Experimental investigation and optimization of fused filament fabrication parameters for enhanced mechanical performance of PETG and carbon fiber-reinforced PETG composites

Abstract

This study investigates the influence of key additive manufacturing parameters on the mechanical performance of polyethylene terephthalate glycol (PETG) and carbon fiber-reinforced PETG (CF-PETG) composites fabricated through fused filament fabrication (FFF). Eighteen specimens were printed with varying infill densities (50%, 75%, 90%), infill patterns (gyroid, cubic, triangle), printing speeds (40–50 mm/s), and layer thicknesses (0.1–0.3 mm). Mechanical testing showed that CF-PETG printed at 90% infill, triangular pattern, 45 mm/s speed, and 0.1 mm layer thickness achieved superior compressive strength, recording 47.21 MPa at 10% and 61.56 MPa at 50% deformation-a 42% improvement over the best PETG counterpart. Flexural strength of CF-PETG printed at 75% infill, cubic pattern, 50 mm/s, and 0.1 mm layer thickness increased by 23%, reaching 59.2 MPa, while tensile strength gains remained marginal. In contrast, PETG printed at 90% infill, triangular pattern, 45 mm/s, and 0.1 mm layer thickness exhibited higher impact strength (2.55 kJ/m²) than CF-PETG (1.44 kJ/m²), reflecting a stiffness–toughness trade-off. Design of Experiments analysis identified infill density as the most influential factor for both materials. PETG performance improved with increasing infill and pattern complexity, whereas CF-PETG reached optimal properties at moderate speeds and simpler patterns, likely due to sensitivity in fiber dispersion and matrix bonding. Interaction plots showed that parameter combinations significantly affect tensile and flexural behaviour, with CF-PETG exhibiting more irregular responses due to complex fibre–matrix interactions. Optimization using desirability functions produced composite desirability values of 0.8518 for PETG and 0.9872 for CF-PETG, confirming the superior performance potential of fibre-reinforced systems when appropriately tuned. Overall, carbon fiber reinforcement enhances mechanical properties, especially compression and modifies process-property relationships, requiring more precise control of FFF parameters for structural applications.

Introduction

Additive manufacturing (AM), especially Fused Deposition Modelling (FDM), has emerged as a transformative technology in engineering, enabling the rapid fabrication of complex, customized parts with reduced material waste and shorter production cycles. The success and performance of FDM are significantly influenced by the material selection, as the thermoplastic filament used determines not only the mechanical properties of the printed part but also its printability, dimensional accuracy, and long-term durability.^1,2

Traditionally, materials such as Polylactic Acid (PLA), Acrylonitrile Butadiene Styrene (ABS), and Nylon have been the mainstays in FDM printing due to their commercial availability and ease of processing. ³ PLA, a biodegradable thermoplastic derived from renewable sources, is known for its ease of printing, low warpage, and good dimensional stability. However, it suffers from brittleness, low thermal resistance, and limited mechanical strength, making it less suitable for functional or load-bearing applications. ⁴ ABS, on the other hand, offers better toughness and heat resistance than PLA but is prone to warping and requires a heated build platform and enclosed chamber to mitigate thermal deformation during printing. ⁵ Nylon offers excellent mechanical strength and flexibility but is highly hygroscopic and demands precise storage and handling conditions. Furthermore, its processing requires high extrusion temperatures and presents challenges in achieving dimensional accuracy and strong interlayer bonding. ⁶

In the quest for materials that combine the printability of PLA, the toughness of ABS, and the durability of Nylon, Polyethylene Terephthalate Glycol (PETG) has emerged as a balanced and versatile alternative.^7,8 PETG is a glycol-modified version of Polyethylene Terephthalate (PET), designed to eliminate the brittleness and enhance the processability of the base polymer. It exhibits good impact resistance, chemical resistance, and durability, with minimal shrinkage and warping during printing. PETG offers an excellent middle ground-easy to print like PLA, with superior strength and toughness akin to ABS, and with less moisture sensitivity compared to Nylon. Its semi-amorphous structure results in smooth prints and better interlayer adhesion, making it increasingly popular for functional prototyping, enclosures, structural parts, and even biomedical and food-grade applications. ⁹

While PETG meets the requirements of many industrial applications, the demand for higher performance in structural and functional components-particularly in automotive, aerospace, and defence sectors-has driven the need for reinforced polymer composites. ¹⁰ Among the various reinforcements available, short carbon fibers (CF) are widely adopted due to their high strength-to-weight ratio, stiffness, and thermal stability. The inclusion of carbon fibers into a PETG matrix results in carbon fiber reinforced PETG (CF-PETG) composites, which exhibit significant improvements in mechanical strength, stiffness, dimensional stability, and thermal resistance. ¹¹ These composites maintain the printability advantages of PETG while offering enhanced load-bearing capacity, making them suitable for applications where mechanical performance is critical. ¹²

Additive Manufacturing (AM), particularly Fused Deposition Modeling (FDM), has emerged as a versatile technique for fabricating complex geometries with reduced material wastage and design flexibility. ¹³ Among the wide range of thermoplastics and composite filaments available for FDM, the mechanical and functional performance of the printed parts is significantly influenced not only by the choice of material but also by the printing parameters employed during fabrication. Parameters such as layer thickness, raster angle, infill density, printing speed, and nozzle temperature govern the degree of inter-layer adhesion, porosity, residual stresses, and surface finish, which in turn dictate the overall strength, durability, and reliability of the printed structures. ¹⁴

Optimization of these parameters is essential to balance the trade-offs between mechanical performance, surface quality, build time, and material usage. ¹⁵ Previous studies have demonstrated that variations in process parameters can lead to substantial changes in tensile, flexural, and impact properties, highlighting the importance of systematic optimization. For instance, higher infill densities improve load-bearing capacity but increase weight and print time, while finer layer thickness enhances surface finish at the expense of longer fabrication cycles. Similarly, nozzle temperature and printing speed strongly affect polymer flow behavior and interfacial bonding, directly influencing part anisotropy and defect formation.^16–18

Recent work on FDM-printed PLA reinforced with milled carbon fiber (MCF) showed that print speed and layer thickness strongly influence tensile strength and stiffness. Optimized parameters significantly improved interlayer bonding and fracture behavior, underscoring the importance of process control for achieving high-performance PLA-MCF composites. ¹⁹

Recent investigations on FDM-based PLA composites reinforced with milled carbon fibers have demonstrated that print parameters such as temperature, layer thickness, infill pattern, and speed significantly influence flexural and compressive behavior. Advanced predictive models using Levenberg-Marquardt and Scaled Conjugate Gradient algorithms achieved high accuracy (R² > 0.98), effectively correlating process parameters with mechanical performance. ²⁰

