Scaling up the 3D printing of surgical guides: repeatability and energy efficiency

Additive manufacturing (AM) of polymer materials is finding increasing use in such a wide variety of industries including construction, biomedical, electronics, aerospace and textile industries. It is highly used for biomedical applications because it allows for product customization and can be altered for different medical contexts (Lakkala et al., 2023; Mader et al., 2021; McAlister and Wood, 2014). One of the advantages of 3D printing is its wide range of methods and materials that can be used (Mami and Rev, 2017). Actually, this is of paramount importance in the medical field as cases require patient-specific applications. Printing customized parts and supports for surgery planning and during surgery to aid the surgeon (e.g. guides to help the cutting operations) represents one of the most widely spread trends of the last decades. Planning before surgeries is crucial for surgeons and having patient-specific models has proven highly beneficial (Bozkurt and Karayel, 2021). These 3D-printed models can be used not only by the surgeon but also by medical students to enhance training in the operating room. Nevertheless, 3D-printed models are not limited to operating room applications. They can also be used outside of the operating room for disease modeling as well as medical device prototyping. For example, hearing aids in pediatric cases are complex and require many adjustments over short periods. Since standard production time can take longer than what is required, 3D-printed hearing aids have proven to be more efficient to produce given their quick turnaround time (Dodziuk, 2016).

This increasing number of applications, coupled with the diffusion of these techniques, led to the formation of an interesting market for additively manufactured parts for surgical applications. The production of these parts is not gathered in a few big companies but is more likely to take place directly in hospitals, schools, clinics or small firms that work close to doctors. Those players, being much more experts in surgery rather than 3D printing, will most probably use the process parameters suggested by the equipment manufacturer for a given material. Financial benefits are also included in the usage of 3D-printed parts in biomedicine. A recent study found that nearly $4,000 would be saved by using a surgical guide in a maxillofacial surgical operating room that costs $179/minute to run (Ballard et al., 2019). Additionally, 3D printing modeling is not only cost-efficient for hospitals, but it is also beneficial for patients who are receiving higher levels of pre-operative and post-operative outcomes (Hanasono and Skoracki, 2013).

In an always-changing environment, organizations need to keep their market position high. For this purpose, they need to make use of their intellectual capital to transform passive consumers into prosumers for their own benefit will gain much more than a satisfied consumer. As a result, they will increase their profitability score and lifetime value and will become more visible and trustworthy in attracting new customers. The innovation appeared once prosumers were discovering new uses for the platform, moving away from just finding relatives through the web to connecting with them in engaging ways. Prosumers bring innovation, and that is because human capital results in creativity and openness, so prosumers bring innovation (Izvercian et al., 2025).

Thus, the production of AM parts for biomedicine represents a prosumer-oriented phenomenon (Santos, 2017). Nevertheless, energy consumption to manufacture 3D-printed parts on a large scale represents a non-negligible aspect that needs to be considered in this regard.

These considerations lead to the following necessities:

• The responsible energy consumption in manufacturing goods is a key factor to achieve sustainable development, to this aim the reduction of the energy required to print the parts is a very important objective.

• One of the critical points of AM techniques is the repeatability and the standardization of the process, when the same part is manufactured in a large number using several machines, even if of the same model, spread in different manufacturing environments and relying on the same process parameters this can be an issue.

It is clear, according to the existing literature, that these two necessities should be solved to enable the application of AM techniques for biomedical applications on a large scale. The scope of this piece of work lies on this path, and two research questions will be addressed:

To tackle these objectives a detailed experimental campaign, involving the printing of several jobs with different printers, has been conceived and carried out in this work. Stereolithography (SLA) is the 3D printing technique adopted to print the parts. It uses a powered laser to harden layers of resin that sit in the printer’s reservoir tank, it is also one of the most popular methods due to its high accuracy and smooth finish (Xu et al., 2020; Huang et al., 2020; Cozzolino et al., 2022). The overall process begins with the heating of the printer followed by the build plate lowering into the liquid resin vat. Computer-generated slices are then sent to the UV laser to begin the photopolymerization of the resin. Once the 2D layer is fully exposed to the laser, the resin adheres to the build plate and polydimethylsiloxane release layer. As the build plate lifts, the previous layer is peeled from the polydimethylsiloxane layer leaving it adhered to the build plate. These steps repeat until the 3D product is finished (Maines et al., 2021). Overall, compared to other 3D-printed methods such as Fused Deposition Modeling (FDM), SLA is less time-consuming, has better resolution and can produce complex parts that are crucial for biomedical prototyping and device production (Shahrubudin et al., 2019).

