Development of GaN-Based UV LEDs: Challenges and Emerging Solutions

1. Introduction

Additive manufacturing (AM) or 3D printing (3DP), is a technology that is widely used to fabricate parts layer by layer from a computer-aided design (CAD) model ^[1]. There are seven different AM categories: material extrusion (ME), vat photopolymerization (VAT), powder bed fusion (PBF), direct energy deposition (DED), sheet lamination (SL), material jetting (MJ), and binder jetting (BJ) ^[2]. ME is the most widely used AM technology because of its numerous advantages, such as its use of less material and time to produce complex parts, its low cost, and its environmental friendliness ^[3-5]. ME is already applied in several areas, such as the food industry, medicine, and aerospace industry ^[6-8]. ME is also now used in metal object manufacturing now ^{[9, 10]}.

Despite the wide use of ME, Metal ME presents a challenge due to working temperature requirements ^[3]. The working temperature of most ME 3D printers varies from 200 to 280°C, and the melting points of most metal materials are much higher. In recent years, new metal-infused polymer filaments have been developed as a feedstock material for ME and low-cost metal material extrusion (LCMME) isa new method for fabricating parts with this new type of material ^{[11, 12]}.

Surface roughness (SR) refers to deviations in the direction of the normal vector of a real surface from its ideal form ^[13]. It is an important mechanical property of metal. The SR of parts manufactured using AM processes has been studied in some works. Ciracì et al. performed research on the impact of SR in several metallic systems ^[13]. Due to the fabrication-induced surface roughness, most metallic systems suffer from some degree of inhomogeneity. He et al. reviewed the influencing factors and related modeling methods ^[15]. This work aimed to generate a comprehensive understanding of the turned surface roughness in theoretical modeling. Alfieri et al. studied the influence of SR in AM applications ^[16].

Machine learning (ML) is a subset of artificial intelligence that can be used to predict the mechanical properties of AM-fabricated parts ^{[17, 18]}. ML has plenty of applications in AM; for example, ML is used to improve the quality of design ^[19], it can help manufacturers perform anomaly detection ^[20], and, with the help of ML, doctors can read the CT images and answer questions accurately ^[21]. However, in specimen fabrication, ML has not been sufficiently studied. Since SR is an important parameter in specimen fabrication, its control is a key question in the industry. After reviewing previous researches, it was found that there is no research on using ML to predict and improve the SR of LCMME-fabricated parts. Thus, this work will compensate for this shortcoming. In this study, the author chose different manufacturing parameters, and the SR was collected based on different parameters. Totally, 327 groups of data were collected from LCMME-fabricated parts. The influence of different manufacturing parameters was analyzed, and two ML algorithms—support vector regression (SVR) and neural network (NN)—were generated to predict the SR values.

With the development of 3DP technology, metal 3DP is a hard point in this research area. Also, ML algorithms already have been widely used in quality evaluation in 3DP. Thus, this work provides a newly developed metal 3DP technology and the SR was chosen to be evaluated.

2. Materials and Methods

2.1. Manufacturing Process

Figure 1 presents a sketch of LCMME. The first several steps are the same as those in traditional AM, but after the AM process, parts fabricated by using metal-infused polymer filaments are sintered to melt out the internal polymer. The sintered part is then composed of pure metal.

In Figure 2, the process of sintering is introduced. In the sintering process, the sintering temperature is lower than the melting temperature. As the temperature increases, the plastic melts out, and the metal powders gather together to form pure metal parts.

2.2. Material Used in the Study

As a widely used plastic material, polylactic acid (PLA) has plenty of applications in both research and manufacturing ^{[22, 23]}. Bronze is a metallic material that also has many applications ^[24] having been used by human beings for more than 5000 years ^[25]. Thus, in this study, the samples were fabricated using bronze-PLA filaments. The filament was made by The Virtual Foundry, it has a diameter of 2.85mm, contains 90% bronze, and has a density of 4.5g/cc. Figure 3 shows the metal-composite part as a CAD model, when it was 3D printed, and after sintering. In this research, the author used LCMME technology fabricated several samples and the micro view of four different samples are shown in Figure 4. The CAD model is designed in SolidWorks and sliced using Cura. The bronze-PLA parts were 3D printed using an Ultimaker S5 3D Printer, and a KSL-1100X muffle furnace was used to sinter the parts. The SR of the pure bronze parts after sintering was measured using a SJ-210 SR Tester. Figure 5 shows the SR tester used in this study.