The study on thermal, structural, and mechanical properties of carbon fiber reinforced PLA composites: influence of FDM print speed and comprehensive analysis explores how print speed affects the performance of FDM-printed carbon fiber reinforced PLA (CFRPLA) composites. It reveals that higher print speeds increase crystallinity but compromise thermal stability and mechanical strength due to greater porosity and fiber–matrix debonding. While stiffness and strength decrease, ductility and toughness improve. The findings highlight a trade-off between crystallinity and porosity, emphasizing the need to optimize print speed for achieving balanced thermal and mechanical properties in high-quality CFRPLA components. ²¹

The study on the relationship between 3D printing parameters, porosity, and mechanical behavior in a carbon fiber reinforced polymer investigates how 3D printing parameters-specifically print speed, temperature, and orientation-affect the porosity and mechanical behavior of carbon fiber reinforced polymer composites. Results show that porosity increases with higher print speed but decreases with increased temperature and optimized orientation. A strong link between porosity and Young’s modulus is established, indicating that higher porosity generally reduces stiffness. The optimal conditions for minimal porosity and improved mechanical performance were found at a printing temperature of 260°C and a 0° printing orientation. ²²

In recent years, numerous optimization approaches such as Design of Experiments (DOE), Response Surface Methodology (RSM), Taguchi techniques, and Artificial Intelligence (AI)-driven models have been employed to identify parameter combinations that maximize performance while minimizing drawbacks.^2,23,24 Such methods not only reduce experimental effort but also provide predictive insights into the complex relationships between process parameters and material response. Despite significant progress, achieving an optimal balance remains challenging due to the multi-objective nature of the problem, which often involves conflicting requirements such as maximizing strength while minimizing weight and fabrication time.^25–28

However, the full potential of PETG and CF-PETG composites can only be realized through optimal control of the FDM process parameters. Parameters such as nozzle temperature, bed temperature, layer height, print speed, infill density, and raster orientation play a crucial role in determining the quality and mechanical properties of the printed parts. Improper parameter settings can lead to poor layer adhesion, void formation, fiber misalignment, and surface defects, which collectively degrade the mechanical performance. For fiber-reinforced composites like CF-PETG, the complexity increases due to the influence of fibers on melt flow behavior and interlayer bonding, necessitating a more refined optimization strategy.

Although several studies have investigated the influence of printing parameters on common FDM materials like PLA and ABS, comprehensive studies focusing on PETG and its fiber-reinforced variant CF-PETG are limited. More importantly, there is a lack of systematic frameworks that utilize design of experiments (DOE), statistical analysis, or multi-objective optimization techniques to establish the relationships between printing parameters and mechanical performance for these materials.

In this context, the present study aims to optimize the FDM printing parameters for both PETG and CF-PETG composites to enhance their mechanical properties, including tensile strength, flexural strength, and impact resistance. Using a structured experimental approach such as the Taguchi method, the study identifies the most influential process parameters and determines their optimal combinations. Statistical tools like Analysis of Variance (ANOVA) are used to quantify the contribution of each parameter and to understand the interaction effects, particularly in the context of fiber reinforcement.

This research contributes to the growing body of knowledge in high-performance additive manufacturing by establishing a process-property-performance relationship for PETG and CF-PETG composites. The insights gained will aid in the production of lightweight, strong, and dimensionally stable components tailored for engineering applications. By addressing both the material and process dimensions, the study sets a foundation for reliable, repeatable, and high-quality manufacturing using PETG-based composite filaments in FDM technology.

Materials and methods

Materials used

In this study, two filament types were selected for investigation: neat Polyethylene Terephthalate Glycol (PETG) and carbon fiber-reinforced PETG (CF-PETG). The PETG filament, with a nominal diameter of 1.75 mm and a density of approximately 1.27 g/cm³, was procured from Tesseract. The CF-PETG filament, reinforced with 10 wt% short chopped carbon fibers, had a slightly higher density of ∼1.3-1.35 g/cm³ and was obtained from the same supplier. Prior to printing, both materials were dried in a convection oven at 60°C for 6 hours to eliminate any moisture content that could negatively affect extrusion quality, surface finish, and mechanical performance. The PETG and CF-PETG filaments used in the present study are presented in Figures 1 and 2 respectively.

Figure 1.

PETG and CF-PETG filaments used in the present study.

Figure 2.

3D printer used for printing the specimen.

The properties of PETG and CF-PETG are provided in Table 1.

Table 1.

Properties of PETG and CF-PETG.

Particulars	PETG	CF-PETG
Density (g/cm3)	1.27	1.3
Tensile strength (MPa)	50–60	60–75
Tensile modulus (GPa)	2.0 GPa	5–7 GPa
Flexural strength (MPa)	70–80	100–120
Flexural modulus (GPa)	2–2.5	6–7.5
Impact strength (kJ/m2)	8–10	5–7
Elongation at break (%)	6–8	1.5–3

3D Printing equipment and experimental setup

All specimens were fabricated using a commercial-grade Fused Deposition Modeling (FDM) 3D printer equipped with a 0.4 mm brass nozzle and a heated build platform. The slicing of CAD models was done, where parameters such as layer thickness, infill density, Raster angle and printing speed were carefully controlled. A raster angle of 45° was selected to achieve balanced mechanical properties and consistent layer bonding in the printed specimens. This orientation was chosen because it provides a balanced distribution of stresses along both the longitudinal and transverse directions, leading to improved inter-road bonding and enhanced mechanical integrity of the printed specimens. Moreover, several studies have reported that a 45° raster angle minimizes anisotropy and promotes uniform load transfer under mechanical loading. Printing was conducted under controlled ambient conditions (25°C and ∼50% relative humidity) to minimize environmental variation. The same setup was used for both PETG and CF-PETG to maintain consistency across all experimental runs. Figure 2 shows the 3D printer used and Figure 3 shows the printed specimen for various mechanical testing.

Figure 3.

Printed coupons for mechanical testing.

Experimental design and parameter selection

A Taguchi L9 orthogonal array was employed to systematically investigate the impact of key FDM parameters on mechanical performance. The study considered four factors: layer thickness, infill density, pattern, and printing speed, each at three levels, as presented in Table 2. These parameters were chosen based on their significant influence on bonding, porosity, and fiber alignment in printed parts. This design yielded nine experimental conditions per material, totalling 9 runs. This structured approach allowed efficient identification of optimal parameter combinations with minimal experimental effort. Table 3 provides Taguchi’s L9 Orthogonal array used to study the effect of factors on the mechanical properties of the developed composite.

Table 2.

Factors and their levels used to print the specimen.

Factors	Level 1	Level 2	Level 3
Infill density (A)	50	75	90
Pattern (B)	Gyroid	Cubic	Triangle
Printing speed (C)	40	45	50
Layer thickness (D)	0.1	0.2	0.3

Table 3.

Taguchi’s L9 OA used in the present study.