The potential biomedical applications of SLA are numerous. However, one of its challenges is the absence of biocompatible and biodegradable resin for biomedical applications. What is more, a significant limitation of SLA is that the complexity of various printing parameters poses a challenge in predicting and optimizing the performance of the printed parts. Also, energy consumption to manufacture 3D printed parts on a large scale represents a non-negligible aspect that still has not been well considered in the literature. Thus, there is an urgent need for advanced optimization and modeling tools capable of accurately predicting mechanical and environmental properties and optimizing functional parts tailored to specific applications.

This work aims to respond to that need. In fact, a joint investigation of energy consumption, mechanical properties and dimensional accuracy to evaluate the repeatability and energy efficiency of surgical guides, typically printed by SLA, represents the novelty of this study.

To make the results more useful for practical applications, different items have been printed: tensile specimens, specimens for geometrical accuracy analysis and case studies representative of real parts. The material used varies depending on the biomedical application required. In this study, the attention has been focused on two materials: a well-known one and an innovative one. As a case study, surgical guides have been considered. A surgical guide is printed for cases involving mandibular resection and reconstruction. Using a portion of the patient’s fibula and nearby vessels, the surgeon will reconstruct the patient’s mandible. This procedure is called a free flap mandible reconstruction and often uses surgical guides to help the surgeon place screws or make cuts where it is needed (Lakkala et al., 2023). 3D-printed surgical guides have proven to offer preoperative planning, an increase in reconstruction accuracy, and less time spent on the operation itself (Hanasono and Skoracki, 2013).

To fulfill the aims declared in the Introduction section, a two-step experimental campaign has been carried out. In the first stage, surgical guides were printed with different positions and orientations with respect to the printing plate to investigate the link between energy consumption and printing configuration. In the second step, dog bone specimens, accuracy cubes and surgical guides have been printed with two printers operating in two different production environments to investigate repeatability of the process. Multiple jobs have been printed with each printer, and energy consumption has also been recorded in this stage.

To replicate the design of traditional medical 3D printing, two biocompatible resin materials were used: Biomed Amber and Biomed White. Biomed Amber resin is a rigid material for biocompatible applications requiring usually short-term contact. Parts printed with Biomed Amber Resin are compatible with common solvent disinfection and sterilization methods. It is widely used for pre-operation planning and modeling; it is also certified for long-term topical skin contact and short-term mucosal contact. Biomed White is certified for the same applications as Biomed Amber but has additional certification for short-term contact with tissues, bone and dentin (Formlabs, 2023; Medical resins.). Table 1 contains the chemical composition of both Biomed Amber and Biomed White resins.

The Form 3B is the 3D printer used to manufacture the parts. It is a Low Force Stereolithography (LFS) biomedical printer manufactured by Formlabs. The printer uses a flexible resin tank along with a Light Processing Unit to create parts out of resin. Having a flexible resin tank allows for lower peeling forces during printing which can produce accurate and consistent prints (Formlabs, 2019; 4 Ways LFS 3D Printing Produces Better Parts.). It has a build volume of 145×145×185mm and a laser spot size of 85 µm. The power requirements for the Form 3Bs used in this experiment were 100–240V, 2.5A, 50 / 60 Hz and 220 W.

This part of the study consists of printing the same object – a surgical guide in 4 different orientations on the Form 3B 3D printer at NanoTech West (NTW) building, by using Biomed Amber Resin as material, and by measuring the energy consumption. The mass of the single object and its supports has also been measured. Four different orientations have been investigated, as reported in Figure 1.

