Investigation of Thermomechanical and Dielectric Properties of PLA-CA 3D-Printed Biobased Materials

: Renewable dielectric materials have attracted the attention of industries and stakeholders, but such materials possess limited properties. This research focused on studying polylactic acid (PLA)/cellulose acetate (CA) blends produced by 3D printing to facilitate their integration into the electrical insulation field. The dielectric findings showed that a blend containing 40% of CA by weight had a dielectric constant of 2.9 and an electrical conductivity of 1.26 × 10 − 11 S · cm − 1 at 100 Hz and 20 ◦ C while exhibiting better mechanical rigidity in the rubbery state than neat PLA. In addition, it was possible to increase the electrical insulating effect by reducing the infill ratio at the cost of reduced mechanical properties. The differential scanning calorimetry, broadband dielectric spectroscopy, and dynamic mechanical analysis results showed that the PLA plasticizer reduced the energy required for PLA relaxations. These preliminary results demonstrated the benefits of using a combination of PLA, CA, and 3D printing for electrical insulation applications.


Introduction
Dielectric materials are increasingly needed and produced for electronic applications in an increasingly connected world.The global electronic materials market has been estimated at USD 65.7 billion in 2022, with a 6% increase over the next 5 years [1].This fast-paced production also raises issues such as the end of life of such materials.According to Ankit et al., 54 million tons of electronic waste have been produced in 2019, with a forecasted production of 75 million tons by 2030 [2].Current e-waste management solutions are still inadequate, mainly based on landfill and incineration processes, with the associated pollution problems.Biobased materials would be highly desirable to mitigate such problems.These materials are more environmentally friendly than their synthetic counterparts in their manufacture, renewability, and biodegradability properties.According to the European Bioplastics association, 87 thousand tons of biobased materials were used in 2023 by the electronics industry, with an exponential increase predicted [3].
Two of the most well-known biobased polymers are polylactic acid and cellulose acetate.Nakatsuka produced plasticized PLA-coated electrical cables [4].The plasticized PLA showed an electrical conductivity of 10 −12 S•cm −1 , and the cable presented a dielectric strength (E BR ) of around 45 kV with good flexural properties while also showing insufficient plasticizer stability over time.By controlling the alignment of electrospun cellulose acetate (CA) fibers, Meng et al. controlled the CA/aluminum bimetal deformation [5].These devices were used as actuators in a non-contact sensor for detecting the moisture content of a human hand, without showing any degradation in the electrical resistance of the aluminum part.This showed high potential in wearable health monitors and advanced non-contact human-machine interactions.However, despite properties that make them attractive for use in electrical insulation applications as previously shown, these polymers face problems that limit their large-scale use, such as the drastic drop in the mechanical properties of PLA at temperatures above 60 • C [4], and the polar nature of CA [5].One approach to mitigate such problems involves the use of polymer blending.In recent research, Coltelli et al. found that PLA/plasticized CA blends with a high PLA content gave heterogeneous blends with a pronounced tendency to form a fibrous-like morphology [6].In addition, previous research carried out on PLA/plasticized CA showed that there was an interaction between the carbonyl groups of PLA and the hydroxyl groups of CA, greatly increasing the complex viscosity of the blends.In view of the applications of such blends in the dielectric field, Henning et al. proposed a PLA/CA blend filled with zinc pyrophosphate (ZnPP) as a biobased alternative to conventional printed circuit board (PCB) [7].PLA/CA and PLA/CA blended with a ZnPP weight content (W ZnPP ) of 5% presented comparable dielectric properties to FR4 and FR2 glass/epoxy laminates, except for a significative lower dielectric constant and higher surface resistivity.They have been successfully used for PCB demonstrators, but issues remain unsolved, such as a warping effect during the soldering of components on its surface.
Three-dimensional printing is a promising technology in this sector as it allows extremely rapid and economical prototyping compared with conventional processes such as injection molding.The application of 3D printing for PLA/CA blends has been briefly studied for dielectric applications.Previously published results have demonstrated that such blends exhibited the results of dielectric and mechanical properties ranging between PLA and CA, offering design flexibility depending on the desired properties [4].Furthermore, the 3D printing process can be used to directly tailor dielectric and mechanical properties by tuning the infill ratio [8].Even if PLA/CA is a promising biobased polymeric for electronic applications, its application with the fused filament fabrication (FFF) technology is still largely undeveloped, and there is a lack of knowledge for electrical insulation applications.Their dielectric and thermomechanical performances should be optimized.To promote the use of biobased materials, this study proposes investigating the use of PLA/plasticized CA blends by 3D printing for electrical insulation applications to understand the advantages of such materials and their noted limitations.