Run number	Infill density	Pattern	Printing speed	Layer thickness
1	50	Gyroid	40	0.1
2	50	Cubic	45	0.2
3	50	Triangle	50	0.3
4	75	Gyroid	45	0.3
5	75	Cubic	50	0.1
6	75	Triangle	40	0.2
7	90	Gyroid	50	0.2
8	90	Cubic	40	0.3
9	90	Triangle	45	0.1

Dimensions of the specimens used for tensile, flexural, compressive, impact and hardness testing along with the printed samples are presented in Figure 4.

Figure 4.

Schematic and printed samples of (a) Tensile; (b) Flexural; (c) Compression and (d) Impact specimen.

Mechanical characterization

Mechanical properties are assessed through tensile test, flexural test, compression test and impact test.

Tensile testing

Tensile properties of the developed composite specimens were evaluated in accordance with ASTM D638 (Type I) as presented in Figure 5. The specimens were prepared with standard dimensions and tested using a Universal Testing Machine (UTM). The tests were performed at a crosshead speed of 5 mm/min under ambient laboratory conditions. During the test, parameters such as Ultimate Tensile Strength (UTS), Young’s Modulus, and Elongation at Break were recorded. UTS was determined from the maximum load sustained before fracture, Young’s Modulus was calculated from the initial linear region of the stress-strain curve, and Elongation at Break was measured as the strain corresponding to the point of specimen failure. These values provided critical insights into the composite’s load-bearing capacity, stiffness, and ductility.

Figure 5.

Specimen subjected to tensile testing.

Flexural testing

Flexural properties were assessed using the three-point bending test in line with ASTM D790 standards and presented in Figure 6. The test specimens were positioned over a support span of 64 mm, and the load was applied at the midpoint using a UTM equipped with a suitable bending fixture. The crosshead speed was maintained at 2 mm/min to ensure controlled loading. From the resulting load-displacement data, the Flexural Strength and Flexural Modulus were computed. Flexural Strength indicated the material’s resistance to bending-induced failure, while Flexural Modulus reflected its stiffness under flexural loading conditions.

Figure 6.

Specimen subjected to flexural testing.

Impact testing

The impact resistance of the composites was evaluated using the Charpy impact test, as specified in ASTM D6110 as shown in Figure 7. Notched specimens were used to focus stress and ensure consistent crack initiation. The test involved striking the specimen with a pendulum hammer to determine the amount of energy absorbed during fracture. The energy absorption values obtained provided a measure of the composite’s toughness and its ability to resist sudden impact loads, which is crucial for applications subjected to dynamic or accidental loading conditions.

Figure 7.

Charpy impact testing facility.

Compression testing

Compression testing of the developed polymer composites was performed in accordance with ASTM D695 to assess their compressive strength, stiffness, and failure characteristics and the testing arrangement is presented in Figure 8. The ASTM D695 standard is extensively used for evaluating neat polymers as well as particulate- and short-fiber-reinforced composites, ensuring reliable comparison across material systems. For this purpose, specimens were prepared in a rectangular geometry with smooth and parallel surfaces to promote uniform stress distribution and avoid premature failure due to stress concentrations. The tests were conducted using a universal testing machine (UTM) under uniaxial compressive loading, with the applied load delivered at a constant crosshead displacement rate as prescribed by the standard. This procedure enabled accurate measurement of the composites’ load-bearing capacity and facilitated the identification of their characteristic compressive failure modes.

Figure 8.

Compression testing facility.

Scanning electron microscopy (SEM)

Fracture surfaces from tensile and impact specimens were gold-coated and examined using SEM to assess the dispersion of fillers, interfacial adhesion, and failure mechanisms such as fiber pull-out, matrix cracking, or void formation.

Results and discussions

The overview of the mechanical properties of the 3D printed PETG and CF-PETG at varied printing parameters is presented in Table 4.

Table 4.

Overview of mechanical properties.

Composite	Infill density	Pattern	Printing speed	Layer thickness	Tensile strength (MPa)	Flexural strength (MPa)	Compression strength (MPa)			Impact strength (kJ/m2)
							10% deformation	25% deformation	50% deformation
P1	50	Gyroid	40	0.1	21	42.8	7.64	11.25	16.1	0.79
P2	50	Cubic	45	0.2	19.2	44.5	13.27	12.16	15.68	1.3
P3	50	Triangle	50	0.3	29.4	52.1	22.64	23.22	27.35	1.04
P4	75	Gyroid	45	0.3	30.4	52.6	24.02	23.38	34.4	1.86
P5	75	Cubic	50	0.1	30.3	56.5	25.91	27.66	37.79	1.42
P6	75	Triangle	40	0.2	24.5	53.5	30.63	29.92	37.86	1.12
P7	90	Gyroid	50	0.2	33.7	54.7	30.47	31.62	43.44	1.3
P8	90	Cubic	40	0.3	30.5	58.5	28.46	30.3	43.04	1.31
P9	90	Triangle	45	0.1	22	49.1	32.17	35	43.29	2.55
CF1	50	Gyroid	40	0.1	25.9	57.6	22.25	28.27	42.58	1.01
CF2	50	Cubic	45	0.2	21.8	41.1	17.93	24.26	37.8	0.95
CF3	50	Triangle	50	0.3	20.8	38.9	18.62	26.06	36.58	0.91
CF4	75	Gyroid	45	0.3	25.8	42.8	24.38	36.4	58.38	1.37
CF5	75	Cubic	50	0.1	32	59.2	30.91	37.35	55.72	1.33
CF6	75	Triangle	40	0.2	23.3	44.1	26.32	34.4	51.41	0.98
CF7	90	Gyroid	50	0.2	28.9	48.3	34.42	43.44	66.62	1.13
CF8	90	Cubic	40	0.3	24	40.6	25.58	34.26	55.24	1.03
CF9	90	Triangle	45	0.1	29.2	55.4	47.21	48.57	61.56	1.35

Tensile strength

Tensile strength reflects the ability of a material to resist stretching under uniaxial load, and in 3D-printed composites, it is strongly influenced by infill density, printing pattern, layer thickness, and the addition of reinforcement. The results presented in Figure 9 showed a clear trend of increasing tensile strength with increasing infill density for both PETG and CF-PETG specimens. For example, PETG P1 with 50% gyroid infill recorded only 21 MPa, whereas PETG P7 with 90% gyroid infill achieved 33.7 MPa. This improvement is attributed to the reduction of voids and improved continuity of load paths at higher densities, which enhances stress transfer between the polymer layers. A similar effect was observed for carbon fiber reinforced PETG, where CF5 at 75% cubic infill reached 32 MPa compared to CF2 at 50% cubic infill, which remained at 21.8 MPa.

Figure 9.

Variation in tensile strength of 3D printed PETG and CF-PETG.

The choice of infill pattern also had a significant influence. Triangular and cubic patterns generally produced higher tensile strengths than gyroid patterns at the same density. For example, PETG P3 with a triangular infill reached 29.4 MPa compared to PETG P1 with a gyroid infill, which achieved only 21 MPa at the same 50% density. The triangular and cubic patterns provide more direct load transfer paths, which reduce stress concentration and enhance stiffness. In contrast, the gyroid infill, with its wavy architecture, is more suitable for energy absorption rather than tensile loading.

