PET-CF Annealing: How Can It Improve FDM 3D Prints?

FDM 3D-printed parts can behave differently from components made by conventional processes because they are built from layers and may contain pores or weaker bonding zones. For engineers considering carbon-fibre-reinforced PET, or PET-CF, an important question is whether post-print heat treatment can improve strength and stiffness—and whether the same annealing settings benefit every mechanical property.

Key findings

• The study tested annealing temperatures of 80 °C, 130 °C and 180 °C, for 30, 120 and 210 minutes.
• Its conclusion reports maximum increases of 18.34% in tensile strength, 18.7% in tensile modulus, 30.85% in flexural strength and 15.87% in flexural modulus.
• Temperature was the most influential factor across all four mechanical properties studied.
• The highest reported flexural strength and flexural modulus, 5.6 GPa, occurred at 130 °C for 120 minutes.
• The tensile-strength prediction model was weak: its R-sq was 56.73%, and its predicted R-sq was 2.65%.

Why study annealing of PET-CF?

Additive manufacturing makes complex parts by adding material layer by layer. In fused deposition modeling (FDM), thermoplastic filament is heated and deposited through a nozzle. Rapid cooling, bonding between layers and internal pores can all affect the finished part’s load-bearing capacity. PET-CF is polyethylene terephthalate reinforced with short carbon fibres, developed for rigid and dimensionally stable parts. Its manufacturer states that annealing can further alter its performance, but the outcome depends on the material and the heat-treatment conditions.

Annealing is controlled heating after printing, followed by cooling. In polymers, it can relieve residual stresses and promote changes in the material structure. However, excessive or uneven heating may cause distortion or brittleness. Testing a particular filament is therefore more useful than assuming that a heat-treatment recommendation for another polymer will produce the same outcome.

What the study tested

The paper by Adi Pandžić, PhD, was published in 2025 in the Proceedings of the 36th DAAAM International Symposium on Intelligent Manufacturing and Automation.

The study tested UltiMaker PET CF filament with a diameter of 2.85 mm. Tensile specimens were designed according to ISO 527-2, and flexural specimens according to ISO 178. The reported printing settings included a 0.6 mm nozzle, 0.15 mm layer height, 100% infill density, 270 °C printing temperature, 80 °C build-plate temperature and 25 mm/s printing speed. Mechanical tests were carried out on a Shimadzu AGS-X universal testing machine at a crosshead speed of 5 mm/min.

The study tested 100 specimens: 50 for tensile tests and 50 for flexural tests. Five specimens were used for each combination of the three temperatures and three annealing durations. A further five specimens for each test type formed the untreated control group. After oven annealing, the specimens were allowed to cool at room temperature. The data were assessed statistically, including with analysis of variance (ANOVA), which estimates the effects of individual factors and their interactions.

Results: temperature mattered more than duration alone

The clearest result from the ANOVA was that temperature was the dominant factor. For tensile strength, its contribution to the model was 36.06%, with a p-value of 0.004. Annealing time and the temperature-by-time interaction were not statistically significant for this response. However, the tensile-strength model did not capture the full variation reliably: it explained 56.73% of the total variation, while its predicted R-sq was only 2.65%. The model should therefore not be treated as a precise prediction tool.

The untreated specimens had a reported average tensile strength of 47.5 MPa. The results table gives 56.3 MPa for specimens annealed at 130 °C for 210 minutes. The paper’s conclusion reports a maximum tensile-strength increase of 18.34%, while also noting that the model for this property was unsatisfactory. These points should be considered together: measured improvements were reported, but confidence in the model’s predictive performance is limited.

For tensile modulus, a measure of stiffness under tension, the model explained 97.88% of the variation. Temperature contributed 95.86%; time alone was not statistically significant, while the temperature-by-time interaction was significant but had a smaller contribution. The conclusion reports a maximum tensile-modulus increase of 18.7%.

For flexural strength, or resistance to loading in bending, temperature contributed 92.71% to the model. Time and the temperature-by-time interaction were also statistically significant, but contributed less. The model’s R-sq was 96.68%. The paper reports that 130 °C for 120 minutes achieved the highest flexural strength. The highest recorded flexural modulus, a measure of stiffness in bending, was 5.6 GPa at the same setting. The conclusion reports maximum increases of 30.85% for flexural strength and 15.87% for flexural modulus.

What the findings mean in practice

For product development and functional-part manufacturing, the results suggest that annealing should be selected according to the property that matters for the application. In this experiment, 130 °C for 120 minutes was the best reported setting for flexural strength and flexural modulus. That does not establish the same setting as optimal for every PET-CF part, geometry or production setup. The findings apply to the tested filament, FDM printing parameters and specimen designs.

For students and researchers, the paper illustrates why statistical significance and model quality need to be considered alongside average test results. A statistically significant factor does not automatically mean that a model predicts outcomes reliably. This is especially important for tensile strength, where predicted R-sq was very low. Tensile and flexural tests also represent different loading conditions, so an improvement in one response should not automatically be treated as an equal improvement in every function of a part.

Limitations and next steps

The experiment used five specimens for each treatment condition. The paper suggests that more replicates are needed to improve the reliability of the tensile-strength model. It also identifies outliers and inconsistent cooling conditions as possible contributors to its weak predictive performance. Differences from manufacturer specifications are attributed to differences in printer models, filament storage conditions and PET-CF’s sensitivity to moisture.

The paper proposes further work on moisture, controlled cooling after annealing and the effects of annealing on other FDM filaments. Such studies could help distinguish the effects of material condition, printing and thermal post-processing.

Conclusion

In this study, annealing improved measured mechanical properties of FDM-printed PET-CF, with temperature having a greater influence than duration alone. The highest reported flexural strength and flexural modulus were achieved at 130 °C for 120 minutes. For engineering design, the findings should be validated on the actual part and with controlled cooling, and the tensile-strength model should be used with particular caution.

Frequently asked questions

Author: Adi Pandžić, PhD in Mechanical Engineering

This article summarises a published research paper; please refer to and cite the original paper. About the author

← All articles