Experimental Procedures
All materials were oven-dried for at least 4 h at 60 • C for all the processing steps to avoid internal moisture.A first extrusion processing was performed with a SCAMEX 25-20D (SCAMEX, Isques, France).A total of 5 conditions were produced, ranging from neat PLA to a PLA/CA blend containing 40% of CA by weight (named CA-40) and detailed in Table 1.A screw speed of 30 rpm and an extrusion profile temperature of 160-170-180 • C were used according to the preliminary tests.
The 3D filaments were produced with a 3DEVO filament maker (3DEVO, Utrecht, NLD) at the ITHEMM laboratory (Charleville-Mézières, France), with a screw speed of 5 rpm and an extrusion profile of 170-180-180-170 • C. The filament diameter was fixed at 2.85 mm as a compromise between diameter stability and filament quality.The 1.75 mm filaments manufactured were unsuitable for ensuring good printing quality.The fused filament fabrication (FFF) was performed with a SIGMAX R19 3D printing device (BCN3D, Gavà, SPA).The parameters used were resumed in Table 2, and the produced samples were presented in Figure 1.Printing problems have been noted when the filaments were not stored in desiccators, so during the entire printing stage, the filaments were stored in desiccators.All samples achieved good quality without warping or cracking.A glue stick had to be applied to the glass bed to increase adhesion only for CA-30 and CA-40 conditions.Before their characterization, all samples were stored in a desiccator for 3 h.The 3D filaments were produced with a 3DEVO filament maker (3DEVO, Utrech NLD) at the ITHEMM laboratory (Charleville-Mézières, FRA), with a screw speed of rpm and an extrusion profile of 170-180-180-170 °C.The filament diameter was fixed 2.85 mm as a compromise between diameter stability and filament quality.The 1.75 m filaments manufactured were unsuitable for ensuring good printing quality.
The fused filament fabrication (FFF) was performed with a SIGMAX R19 3D printin device (BCN3D, Gavà, SPA).The parameters used were resumed in Table 2, and the pr duced samples were presented in Figure 1.Printing problems have been noted when t filaments were not stored in desiccators, so during the entire printing stage, the filamen were stored in desiccators.All samples achieved good quality without warping or crac ing.A glue stick had to be applied to the glass bed to increase adhesion only for CAand CA-40 conditions.Before their characterization, all samples were stored in a desicc tor for 3 h.

Characterization Methods
All the figures presented below were produced using the Python 3.10.11program ming language and the Pandas' module to read the dataset and Matplotlib to produce t different graphs.

Characterization Methods
All the figures presented below were produced using the Python 3.10.11programming language and the Pandas' module to read the dataset and Matplotlib to produce the different graphs.

Differential Scanning Calorimetry (DSC) Analysis
Differential scanning calorimeter (DSC) analyses were proposed to observe the thermal properties and the crystallinity of the materials.The measurements were performed on a DSC 214 Polyma (Netzsch, Selb, Germany) with 15 mg of material, inserted in Al pans.The used protocol consisted of one heating and one cooling ramp between 20 and 180 • C with a heating rate of 10 K•min −1 and an argon flow of 20 mL•min −1 .The crystallinity content (X cr ) were determined according to Equation (1) developed for PLA/CA blends [9]: where dH is the measured crystallization enthalpy, dH 0 is the melting enthalpy of the theoretical 100% crystalline polymer 100%, both expressed in J•g −1 .W PLA is the weight percentage of PLA.