Layer thickness was another important parameter. Specimens with thinner layers (0.1 mm) exhibited better tensile strength due to stronger inter-layer adhesion and reduced voids at the interfaces. PETG P5 with 0.1 mm layers exhibited a tensile strength of 30.3 MPa, compared to PETG P4 at a 0.3 mm layer thickness, which had a similar density but slightly reduced interlayer adhesion. Larger layer thickness increases the risk of poor bonding between successive layers, leading to premature crack initiation during tensile loading.

Carbon fiber reinforcement further improved tensile strength in most cases, particularly at medium to high densities. CF5 (75% cubic infill) showed higher tensile strength (32 MPa) compared to its PETG counterpart P5 (30.3 MPa). The improvement arises from the ability of short carbon fibers to act as stress bridges, transferring load effectively across the matrix and delaying crack propagation. However, at low density, such as CF2 (50% cubic infill), the benefit of reinforcement was limited because voids dominated the failure mechanism and reduced the effectiveness of fiber–matrix load transfer. Overall, tensile strength was maximized with high infill density, triangular or cubic infill patterns, thinner layers, and the incorporation of carbon fiber reinforcement.

The fracture surface analysis of the tensile-tested specimens (P1-P9) as presented in Figure 10 reveals distinct failure mechanisms governed by infill density, pattern, and layer thickness. At lower infill (P1-P3, 50%), the specimens exhibited brittle fracture with clear signs of inter-layer separation and void formation, more pronounced in P2 and P3 due to cubic and triangular patterns combined with thicker layers, which weakened bonding. In P4-P6 (75% infill), the fracture surfaces showed relatively denser failure features with reduced voids, indicating improved load-bearing capacity; among them, P5 and P6 displayed progressive crack propagation with better adhesion, while P4 revealed mixed brittle–ductile behavior due to its thicker layers. At higher infill (P7–P9, 90%), the fracture morphology was markedly more compact, with evidence of ductile tearing and strong resistance before final failure. P7 exhibited uniform tearing attributed to the gyroid pattern, P8 showed brittle cracking and incomplete bonding due to larger layer thickness, and P9 revealed dense shear bands and delayed failure, signifying superior inter-layer adhesion and toughness. Overall, the transition from brittle to ductile-dominant fracture with increasing infill density and finer layer thickness underscores the crucial role of printing parameters in governing tensile failure behavior.

Figure 10.

Tensile-tested specimens P1–P9 with corresponding fracture surfaces.

The fracture surface analysis of CF-PETG specimens (CF1-CF9), as presented in Figure 11, highlights the combined influence of infill density, infill pattern, layer thickness, and printing speed on both micropore formation and tensile failure mechanisms. The size and distribution of micropores were found to be strongly dependent on the selected printing parameters. At lower printing speeds and finer layer thicknesses, the extruded filaments experienced adequate heat transfer and inter-layer fusion, resulting in smaller and fewer micropores with smoother fracture surfaces. In contrast, higher printing speeds and thicker layers led to incomplete bonding and rapid solidification, producing larger and irregular micropores that acted as crack initiation sites and reduced the effective load-bearing area.

Figure 11.

Tensile-tested specimens CF1–CF9 with corresponding fracture surfaces.

At 50% infill (CF1–CF3), weak inter-layer bonding and the presence of coarse micropores contributed to a brittle-dominant fracture, although CF1 exhibited partial ductile tearing due to finer layer deposition. CF2 and CF3 showed irregular cracking and larger voids, consistent with poor adhesion from thicker layers. With 75% infill (CF4–CF6), the fracture surfaces appeared denser and better consolidated, reflecting improved load transfer and reduced microporosity. CF5 and CF6 demonstrated evident fiber pull-out and compact tearing, indicating enhanced reinforcement efficiency, whereas CF4 displayed mixed brittle–ductile failure due to its relatively larger layer thickness. At 90% infill (CF7–CF9), the fracture morphology became markedly compact with minimal micropores, extensive ductile tearing, and strong fiber-matrix interactions. CF7 showed uniform tearing associated with the gyroid pattern, CF8 revealed localized brittle cracking due to bonding inconsistencies at 0.3 mm layers, and CF9 exhibited dense fiber pull-out and ductile fracture, confirming superior inter-layer adhesion and tensile resistance. Overall, the observed reduction in micropore size and transition from brittle to ductile-dominant fracture with increasing infill density and optimized printing parameters underscores the importance of interfacial fusion and controlled printing speed in enhancing the crack resistance, toughness, and structural integrity of CF-PETG composites.

Flexural strength

Flexural strength, which evaluates a material’s ability to withstand bending stresses, showed a strong dependence on infill density and pattern, similar to its tensile properties. Figure 12 shows the variation in flexural strength of PETG and CF-PETG samples. PETG specimens demonstrated an increase in flexural strength with increasing density, with P2 at 50% cubic infill showing 44.5 MPa, while P8 at 90% cubic infill reached 58.5 MPa. The higher density reduces porosity and provides a more rigid internal structure that resists deflection under bending loads.

Figure 12.

Variation in flexural strength of 3D printed PETG and CF-PETG.

In terms of infill patterns, cubic structures consistently provided the highest flexural strength. At 90% density, P8 (cubic) exhibited 58.5 MPa, which was greater than P7 (gyroid, 54.7 MPa) and P9 (triangle, 49.1 MPa). The cubic infill distributes stresses more uniformly between tensile and compressive zones during bending, whereas triangular patterns create sharp edges that act as stress concentrators, and gyroid structures deform more easily due to their curved geometry.

Layer thickness also played a notable role. Specimens with thinner layers showed enhanced flexural performance, as observed in P5 (75% cubic, 0.1 mm, 56.5 MPa), which outperformed P4 (75% gyroid, 0.3 mm, 52.6 MPa). Thin layers enhance interlayer bonding, thereby reducing the likelihood of delamination when subjected to bending stresses. Conversely, thicker layers contain more interfacial defects that act as crack initiation points, thereby lowering flexural strength.

Carbon fiber reinforcement further enhanced bending performance. CF5 (75% cubic) achieved the highest flexural strength of 59.2 MPa, surpassing its PETG equivalent P5 (56.5 MPa). The embedded fibers resist tensile stresses on the outer bending surface and delay crack initiation. However, the extent of reinforcement effectiveness varied with pattern; triangular infills in CF3 and CF6 provided lower gains because their geometry concentrated stresses and reduced the contribution of fiber bridging. Overall, flexural strength was optimized with cubic infill, higher density, thinner layers, and carbon fiber reinforcement, highlighting the combined influence of geometry and reinforcement in resisting bending loads.