Each different orientation involves a different support strategy and results in a different printed mass, the printed mass includes both the mass of the part and the mass of the supports. So, a balanced analysis will include both energy considerations and material use considerations.

Two different printers in two different production environments have been used. A total of nine print jobs were completed in two different industrial facilities: the Research Center (RschC) and the NTW. The two facilities are approximately a kilometer apart and use separate incoming power feeds with dedicated voltage conversion transformers. The RschC has dedicated step-down transformers that convert 11.500-volt industrial power down to 480-volt 3-phase power, which is further stepped down to 240-volt and 120-volt near the point of use. NTW similarly has a dedicated high voltage feed, however the high voltage is stepped down directly to 240-volt 3 phase. The result of this difference in power causes a couple of unique variations in the point of use/outlet power in the two buildings. The RschC has more system inductance due to the number of transformers connected both series and parallel. While the RschC power is more stable, the voltage at the point of use is unique to the transformer connected to that specific outlet. NTW is also running off an industrial connection, but it performs closer to a commercial configuration, where all the output voltage is consistent, and the building has far fewer transformers. For this program, the goal of incorporating different buildings with different power system designs was to determine if the stability and type of power system had a measurable impact on power usage. Each print job of this study consisted of three ASTM D638-22 Type I dog bones, 1 quality control cube and 1 cutting guide. A single print job, including the 5 parts previously mentioned, was printed on the same build plate using the same material. In particular, 3 jobs of Biomed Amber and Biomed White were run in the RschC and 3 jobs of Biomed Amber at NTW, for a total of 9 print jobs. It is important to underline that the two locations used separate printers, but both were of the same model.

Using PreForm software, STL models were loaded onto build plates using 100 µm layer thickness and auto-generated supports for the most efficient print. Figure 2 depicts the job to be printed in the main study modeled by using PreForm software.

Once finished, samples were post-processed using a FormWash machine with 99% IPA (Isopropyl alcohol) to rid the samples of excess resin. They were then cured using UV light in a Form Cure. Once cured, samples were bagged and prepared for sterilization. The samples were autoclave sterilized using a pre-vacuum cycle for 4 min at 132 °C.

The energy consumption has been continuously measured during all the jobs.

The two parts of this experimental campaign can be summarized in Tables 2 and 3. As mentioned above, in the first part of the study, just the object was printed in different orientations (Figure 1) whereas in the second part of the study the printed job included the part, the accuracy cube and the tensile specimens (Figure 2).

D638 Type I dog bone samples were tested on Instron 5958 pneumatic load frame at a rate of 5 mm/min. Each print job contained 3 dog bone samples; therefore 27 samples were measured over 9 prints. The tensile testing operating procedure was programmed in Bluehill software version 3.72 to include sample sizes. Mechanical properties deriving from the tensile tests were calculated from the resulting stress-strain curves for each sample.

The accuracy of test cubes was analyzed by measuring 12 points per cube, using 9 cubes in total (3 at NTW in Biomed Amber, 3 at NTW in Biomed White and 3 at RschC in Biomed Amber). Each cube measured 30 mm in all directions. One data point was recorded for each corner in the X, Y and Z directions. These were measured using a Mitutoyo Absolute Digimatic Indicator ID-C112EXB with recording capability and data were recorded using the U-WAVEPAK software. Cubes were measured before autoclaving and after autoclaving to analyze the accuracy regarding sterilization. When printing surgical guides and models, cubes are measured in quality control processes to analyze whether the printer is calibrated and running smoothly, therefore reassuring the quality of the guide or model set to enter the operating room.

Power and energy consumption of the Form 3B printer have been recorded using the Extech 382100 3-Phase Power Analyzer. It uses current sensors and tension cables with crocodile clips to measure current and tension data, respectively. As a result, it gives power and energy consumption over time, according to the sampling time chosen. Form 3B is provided with a single-phase two-wire connection. Thus, current and tension have been measured using one current sensor and two tension cables with two crocodile clips, respectively. The sampling period has been chosen equal to two seconds.