Dynamic Mechanical Analysis (DMA)
DMA was carried out on 60 × 10 × 2 mm and 60 × 10 × 4 mm samples with an Artemis DMA 242 E (Netschz, Selb, Germany).The double cantilever method was chosen over the 3-point bending method due to the very low stiffness of PLA once its glass transition temperature is reached (T > 60 • C).The protocol consisted of a ramp between 30 and 140 • C at 5 • C•min −1 .The controlled force has been set at 1 Hz, which is the frequency usually used in the literature.

Broadband Dielectric Spectroscopy (BDS) Analysis
BDS tests were carried out using a Keysight E4980A Precision LCR Meter (Agilent Technologies, Santa Rosa, CA, USA) on 25 mm disk samples.The heating program consisted of an isothermal rise between 0 and 160 • C with 4 • C per step and measured at 100 Hz.The dielectric constant ε ′ is expressed according to Equation (2) [10]: where C p is the electrical capacitance given in Farad, S is the cross-sectional area of the sample (in m 2 ), ep is the distance between electrodes (in m), and ε 0 is the vacuum permittivity (given as 8.541878 × 10 −12 F•m −1 ).The electrical conductivity σ AC can be determined with Equation (3) [10]: where σ AC is expressed in S•m −1 , ω is the angular frequency, and f is the applied electrical frequency (in Hz).