Figure 13 shows the flexural tested samples of both P and CF. For the P specimens (P1-P9), the flexural failure behaviour and macroscopic fracture features indicate a clear dependence on infill density, pattern and layer thickness. At 50% infill (P1-P3) the samples bent with relatively large mid-span deflection and then failed with brittle-looking breaks or layer delamination - P1 (gyroid, 0.1 mm) showed the most gradual bending before fracture due to better interlayer contact from fine layers, whereas P2 and P3 (cubic/triangle with thicker layers and higher speeds) displayed abrupt cracking and visible separation between layers. The 75% group (P4-P6) carried higher bending loads with smaller deflections and exhibited more localized shear/tensile failure on the tensile face; P5 and P6 showed progressive crack growth and less surface splitting, while P4 (0.3 mm layers) retained some step-like delamination. At 90% infill (P7-P9) the specimens behaved stiffly under bending and failed at higher loads with ductile-looking edge tearing or fiber-like filament bridging (depending on pattern); P9 (90%, triangle, 0.1 mm) combined high density and fine layers to give the best flexural integrity, while specimens printed with coarse layers (0.3 mm) still showed brittle surface spalling despite high infill. Overall, flexural strength increased with infill density and finer layer thickness. The gyroid produced a more uniform bending response, the triangle gave high stiffness but became brittle at low infill, and the cubic tended to concentrate damage, leading to delamination.

Figure 13.

Flexural-tested specimens P1-P9 and CF1-CF9.

For the CF-PETG specimens (CF1-CF9), the presence of carbon fibre noticeably altered the bending response and fracture morphology: carbon fibre reinforcement increased stiffness, reduced mid-span deflection, and promoted crack-bridging and pull-out mechanisms that delay catastrophic failure. At 50% infill (CF1-CF3) CF1 (gyroid, 0.1 mm) showed the most ductile-looking bend-to-break with some matrix tearing and limited fibre pull-out; CF2 and CF3 with thicker layers showed earlier delamination and brittle skin cracking, with CF3 worst due to the combined effect of large layer height and high speed. In the 75% set (CF4-CF6) specimens carried higher bending loads and failed with mixed mode features-compact tensile-face cracking accompanied by clear fibre pull-out and bridging in CF5 and CF6, while CF4 showed non-uniform failure where layer steps initiated cracks. In the 90% group (CF7–CF9) the bending performance was best: CF7 (gyroid) exhibited smooth bending and ductile tearing, CF8 (0.3 mm layers) still suffered from some brittle splitting despite the reinforcement, and CF9 (90%, triangle, 0.1 mm) showed the highest resistance to flexural fracture with extensive fibre–matrix interaction and delayed crack propagation. In summary, carbon fibre improved flexural stiffness and toughness, finer layers and higher infill maximised these benefits, and patterns that promote isotropic load distribution (gyroid/triangle at high density) produced the most favourable flexural failure modes.

Impact strength

Impact strength is a measure of toughness, representing the ability of a material to absorb energy under sudden loading. Figure 14 shows the variation in impact strength of PETG and CF-PETG samples. Unlike tensile and compressive strengths, the impact performance did not consistently improve with increasing density. PETG P1 with 50% gyroid infill exhibited an impact resistance of 0.79 kJ/m², while relatively denser specimens such as P8 (90% cubic) achieved 1.31 kJ/m². This trend indicates that low-density structures did not always guarantee superior energy absorption, as their higher deformability could also lead to premature localized failure under impact.

Figure 14.

Variation in impact strength of 3D printed PETG and CF-PETG.

The choice of infill pattern had a significant impact on toughness. Gyroid infills generally offered better energy absorption than cubic and triangular ones because of their continuous geometry, which facilitated stress distribution and delayed crack propagation. In contrast, cubic and triangular infills, although effective under static loading, were prone to brittle fracture during impact due to their sharp corners, which acted as stress concentrators.

Layer thickness exerted a secondary influence. Thinner layers improved interlayer bonding, slightly enhancing impact resistance, though in high-density samples the inherent brittleness of rigid geometries limited this benefit.

Carbon fiber reinforcement reduced impact resistance when compared to neat PETG. For instance, CF5 recorded 1.33 kJ/m² compared to its PETG counterpart P5 (1.42 kJ/m²), and CF9 showed only 1.35 kJ/m², both of which remained higher than P1 (0.79 kJ/m²). The brittle nature of carbon fibers encouraged early crack initiation and constrained plastic deformation of the polymer matrix, reducing overall energy absorption. This illustrates the inherent trade-off between mechanical strength and toughness that occurs when reinforcement is introduced. Overall, while gyroid infills aided in distributing impact stresses, neat PETG specimens at lower densities did not always provide superior toughness, and carbon fiber reinforcement further compromised impact resistance despite its positive effect on static strength.

The impact fracture morphology of the 3D-printed PETG specimens (Figure 15) clearly demonstrates the influence of infill density, pattern, printing speed, and layer thickness on the failure behavior under impact loading.

Figure 15.

Impact-tested specimens P1–P9 with corresponding fracture surfaces.

At 50% infill density (P1-P3), the specimens exhibited relatively smooth and planar fracture surfaces, characteristic of brittle failure. The limited energy absorption is attributed to lower material continuity and poor interlayer bonding. Among these, P1 (Gyroid pattern, 0.1 mm layer) shows slightly higher resistance due to its continuous cellular geometry, whereas P3 (Triangle pattern, 0.3 mm layer) reveals pronounced interlayer delamination and void formation, indicating poor filament fusion at thicker layers.

For 75% infill density specimens (P4-P6), the fracture surfaces appear rougher and more irregular, revealing improved ductility and enhanced impact resistance. The rough, tortuous crack paths indicate localized yielding and better stress distribution. P4 (Gyroid pattern) exhibits distinct shear lips and microvoid coalescence, suggesting a mixed brittle–ductile failure mode. P6 (Triangle pattern) shows multiple filament pull-outs, implying partial debonding yet higher energy dissipation compared to 50% infill samples.

At 90% infill density (P7–P9), the specimens demonstrate the most rugged fracture morphology with significant roughness, filament tearing, and evidence of plastic deformation. The improved interlayer adhesion and reduced void content allow for greater load transfer and energy absorption before failure. Particularly, P7 (Gyroid, 0.2 mm layer) shows a well-integrated structure with minor delamination, while P9 (Triangle, 0.1 mm layer) exhibits the most ductile behavior with stretched filaments and a highly irregular fracture surface.

The impact fracture surfaces of the carbon fiber reinforced PETG (CF-PETG) composites (Figure 16) reveal a distinct shift in failure mechanism compared to the unreinforced PETG samples. The incorporation of short carbon fibers significantly influences not only the fracture morphology but also the formation and distribution of micropores, primarily governed by the printing parameters such as infill density, layer thickness, printing speed, and infill pattern.

Figure 16.

Impact-tested specimens CF1–CF9 with corresponding fracture surfaces.

