direct relationship between substrate surface roughness and conductive film roughness, but it could be said that films on rougher surfaces tend to be slightly rougher. .A major constraint in PE-AM integration is the use of polymer substrates that limits the curing temperature of conductive paste and consequently electrical performance due to low heat deflection and T g of such materials as warping occurred on ABS substrates.
PC interactions with silver paste were found to be noticeably stronger than the other polymers substrates investigated.This is because films on PC demonstrated better conductivity with thinner structures as well as the best structural integrity among all specimens with good repeatable results.This was a result of PC surface properties since it was the roughest with the best wettability.
It should however be noted that within this investigation: . The effect of curing time on the film's dry thickness, conductivity, and adhesion was not within the scope of this experiment.
. The effect of deposition parameters on ink deposition and electrical properties was not assessed. .Substrate choices were limited to four of the most common and low-cost ME materials and only one type of ink, as such, further opportunities for investigation are available. .Substrate roughness and contact angle measurements were not dependent on directions, rather an average value was used to calculate these properties. .The adhesion evaluation method used was highly relative since it was performed by hand-peeling and was highly reliant on the skill of the operator rather than being mechanically repeatable. .The findings of this paper are from results of 2 dimensions and therefore applicability of such structures into various 3D scenarios needs further work.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Figure 1 .
Figure 1.Schematic of multi-material test specimen design.

Figure 3 .
Figure 3. Substrate and conductive film thickness measurement locations.

Figure 7 .
Figure 7. SEM analysis of a conductive film on PC substrate.

Figure 8 .
Figure 8.Comparison of the substrates upon curing.

Figure 9 .
Figure 9. SEM image of a conductive film on ABS substrate.

Figure 10 .
Figure 10.Mean surface roughness values of specimens.Figure11.Contact angles of the substrates.

Table 1 .
Material properties of Sun Chemical's C2180423D2.

Table 4 .
Adhesion test results of multi-material specimens with the standard deviation.