Results
Figure 2 presents the thermal properties of neat PLA and PLA/CA blends.Neat PLA exhibited all the expected relaxations for a PLA: a glass temperature (T g ) of 61.6 • C, a cold crystallization temperature (T cc ) of 126.8 • C, and a melting temperature (T m ) of 156.3 • C [11].Adding CA decreased all the observed relaxations.At W CA = 40%, the T g , T cc , and T m of PLA decreased by 21.3, 23.9, and 8.5 • C, respectively.Kang et al. have also observed a decrease in T g by adding plasticizers into PLA [11].According to the latter, adding plasticizers, generally small molecules, increased the free volume of macromolecular chains, thereby improving the polymer's mobility.It is also worth noting that the supplied PLA was amorphous, with a measured X cr below 1%.Furthermore, while adding the plasticizer improved the mobility of the macromolecular chains, it did not significantly change the crystallinity of the blends, with X cr still below 1%.The only visible change is the presence of a second slight melting peak at around 142 • C, proof that a second crystal lattice appeared.
As the DSC findings demonstrated, the addition of CA altered the thermal properties of PLA by lowering the relaxation energies.Dielectric materials were subjected to various temperature fluctuations due to electric current, which could affect their dielectric properties.An understanding of the thermal dielectric stability would be important to qualify the presented materials.Figure 3 presents the dielectric constant (ε ′ ) and electrical conductivity (σ AC ) of the 3D-printed neat PLA and CA-40 with an infill ratio of 100% versus temperature.At 20 • C, PLA and CA-40 presented a ε ′ of 2.53 and 2.81, respectively, and a σ AC of 4.8 × 10 −13 and 1.4 × 10 −12 S•cm −1 , respectively.We thought that the increase in dielectric properties was attributed to the polar nature of CA, which could contain more hydroxyl functions than PLA [5].Although the addition of CA increased the electrical conductivity of the blends, the measured properties at 100 Hz for all materials were still in the electrically insulating range [12], validating their use for dielectric applications.However, the produced materials showed notable temperature sensitivity.At 160 • C, PLA exhibited a ε ′ and σ AC of 1.92 and 503 times greater, respectively, and CA-40 exhibited a ε ′ and σ AC of 1.85 and 145 times greater, respectively, than measured at 20 • C. The observations for neat PLA are in perfect agreement with the literature [5,13].Badia et al. have also noticed that measured at 100 Hz, the dielectric constant of neat PLA increased from 3.05 to 3.95 when the temperature shifted from 20 to 130 • C [13].The temperature-dependent behavior observed in both electrical conductivity and dielectric constant can be associated with the increased ionic mobility of macromolecular chains, as mentioned in the literature [5].As the melting temperature approached, the σ AC increased faster for CA-40 than neat PLA.This could also be due to the increased free volume, greatly improving the mobility of the carrier charges [14].CA reduced the energy required for α-relaxation measured by BDS, as observed in the literature when a plasticizer is added to a plastic matrix.CA-40 showed an α-relaxation shift of −20 • C. While ε ′ and σ AC were slightly higher for CA-40 than for PLA in the measurement range tested, these results remained comparable and low enough to be suitable for electrical insulation applications.As the DSC findings demonstrated, the addition of CA altered the thermal properties of PLA by lowering the relaxation energies.Dielectric materials were subjected to various temperature fluctuations due to electric current, which could affect their dielectric properties.An understanding of the thermal dielectric stability would be important to qualify the presented materials.Figure 3 presents the dielectric constant (ε′) and electrical conductivity (σAC) of the 3D-printed neat PLA and CA-40 with an infill ratio of 100% versus temperature.At 20 °C, PLA and CA-40 presented a ε′ of 2.53 and 2.81, respectively, and a σAC of 4.8 × 10 −13 and 1.4 × 10 −12 S•cm −1 , respectively.We thought that the increase in dielectric properties was attributed to the polar nature of CA, which could contain more hydroxyl functions than PLA [5].Although the addition of CA increased the electrical conductivity of the blends, the measured properties at 100 Hz for all materials were still in the electrically insulating range [12], validating their use for dielectric applications.However, the produced materials showed notable temperature sensitivity.At 160 °C, PLA exhibited a ε′ and σAC of 1.92 and 503 times greater, respectively, and CA-40 exhibited a ε′ and σAC of 1.85 and 145 times greater, respectively, than measured at 20 °C.The observations for neat PLA are in perfect agreement with the literature [5,13].Badia et al. have also noticed that measured at 100 Hz, the dielectric constant of neat PLA increased from 3.05 to 3.95 when the temperature shifted from 20 to 130 °C [13].The temperature-dependent behavior observed in both electrical conductivity and dielectric constant can be associated with the increased ionic mobility of macromolecular chains, as mentioned in the literature [5].As the melting temperature approached, the σAC increased faster for CA-40 than neat As the field of dielectric materials covers applications requiring high flexibility or high rigidity, it is also important to know the thermomechanical behavior of the proposed materials.PLA suffers from an important loss of mechanical properties beyond its T g due to its rubbery state.As the incorporation of plasticized CA into PLA will result in a complex behavior, it is important to understand its behavior [5]. Figure 4 presents the storage modulus (E ′ ) and loss factor (tan δ) of the 3D-printed neat PLA and CA-40 with an infill ratio of 100% versus temperature.PLA also presented a typical amorphous behavior.A rigid glassy behavior is noted at T < 60 • C, with an E ′ value of 1.81 GPa at 30 • C. Once reaching the T g , the storage modulus sharply decreased, reaching its lowest value at around 96 • C. At this temperature, the E ′ reached 4.0 MPa, i.e., a reduction of 450 times in stiffness between 30 and 96 • C. Once the minimum has been reached, the mechanical rigidity increased due to the formation of the crystal lattice resulting from the cold crystallization of PLA.The addition of CA slightly reduced the E ′ of PLA, with an E ′ of 1.69 GPa obtained for CA-40, i.e., a mechanical rigidity reduction of 7%.At 30 • C, PLA and CA-40 showed an E ′ of 1.81 and 1.69 GPa, respectively.Although the addition of CA in the glassy state decreased the E ′ and increased the tan δ, the CA addition appeared attractive from a mechanical point of view, as it seems to greatly limit the loss of mechanical properties when PLA reached the rubbery state.This improvement effect can be directly linked to the CA and not to an improvement in the blend's crystallinity, as demonstrated by DSC.Moreover, the tan δ was much lower in the rubbery state, so less electromagnetic energy was converted into heat.It is also worth noting that for both DMA and BDS measurements, the addition of CA decreased the α-relaxation and cold crystallization temperature of PLA, in agreement with DSC analysis.As the field of dielectric materials covers applications requiring high flexibility or high rigidity, it is also important to know the thermomechanical behavior of the proposed materials.PLA suffers from an important loss of mechanical properties beyond its Tg due to its rubbery state.As the incorporation of plasticized CA into PLA will result in a complex behavior, it is important to understand its behavior [5]. Figure 4 presents the storage modulus (E′) and loss factor (tan δ) of the 3D-printed neat PLA and CA-40 with an infill ratio of 100% versus temperature.PLA also presented a typical amorphous behavior.A rigid glassy behavior is noted at T < 60 °C, with an E′ value of 1.81 GPa at 30 °C.Once reaching the Tg, the storage modulus sharply decreased, reaching its lowest value at around 96 °C.At this temperature, the E′ reached 4.0 MPa, i.e., a reduction of 450 times in stiffness between 30 and 96 °C.Once the minimum has been reached, the mechanical rigidity increased due to the formation of the crystal lattice resulting from the cold crystallization of PLA.The addition of CA slightly reduced the E′ of PLA, with an E′ of 1.69 GPa obtained for CA-40, i.e., a mechanical rigidity reduction of 7%.At 30 °C, PLA and CA-40 showed an E′ of 1.81 and 1.69 GPa, respectively.Although the addition of CA in the glassy state decreased the E′ and increased the tan δ, the CA addition appeared attractive from a mechanical point of view, as it seems to greatly limit the loss of mechanical properties when PLA reached the rubbery state.This improvement effect can be directly linked to the CA and not to an improvement in the blend's crystallinity, as demonstrated by DSC.Moreover, the tan δ was much lower in the rubbery state, so less electromagnetic energy was converted into heat.It is also worth noting that for both DMA and BDS measurements, the addition of CA decreased the α-relaxation and cold crystallization temperature of PLA, in agreement with DSC analysis.One of the main benefits of 3D printing is the ability to control the infill ratio of 3D specimens.Zhang et al. demonstrated that the dielectric constant was proportionally related to the infill ratio, which can be valuable for producing materials with controlled dielectric properties [15].To investigate this effect on the mechanical and dielectric properties of PLA/CA blends, Figure 5 shows the ε′, σAC, and E′ of CA-20, CA-30, and CA-40 One of the main benefits of 3D printing is the ability to control the infill ratio of 3D specimens.Zhang et al. demonstrated that the dielectric constant was proportionally related to the infill ratio, which can be valuable for producing materials with controlled dielectric properties [15].To investigate this effect on the mechanical and dielectric properties of PLA/CA blends, Figure 5 shows the ε ′ , σ AC , and E ′ of CA-20, CA-30, and CA-40 blends.The ε ′ and E ′ decreased with a decreasing infill ratio, while the σ AC increased with decreasing infill ratio.In the case of CA-40, shifting from 100% to 40%, the infill ratio decreased the dielectric constant and storage modulus by 29% and 61%, respectively.The σ AC was, however, increased by 216%.The decrease in ε ′ and E ′ with a decreasing infill ratio can be directly related to the porosity rate of the samples [15][16][17].Regarding the variation in the dielectric constant with the infill ratio, two studies in the literature can provide information on the involved mechanism.Figure 6 shows the ε ′ versus infill ratio of the work by Zhang et al. and Colella et al., using 3D-printed PLA [15,18].It can be noted that the theoretical model was linear for both studies.Lowering the infill ratio logically produced less dense samples that were, therefore, less likely to be polarized, as the dielectric constant of air approaches 1 [8].It is worth noting that linear regression applied to literature data gives a value very close to 1, validating this statement.The increase in the electrical insulating properties associated with the reduction in the infill ratio can be linked to more conductive paths, enabling the electrical current to bypass areas of high local resistance due to greater local heterogeneity.These results demonstrate that improving the insulating properties of PLA/CA blends would be possible by reducing the infill ratio at the cost of lower mechanical stiffness.To compare these new materials with their synthetic counterparts, Table 3 presents a comparative analysis of the tested properties of CA-40 and LDPE as reported in the literature.These results show that CA-40 with 40% infill had a storage modulus and electrical conductivity equivalent to LDPE, with a lower dielectric constant than LDPE.This would prove that the 3D printing of PLA/CA blends is relevant to replace dielectric applications usually dominated by LDPE, such as electrical insulation, cable insulation, and antistatic and electromagnetic shielding.To compare these new materials with their synthetic counterparts, Table 3 presents a comparative analysis of the tested properties of CA-40 and LDPE as reported in the literature.These results show that CA-40 with 40% infill had a storage modulus and electrical conductivity equivalent to LDPE, with a lower dielectric constant than LDPE.This would To compare these new materials with their synthetic counterparts, Table 3 presents a comparative analysis of the tested properties of CA-40 and LDPE as reported in the literature.These results show that CA-40 with 40% infill had a storage modulus and electrical conductivity equivalent to LDPE, with a lower dielectric constant than LDPE.This would prove that the 3D printing of PLA/CA blends is relevant to replace dielectric applications usually dominated by LDPE, such as electrical insulation, cable insulation, and antistatic and electromagnetic shielding.