At lower printing speeds and finer layers, improved thermal bonding between adjacent filaments resulted in smaller and fewer micropores, contributing to better energy dissipation during impact. Conversely, higher printing speeds and thicker layers caused rapid cooling and incomplete fusion, leading to larger and irregular micropores that acted as micro-stress concentrators and weakened local bonding.

At 50% infill density (CF1-CF3), the fracture surfaces exhibit a relatively porous and irregular morphology with visible interlayer voids arising from inadequate heat transfer and poor consolidation. Despite carbon fiber reinforcement, the combination of low material content and pronounced microporosity limited the energy absorption capability. CF1 (Gyroid, 0.1 mm layer) showed fine striations and partial ductile tearing attributed to better filament fusion at lower layer thickness, whereas CF3 (Triangle, 0.3 mm layer) presented large voids and brittle cracking due to increased micropore formation and insufficient interlayer adhesion.

With 75% infill (CF4-CF6), the fracture morphology became denser, reflecting enhanced fusion and reduced micropore size. The surfaces display microvoid coalescence, fiber pull-out, and crack deflection around carbon fibers-mechanisms that collectively improve energy absorption. CF4 (Gyroid, 0.3 mm layer) demonstrated a mixed failure mode dominated by shear yielding and localized delamination, while CF6 (Triangle, 0.2 mm layer) exhibited multiple fiber pull-outs and rougher topography, indicating efficient stress transfer through the reinforced matrix.

At 90% infill density (CF7-CF9), the fracture surfaces are highly consolidated with minimal microporosity and extensive fiber–matrix interfacial bonding. The increased infill density and optimized printing conditions promote superior interlayer adhesion and uniform load distribution, thereby maximizing impact resistance. CF7 (Gyroid, 0.2 mm layer) exhibited a well-fused structure with fine microcracks and continuous filament alignment, whereas CF9 (Triangle, 0.1 mm layer) showed dense fibrillation, fiber bridging, and ductile tearing zones-confirming maximum energy absorption among all configurations.

Overall, the observed reduction in micropore size and increased ductile tearing with optimized printing parameters highlight the critical role of printing speed, infill density, and layer thickness in governing microporosity-driven impact performance of CF-PETG composites.

Compressive strength

Compressive strength, measured at 10%, 25%, and 50% deformation, is a critical parameter for load-bearing applications, and the results presented in Figure 17 indicated strong dependence on infill geometry, density, and reinforcement as presented in Figure 9. For PETG, compressive strength increased substantially with infill density. At 10% deformation, P1 (50% gyroid) recorded only 7.64 MPa, while P9 (90% triangular) reached 32.17 MPa. Similar improvements were observed at higher deformation levels, where denser structures resisted collapse more effectively due to reduced void space and enhanced load-bearing capacity.

Figure 17.

Variation in compressive strength of PETG and CF-PETG samples.

Among infill patterns, triangular structures exhibited superior compressive resistance, particularly at higher strain levels. At 50% deformation, PETG P9 with triangular infill reached 43.29 MPa, compared to 43.04 MPa for cubic (P8) and 43.44 MPa for gyroid (P7). The truss-like architecture of triangular infills distributes compressive stresses efficiently, minimising localised collapse.

Layer thickness influenced compressive strength by affecting interlayer adhesion. PETG P9 (90% triangle, 0.1 mm) outperformed P7 (90% gyroid, 0.2 mm), demonstrating that thinner layers improve layer fusion and reduce the likelihood of delamination under compression. However, at high strain levels, thicker layers occasionally provided additional stability by minimizing interfacial defects, delaying final collapse.

Carbon fiber reinforcement produced remarkable gains in compressive strength. CF9 (90% triangle, 0.1 mm) exhibited the highest compressive strength, reaching 61.56 MPa at 50% deformation, significantly surpassing the PETG counterpart P9 (43.29 MPa). The improvement is attributed to the role of carbon fibers as micro-columns that resist buckling and distribute compressive loads across the matrix. Even at moderate densities, reinforcement improved load-bearing ability, as seen in CF5 (75% cubic, 55.72 MPa at 50% deformation) compared to P5 (37.79 MPa). In summary, compressive strength was maximized using triangular infill, high density, thin layers, and reinforcement, making CF9 the best-performing sample across all deformation levels.

The compressive behavior of 3D-printed PETG and carbon fiber-reinforced PETG (CF-PETG) specimens was evaluated under varying infill densities, patterns, printing speeds, and layer thicknesses and presented in Figure 18. The visual inspection of post-compression samples revealed distinct differences in deformation and failure modes between the two materials. PETG specimens generally exhibited ductile failure characterized by significant bulging and plastic deformation, particularly in low infill samples (P1–P3). In contrast, CF-PETG samples showed brittle fracture with cleaner failure planes and limited deformation, especially in low-density triangular infill (CF3), indicating the reinforcing effect of carbon fibers increased stiffness at the expense of ductility. Among all infill patterns, the triangle configuration consistently underperformed in both materials, often leading to premature failure, whereas the gyroid and cubic patterns offered more uniform stress distribution and improved compressive resistance. Increasing infill density notably enhanced structural integrity, with 90% infill samples (P7–P9 and CF7–CF9) demonstrating superior load-bearing capabilities. Additionally, a finer layer thickness (0.1 mm) positively influenced mechanical performance by improving interlayer adhesion, as observed in specimens such as P9 and CF5. While higher printing speeds slightly compromised layer bonding, their impact was less significant compared to infill and layer thickness parameters. Overall, CF-PETG exhibited improved compressive strength and dimensional stability under load, making it a suitable candidate for applications requiring higher rigidity and structural performance, provided the brittleness is managed through design optimization.

Figure 18.

Compression-tested specimens P1-P9 and CF1–CF9.

DOE

Main effect plots for SN ratios of PETG and CF-PETG

The experimental characterisation of the 3D-printed composites revealed the combined influence of infill density, infill pattern, printing speed and layer thickness on the overall mechanical behaviour. A comparative evaluation of the PETG and CF-PETG provides a detailed understanding of the structure property relationship in additively manufactured composites.

The main effects plots for the signal-to-noise (S/N) ratios of PETG composites are presented in Figure 19. The plots highlight the influence of four control factors (A, B, C, and D) on the performance characteristics, evaluated under the “larger-the-better” criterion. For PETG, Factor A demonstrated the strongest positive influence, with a clear upward trend from Level 1 to Level 3, indicating that higher settings of this parameter significantly enhance the S/N ratio. Factor B exhibited a mild improvement with increasing levels, suggesting a moderate but positive contribution to performance. Factor C showed a sharp peak at Level 2, implying that the intermediate setting provides the optimal condition, while both lower and higher levels reduce effectiveness. Factor D revealed a shallow U-shaped trend, where Level 2 resulted in reduced performance, whereas Levels 1 and 3 were more favorable. Consequently, the optimal combination for PETG was identified as A3–B3–C2–D1/3.

Figure 19.

Main effect plot for SN ratios for PETG.