2.3. Data Preparation

There are several SR parameters, such as R_a, R_ms, R_z, R_v, and so on. In this research, R_a is used since its accuracy and simplicity. R_a is the arithmetical mean deviation of the assessed profile; the equation is shown below:

$R_{a} = \frac{1}{l_{r}}\int_{0}^{l_{r}}{\mid z(x) \mid dx}$(1)

where l_r is the measured length. In this research, the value of l_r is chosen as 20% of the specimen length. For example, if the length of the part is 20mm, the l_r of this part is 4mm.

In this research, there are five different manufacturing parameters, which are:

Layer Thickness (LT): the height of each layer during the printing process;

Sintering Temperature (ST): the temperature to sinter the bronze-PLA parts;

Ramp Ratio (RR): the temperature increasing ratio from room temperature to ST;

Nozzle Temperature (NT): the temperature of the printing nozzle during the 3DP process;

Printing Speed (PS): the moving speed of the nozzle during the 3DP process.

Table 1 shows the units and values of different manufacturing parameters.

2.4. SVR and VV

In this study, support vector regression (SVR) and a neural network (NN) were used to predict the SR results of LCMME fabricated parts. SVR provides flexibility to define whether the error in a model is acceptable or not. An NN uses a set of network layers to translate an input into an output. These two methods have been proven to be effective tools in real-value estimation. Figure 6 shows the structure of the NN algorithm. Operators input the manufacturing parameters, and the algorithm generates the predicted SRvalue.

3. Results

In this ssction, the results are shown.

Figure 7 shows the top, front, and side views of a sample, and Table 2 shows the SR values of different surfaces of two different specimens. These two specimens that were chosen had different manufacturing parameters; for example, the LT was 0.3mm for the first one and the second specimen had an LT of 0.2mm LT. In this study, the data were firstly divided into three groups, which were the top, front, and side. Different manufacturing parameters lead to different SR values for different surfaces.

3.1. Results of ANOVA

This study used R to apply ANOVA for two main purposes. The first was to determine whether all surfaces had the same SR values for each cubic sample. The second was to determine whether all manufacturing parameters had an influence on the SR values.

A one-way ANOVA model was developed to determine whether the SR is the same on different surfaces. The following hypotheses were set:

H₀: μ_{SR_Top} = μ_{SR_Front} =μ_{SR_Side}

H_a: at least one group of SR values is different

The result of this one-way ANOVA result is shown in Table 3. F crit < F, which means the H₀ is rejected. Thus, the three groups of surfaces do not have the same SR values.

However, during the data collection, the researcher found that the SR_Front and SR_Side are similar. So the research group did another one-way ANOVA model is developed to determine if the SR_Front and SR_Side of a part are the same or not. The following hypothesis is set:

H₀: μ_{SR_Front} =μ_{SR_Side}

H_a: μ_{SR_Front} ≠μ_{SR_Side}

The result of this one-way ANOVA result is shown in Table 4. F crit > F, which means the H0 is not rejected. Thus, the two groups of surfaces have the same SR values.

Thus, the three groups of data can be simplified into two groups, which are SR_Top and SR_Edge (including Front and Side). Then, further ANOVA analyses were generated to find if all manufacturing parameters have influence on SR values or not. Table 5 and 6 show the results.

In an ANOVA, if the P-value is smaller than 0.001, itindicates that this parameter does not influence the final result. The results from the two tables above show that the P-values of ST, NT, and RR on SR_Top are smaller than 0.001, meaning that these three parameters have an influence on SR_Top. In Table 5, the P-values of LT and PS is 0.1962310 and 0.3608456 respectively. Thus, these two parameters do not affect SR_Top. In addition, all parameters have p-values smaller than 0.01 for SR_Edge. Thus, in the manufacturing process, all parameters affect the SR_Edge values.