Conclusions
In this study, we investigated the dielectric and thermomechanical properties of 3Dprinted PLA/CA blends to determine the benefits and limitations of such materials for electrical insulation applications.The CA content and infill ratio were the main criteria in the dielectric and mechanical response of PLA/CA blends.We highlighted that adding CA improved the thermomechanical stability of PLA in the rubbery state.Moreover, electric insulation performances were improved by reducing the infill ratio of the 3D-printed samples.Under the chosen conditions, the infill ratio had a greater effect on the mechanical and dielectric properties than the effect of CA, showing that adjusting the dielectric properties of PLA/CA blends by adjusting these crucial parameters is possible.Our results show that these PLA/CA blends are promising for insulation electrical applications, at the cost of a decrease in mechanical properties in the glassy state and a slight decrease in electrical insulating properties.To go further, it would be possible to analyze in depth the influence of post-crystallization on the mechanical stability.Furthermore, to limit the plasticizing effect of CA, we can reduce the plasticizer content to fully benefit from the insulating effect of PLA and the improved mechanical stability at high temperatures provided by CA.This research also opens the way to numerous research opportunities.Investigations could be pursued by introducing micro or nanosized biobased fillers to increase thermomechanical stability at high temperature.In addition, 3D printing can be exploited for the creation of functional gradient structures, as well as customized design for specific dielectric applications, such as high-voltage insulation and antistatic and electromagnetic shielding.