In contrast, the CF-PETG composites exhibited distinct variations in factor sensitivity as evident from Figure 20. Factor A again emerged as the most dominant parameter, but unlike PETG, the trend peaked at Level 2 and slightly decreased at Level 3, establishing Level 2 as the optimum condition. Factor B showed a clear downward slope, indicating that higher levels deteriorated performance, and thus Level 1 was optimal. Similar to PETG, Factor C peaked at Level 2, confirming the robustness of this condition across both material systems. Factor D displayed a pronounced U-shaped response, with Levels 1 and 3 being more favorable compared to Level 2. Hence, the optimal factor setting for CF-PETG was determined as A2–B1–C2–D1/3.

Figure 20.

Main effect plot for SN ratio for CF-PETG.

A comparative assessment between PETG and CF-PETG indicates that Factor A is consistently the most influential for both materials, although the optimum level shifts due to the presence of carbon fiber reinforcement. Notably, Factor B exhibited opposite behavior in the two systems, where higher levels enhanced PETG performance but diminished that of CF-PETG, underscoring the role of reinforcement in altering parameter sensitivity. Factor C showed consistent optimum behavior at Level 2, while Factor D had a relatively weak effect in both materials, with a common dip at the intermediate level. Overall, these findings demonstrate that reinforcement with carbon fibers modifies the process-parameter interactions, leading to material-specific optimization strategies.

Interaction plots for PETG

The interaction plots illustrate how the selected process parameters (A, B, C, and D) and their combinations influence the mechanical properties of PETG composites. The nature of the slopes and intersections in these plots reveals whether factors act independently or exhibit strong interactive effects on performance.

The interaction plot for tensile strength as presented in Figure 21 shows notable variations across factor combinations. Factor A demonstrates a consistent upward trend, indicating that higher levels of this parameter enhance tensile strength. However, significant crossing of lines in combinations involving Factors B and D indicates strong interactive effects, meaning the influence of one factor depends on the level of the other. For instance, tensile strength is maximized when Factor A is at its higher level and coupled with the optimum levels of B and D. This highlights that tensile performance is governed not only by individual parameter settings but also by their combined effects.

Figure 21.

Interaction plot for tensile strength-PETG.

In the case of flexural strength, the plots reveal multiple instances of intersecting lines, particularly in interactions between Factors A–B and C–D as presented in Figure 22. This suggests that flexural properties are highly sensitive to parameter interactions. While Factor A shows a generally increasing trend with higher levels, its effectiveness is amplified or diminished depending on the settings of Factor B. Similarly, Factor C exhibits a peak response at intermediate levels, but the overall effect is strongly modified by Factor D. These results indicate that optimizing flexural strength requires careful selection of parameter combinations rather than focusing on single factors in isolation.

Figure 22.

Interaction plot for flexural strength-PETG.

For compressive strength, the plots depict relatively steeper slopes for Factors A and B, indicating their dominant role as presented in Figure 23. The presence of crossing lines suggests considerable interaction effects, especially between Factors A–B and B–C. Higher levels of Factor A combined with lower-to-intermediate levels of Factor B yield maximum compressive strength. Interestingly, Factor C shows inconsistent behavior, implying that its influence is conditional upon the levels of the other parameters. These findings confirm that compressive strength is highly dependent on multi-factor synergy.

Figure 23.

Interaction plot for compressive strength-PETG.

The interaction plot for impact strength shows more subtle trends compared to the other properties as shown in Figure 24. Factors A and C exhibit positive slopes, with their highest levels generally favoring improved impact resistance. However, the crossing of lines in A–C and C–D combinations suggests that these improvements are not universal across all factor levels. Factor B appears to have a weaker influence, as reflected by its relatively flat lines, while Factor D shows moderate interactions. Overall, impact strength is more influenced by second-order effects rather than single-factor dominance.

Figure 24.

Interaction plot for impact strength-PETG.

Interaction plots for CF-PETG

The interaction plots illustrate the combined influence of processing parameters (A, B, C, and D) on the mechanical properties of CF-PETG composites. The reinforcement with short carbon fibers significantly alters the sensitivity and interaction trends compared to neat PETG, reflecting the complex role of fiber–matrix interactions in determining mechanical response.

Figure 25 shows the interaction plot for tensile strength of CF-PETG. For tensile strength, the interaction plots reveal less linearity compared to PETG, with multiple crossovers between parameter levels, indicating strong interactive effects. Factor A continues to play a significant role, though its influence varies depending on the levels of Factors B and D. The combined effect of A–C and A–D suggests that the tensile response is maximized only under specific multi-parameter conditions. Unlike PETG, where Factor A showed a clear upward trend, in CF-PETG the reinforcement causes more irregular variations, indicating that fiber dispersion and bonding efficiency are highly sensitive to process parameter combinations.

Figure 25.

Interaction plot for tensile strength of CF-PETG.

Flexural strength in CF-PETG shows pronounced interaction behavior, particularly in the A–B and C-D plots, where strong line crossings highlight interdependence of factors as presented in Figure 26. Unlike PETG, where Factor A showed a relatively smooth effect, the CF-PETG system displays sharp fluctuations, suggesting that flexural performance is more sensitive to local variations in fiber alignment and resin flow during printing. Optimal flexural strength is therefore achieved not by maximizing individual factors, but by carefully balancing them in interaction.

Figure 26.

Interaction plot for flexural strength of CF-PETG.

The compressive strength plots exhibit steep slopes for Factors A and B, confirming their dominant influence as depicted in Figure 27. However, unlike PETG, the CF-PETG composites show more consistent positive trends across factor levels, especially for B–C and A–C combinations. This suggests that fiber reinforcement stabilizes the compressive response by restricting polymer chain mobility and enhancing load transfer efficiency. The reduced extent of crossover in some interactions indicates that compressive performance in CF-PETG is less prone to detrimental parameter conflicts compared to tensile and flexural strengths.

Figure 27.

Interaction plot for compression strength of CF-PETG.

Impact strength plots reveal subtle but notable interactive effects as shown in Figure 28. Factors A and C continue to have measurable influence, with line crossings at A–C and C–D interactions suggesting that fiber reinforcement modifies crack initiation and propagation pathways depending on parameter combinations. The overall values of impact strength remain relatively low compared to tensile and compressive strengths, which is expected due to the inherent brittleness introduced by fiber reinforcement. However, certain factor combinations still enable improved toughness by promoting better energy dissipation at the fiber–matrix interface.

Figure 28.

Interaction plot for impact strength of CF-PETG.

The interaction analysis for CF-PETG highlights that reinforcement with carbon fibers intensifies the dependency of mechanical properties on parameter interactions. While PETG showed clearer main effects with moderate interaction influences, CF-PETG exhibited irregular variations and frequent line crossings, particularly in tensile and flexural responses. This underscores the role of fiber–matrix adhesion, dispersion, and orientation, which are highly sensitive to processing conditions. Compressive strength benefited most from fiber reinforcement, displaying more stable and predictable trends, whereas tensile and flexural strengths showed higher variability due to complex load transfer mechanisms. Impact strength, though limited in magnitude, revealed that optimal parameter combinations can partially mitigate brittleness.