As shown be the ANOVA results, the data can be simplified into two groups, which are SR_Top and SR_Edge. In the ML algorithms, only three parameters were used as independent variables in the SR_Top analysis. All five parameterswere used as independent variables in the SR_Edge analysis.

3.2. Results of ML Algorithms

In this study, the ML algorithms were run in Python and the mean square error (MSE) metric was used to evaluate the performance of each algorithm. Table 7 shows the MSE results, indicating that the NN behaved better in predicting the SR of the LCMME fabricated parts.

It has been proved that as the number of samples increased, the accurcy of ML will raise too^[5]. Thus, in order to get a lower MSE, more data collected from new samples could be added into the database.

4. Conclusions

In the AM process, the SR is an important value for quality evaluation. In this study, a new metal AM technique, LCMME, was developed to fabricate pure metal parts in a low-temperature situation without a laser. Two different ML algorithms were used to predict the SR of LCMME-fabricated parts. The conclusions of this research are as follows:

1. The top side of the LCMME-fabricated parts has a different SR value from that of the edge sides.

2. Only three manufacturing parameters have an influence on SR_Top; ST, NT, and RR.

3. All five parameters affect the SR_Edge values.

4. The MSE of an NN is smaller than that of SVR in overall.

This study focused on analyzing the SR of LCMME-fabricated parts and employed ML algorithms to perform prediction. It can help manufacturers develop a more accurate design before processing. In addition, by controlling the manufacturing parameters, better specimens will be fabricated in the future.

5. Future Work

In this study, the influence of different manufacturing parameters is evaluated using ANOVA. In addition, two ML methods are used to predict the SR values of LCMME-fabricated parts. In future research, more parameters, such as the flow rate, build-plate temperature, fan speed, and material flow rate can be added to increase the accuracy of prediction. However, the ML algorithms used in this study occupy significant time and storage sources to generate results. Thus, if the research group could use better equipments, more data could be used in the ML algorithms and more accurate results will be generated. In addition, a straightforward and simple model, such as linear regression with interactions and k-Nearest Neighbors, can be used. In addition, more metal-polymer materials can be used for fabrication with this newly developed method. Finally, the research group will try to find an industrial application of this technology

Funding

This work was supported by Shandong Agriculture and Engineering University Start-Up Fund for Talented Scholars (BSQJ-202301).

Acknowledgments

This research has been made possible with the help provided by Mr. Xin Liu.

Conflicts of Interest

The author Zhicheng Zhang is an editor of X-Disciplinarity but was not involved in the peer review or decision-making process for this article.

References

1. Z. Zhang, I. Fidan, M. Allen. “Detection of material extrusion In-process failures via deep learning.” Inventions 2020, 5, 3, 25.

2. “ISO/ASTM 52900:2015 - additive manufacturing — general principles — terminology.”

3. S. Kedziora, T. Decker, E. Museyibov, J. Morbach, S. Hohmann, A. Huwer, M. Wahl "Strength properties of 316L and 17-4 PH stainless steel produced with additive manufacturing." Materials 2022, 15, 18, 6278.

4. M. Attaran. "The rise of 3-D printing: The advantages of additive manufacturing over traditional manufacturing." Business Horizons 2017, 60, 5, 677.

5. Z. Zhang, I. Fidan. "Machine Learning-Based Void Percentage Analysis of Components Fabricated with the Low-Cost Metal Material Extrusion Process." Materials 2022, 15, 12, 4292.

6. S. Choton, N. Gupta, J. Bandral, N. Anjum, A. Choudhary. "Extrusion technology and its application in food processing: A review." The Pharma Innovation Journal 2020, 9, 2, 162-168.

7. P. Ravi. "Understanding the relationship between slicing and measured fill density in material extrusion 3D printing towards precision porosity constructs for biomedical and pharmaceutical applications." 3D Print. Med. 2020, 6, 1, 10.

8. J. Najmon, S. Raeisi, A. Tovar. "Review of additive manufacturing technologies and applications in the aerospace industry." Additive manufacturing for the aerospace industry. 2019, 7, 31.