Figure 3 .
Figure 3. Dielectric constant ε′ and electrical conductivity σAC of PLA and CA-40 versus temperature, measured at 100 Hz.The main arrows indicate the axes of the corresponding curves.

Figure 3 .
Figure 3. Dielectric constant ε ′ and electrical conductivity σ AC of PLA and CA-40 versus temperature, measured at 100 Hz.The main arrows indicate the axes of the corresponding curves.J. Compos.Sci.2024, 8, x FOR PEER REVIEW 7 of 10

Figure 4 .
Figure 4. Log (storage modulus E′) and loss tangent tan δ of PLA and CA-40 specimens versus temperature.The main arrows indicate the axes of the corresponding curves.

Figure 4 .
Figure 4. Log (storage modulus E ′ ) and loss tangent tan δ of PLA and CA-40 specimens versus temperature.The main arrows indicate the axes of the corresponding curves.

10 Figure 5 .
Figure 5. (a) Dielectric constant ε′, (b) electrical conductivity σAC and (c) log (storage modulus E′) as a function of infill ratio measured on 4 mm thick samples.Dielectric constant and electrical conductivity are measured at 100 Hz and 20 °C, and storage modulus is measured at 30 °C and 1 Hz.

Figure 5 . 10 Figure 5 .
Figure 5. (a) Dielectric constant ε ′ , (b) electrical conductivity σ AC and (c) log (storage modulus E ′ ) as a function of infill ratio measured on 4 mm thick samples.Dielectric constant and electrical conductivity are measured at 100 Hz and 20 • C, and storage modulus is measured at 30 • C and 1 Hz.

Table 1 .
Produced conditions PLA/CA blends (W PLA correspond to PLA content by weight and W CA correspond to CA content by weight).

Table 1 .
Produced conditions PLA/CA blends (WPLA correspond to PLA content by weight and W correspond to CA content by weight).

Table 3 .
Comparative analysis of the tested properties of CA-40 and LDPE as reported in the literature.