Response optimizer

The response optimizer plots for PETG and CF-PETG composites were generated to identify the optimal combination of printing parameters (A: Infill Density, B: Pattern, C: Printing Speed, D: Layer Thickness) for maximizing mechanical performance metrics namely tensile, flexural, compressive, and impact strengths. The optimizer uses desirability functions to predict parameter settings that yield the highest overall performance.

For PETG, as shown in Figure 29, the optimal composite desirability was 0.8518, with the predicted maximum mechanical properties of tensile strength 34.93 MPa, flexural strength 59.73 MPa, compressive strength 50.90 MPa, and impact strength 1.72 kJ/m². The optimizer indicates that maximum performance is achieved when all parameters (A-D) are at their high levels (level 3), highlighting that increased infill density, complex gyroid pattern, higher printing speed, and thicker layers synergistically enhance overall PETG composite performance. Notably, impact strength exhibits lower desirability (d = 0.5265) compared to other mechanical properties, suggesting a compromise between stiffness and toughness.

Figure 29.

Response optimiser for PETG

For CF-PETG, as shown in Figure 30, the optimizer predicts a higher composite desirability of 0.9872, reflecting the reinforcing effect of carbon fibers. The corresponding maximum mechanical properties are tensile strength 31.63 MPa, flexural strength 58.83 MPa, compressive strength 67.57 MPa, and impact strength 1.44 kJ/m². Here, optimal conditions involve high levels of infill density (A = 3) and layer thickness (D = 3), with pattern (B = 1) and printing speed (C = 2) slightly lower than the maximum. This adjustment reflects the interplay between fiber reinforcement and printing parameters, where excessive printing speed or pattern complexity may adversely affect fiber-matrix bonding. All properties exhibit desirability close to unity, indicating a robust overall optimization.

Figure 30.

Response optimiser for CF-PETG.

While PETG shows a more uniform increase in mechanical performance with increasing parameter levels, CF-PETG exhibits a nuanced response due to the fiber reinforcement, particularly in compressive strength where CF-PETG significantly outperforms PETG. The optimizer plots confirm that parameter tuning can substantially enhance performance for both materials, with CF-PETG benefiting more from parameter interaction effects due to fiber-matrix synergy.

Fractography

The fractographic analysis presented in Figure 31 corroborates the mechanical and DOE results by revealing clear correlations between microstructural features and macroscopic performance. In PETG specimens, the 50% infill sample exhibited distinct interlayer voids, poor filament fusion, and irregular deposition, indicating weak interlayer adhesion and limited polymer chain diffusion that led to brittle fracture. Increasing the infill to 75% resulted in improved interlayer cohesion and partial ductile tearing, while the 90% infill sample displayed a densely fused structure with tearing ridges and fibrillation, signifying enhanced load transfer and plastic deformation. In CF-PETG composites, the 50% infill specimen showed fiber pull-out and interfacial debonding, acting as crack initiation sites, whereas the 75% infill sample exhibited better fiber–matrix interlocking and partial fiber fracture, suggesting efficient stress transfer. At 90% infill, a dense and homogeneous fracture surface with firmly embedded, broken fibers confirmed strong interfacial adhesion, though localized fiber agglomeration and restricted melt flow may have induced stress concentrations and reduced global performance. These observations align with the DOE findings that identified infill density as the most influential parameter and justify the optimum performance of CF-PETG at 75% infill, where balanced fiber dispersion, interlayer bonding, and thermal stability collectively enhance strength and stiffness while slightly compromising impact resistance.

Figure 31.

Fractography of P and CF specimens with different infill densities.

The fractographic observations correlate strongly with the mechanical test results: specimens with higher infill and finer layer thickness exhibit better inter-layer fusion, reduced micropore size and frequency, and denser fracture morphologies-features that directly translate to higher tensile, flexural and compressive strengths. Carbon-fiber reinforcement enhances flexural and compressive performance when fibers are well-dispersed and firmly bonded to the PETG matrix, as evidenced by fiber breakage and pull-out on fracture surfaces; however, excessive stiffness or localized defects (agglomerates or trapped pores) can attenuate low-velocity impact absorption. Overall, a combination of ∼75% infill, 0.1-0.2 mm layer thickness, and moderate printing speed (40-50 mm/s) provides the best compromise between strength, toughness and impact resistance, matching both DOE and SEM/fractography evidence.

Conclusions

The present study systematically investigated the influence of key fused filament fabrication (FFF) parameters on the mechanical performance of PETG and carbon fiber-reinforced PETG (CF-PETG) composites, leading to the following conclusions.

• CF-PETG exhibited superior compressive and flexural strengths compared to PETG. The highest compressive strength was recorded for CF9, reaching 47.21 MPa at 10% deformation and 61.56 MPa at 50% deformation, which represents a 42% improvement over the best-performing PETG specimen. Flexural strength improved by up to 23%, with CF5 achieving 59.2 MPa compared to 48.2 MPa for PETG. However, PETG maintained higher impact strength, reaching 2.55 kJ/m² (P9) against only 1.44 kJ/m² for CF-PETG, highlighting the trade-off between stiffness and toughness.

• Design of Experiments (DOE) analysis identified infill density (Factor A) as the most significant factor influencing compressive, flexural, and tensile properties for both PETG and CF-PETG.

• PETG showed continuous improvement with higher infill density (from 50% to 90%) and more complex infill geometries (gyroid and triangle). In contrast, CF-PETG reached its peak performance at 75% infill density, printing speed of 45 mm/s, and layer thickness of 0.2-0.3 mm, suggesting that excessive complexity or higher infill disrupted fiber dispersion and weakened matrix bonding.

• Two-factor interactions, particularly between infill density and pattern (A-B) and between printing speed and layer thickness (C-D), significantly influenced tensile and flexural responses. CF-PETG exhibited irregular interaction trends due to variability in fiber–matrix wetting and load transfer.

• Desirability function analysis revealed composite desirability scores of 0.9872 for CF-PETG and 0.8518 for PETG, confirming that CF-PETG offers the highest potential for multi-objective optimization when processing parameters are carefully selected.

• The findings demonstrate that carbon fiber reinforcement substantially improves compressive and flexural properties but compromises impact resistance. Furthermore, it modifies the sensitivity of PETG to FFF parameters, underscoring the need for precise parameter optimization. This establishes CF-PETG as a promising material for lightweight, load-bearing applications in aerospace, automotive, and structural components where strength and stiffness are prioritized.

This study highlights that tailoring printing parameters in tandem with fiber reinforcement is essential to unlock the full mechanical potential of FFF-based composites, thereby paving the way for their reliable integration into next-generation structural and lightweight engineering applications.