9. M. Mohammadizadeh, H. Lu, I. Fidan, K. Tantawi, A. Gupta, S. Hasanov, Z. Zhang, F. Alifui-Segbaya, A. Rennie. "Mechanical and Thermal Analyses of Metal-PLA Components Fabricated by Metal Material Extrusion." Inventions 2020, 5, 3, 44.

10. G. Fecher, J. Kübler, C. Felser "Chirality in the solid state: Chiral crystal structures in chiral and achiral space groups." Materials. 2022, 15, 17, 5812.

11. H. Gong, D. Snelling, K. Kardel, A. Carrano. "Comparison of Stainless Steel 316L Parts Made by FDM-and SLM-Based Additive Manufacturing Processes." JOM. 2019, 71, 3, 880-885.

12. Z. Zhang, H. Wang, J. Li, H. Fu. "Recent Applications of Metal Additive Manufacturing: A Review." X-Disciplinarity. 2025, 1, 1, 8.

13. P. Huang, Pao-Ming, C. Lee. "Estimation of Tool Wear and Surface Roughness Development Using Deep Learning and Sensors Fusion." Sensors. 2021, 21, 16, 5338.

14. C. Ciracì, F. Vidal-Codina, D. Yoo, J. Peraire, S.-H. Oh, D. R. Smith. "Impact of surface roughness in nanogap plasmonic systems." ACS Photonics 2020, 7, 4, 908–13.

15. C. L. He, W. J. Zong, J. J. Zhang. “Influencing factors and theoretical modeling methods of surface roughness in turning process: state-of-the-art.” Int. J. Mach. Tools Manuf. 2018, 129, 15–26.

16. V. Alfieri, P. Argenio, F. Caiazzo, V. Sergi. “Reduction of surface roughness by means of laser processing over additive manufacturing metal parts.” Materials 2017, 10, 1, 30.

17. W. Ben Chaabene, M. Flah, M. L. Nehdi. “Machine learning prediction of mechanical properties of concrete: critical review.” Constr. Build. Mater. 2020, 260, 119889.

18. S. Badini, R. Stefano, P. Raffaele. "Unleashing the power of artificial intelligence in materials design." Materials. 2023, 16, 17, 5927.

19. C. Wang, X. P. Tan, S. B. Tor, C. S. Lim. “Machine learning in additive manufacturing: state-of-the-art and perspectives.” Addit. Manuf. 2020, 36, 101538.

20. L. Meng, B. McWilliams, W. Jarosinski, H.-Y. Park, Y.-G. Jung, J. Lee, J. Zhang. “Machine learning in additive manufacturing: a review.” JOM 2020, 72, 6, 2363–77.

21. C. Gobert, E. W. Reutzel, J. Petrich, A. R. Nassar, S. Phoha. “Application of supervised machine learning for defect detection during metallic powder bed fusion additive manufacturing using high resolution imaging.” Addit. Manuf. 2018, 21, 517–28.

22. B. Freeland, E. McCarthy, R. Balakrishnan, S. Fahy, A. Boland, K. Rochfort, M. Dabros, R. Marti, S. M. Kelleher, J. Gaughran "A review of polylactic acid as a replacement material for single-use laboratory components." Materials. 2022, 15, 9, 2989.

23. S. Liparoti, D. Sofia, A. Romano, F. Marra, R. Pantani. "Fused filament deposition of PLA: the role of interlayer adhesion in the mechanical performances." Polymers. 2021, 13, 3, 399.

24. Y. Lu, D. Li, and F. Liu. "Characterizing the chemical structure of Ti3C2Tx MXene by angle-resolved XPS combined with argon ion etching." Materials 2022, 15, 1, 307.

25. B. Salvadori, A. Cagnini, M. Galeotti, S. Porcinai, S. Goidanich, A. Vicenzo, C. Celi, P. Frediani, L. Rosi, M. Frediani, G. Giuntoli, L. Brambilla, R. Beltrami, S. Trasatti "Traditional and innovative protective coatings for outdoor bronze: Application and performance comparison." Journal of Applied Polymer Science. 2018, 135, 12, 46011.