The total energy consumption and the mass of the material used to print the surgical guide in each position are given in Table 4. It is important to note that slight variations in mass among the guides printed may be attributed to the automatic placement of supports by the Preform software and to the tolerances of the equipment itself. It can be observed that orientation 2, where the surgical guide was placed horizontally, required the minimum energy consumption to print the object, including the surgical guide and the supports. On the other hand, this configuration required the maximum amount of printed material. To take into account the material used, total energy consumption related to the printing of the part E_t is expressed by the formula:

where E_pr is the energy due to the printing of the part whereas E_emb is the embodied energy, that is around 125 MJ/kg for resins (Augusto et al., 2021).

For the first part of the study, total mass m_tot is calculated as the result of the following formula:

where m_pr is the mass of the surgical guide whereas m_s is the mass of the supports. Table 4 contains all the results of the first part of the study.

As expressed above, three jobs have been printed, each job includes (as shown in Figure 2) three dog bone specimens, one accuracy cube and one surgical guide (printed in the orientation which minimizes the required total energy consumption E_t). Three jobs have been printed with printer 1 in NTW, and three more jobs have been printed with printer 2 in RschC. Moreover, three more jobs have been printed with printer 1 in NTW but with a slightly different material, namely Biomed White, to further investigate the reliability of the process. The following subsections illustrate the results of the main study in terms of mechanical properties, accuracy and energy consumption.

The mechanical properties of the printed materials were evaluated as described in the 2.4 subsection. Tensile testing results showed that NTW Biomed Amber had the greatest strength and consistency across three builds. The ultimate tensile strength ranged for 60–70 MPa. RschC Biomed Amber had tensile samples from builds 1 and 3 with similar tensile strength when compared to NTW. However, build 2 saw a significant drop in UTS with an average UTS of around 40 MPa. NTW Biomed White was observed to have the lowest UTS, also around 40 MPa. Elongation for NTW Biomed Amber and NTW Biomed White had similar results while RschC had increased elongation with subsequent prints. The results obtained agree with those stated in Formlabs datasheets (Formlabs, 2023; Medical resins.) (Formlabs, 2019; 4 Ways LFS 3D Printing Produces Better Parts.). Figure 3 contains the results of bar charts of tensile stress and tensile strain results at breaks between print jobs, locations and materials used.

As described above, the accuracy of test cubes was analyzed by measuring 12 points per cube, using 9 cubes in total (3 at NTW in Biomed Amber, 3 at NTW in Biomed White and 3 at RschC in Biomed Amber). Each cube measured 30 mm in all directions. One data point was recorded for each corner in the X, Y and Z directions. Table 5 and Figure 4 contain the results obtained in terms of accuracy. In total, there were 108 data points spread over 9 cubes. All measurements were recorded in mm and averaged within 0.5 mm of CAD dimension. It can be said that autoclaving reduces the point measurements but not largely. Previous evidence shows that steam sterilization on SLA-printed surgical guides does not have a significant effect on accuracy measurements (Sharma et al., 2025).

For the main study, the surgical guide, mechanical specimens and accuracy cube were printed in each job. Table 6 contains the results in terms of energy consumption to print the job. The variability of energy consumption has been investigated between two locations, NTW and RschC, and between two materials, Biomed Amber and Biomed White. For each combination of location and material, three jobs were printed to investigate the repeatability.

It was observed that time plays a crucial role in energy consumption, in accordance with the literature (Cozzolino et al., 2023). Thus, the higher the duration of the printing, the higher is energy consumption. According to our results, NTW required more energy than RschC to print the same job by using the same 3D printer and the same material, that is Biomed Amber. In this case, the time required to do it was very similar, so this difference is due to the power consumption, which was slightly higher for NTW.

Figure 5(a) and (b), depict the power trend over time to print the Biomed Amber in NTW and RschC, respectively. Figure 5(c) shows the power trend to print the Biomed White in NTW. Many power peaks over time and many oscillations can be observed for the printed job in NTW; that explains the higher energy consumption in NTW. This phenomenon is observable also by varying material in NTW, as observable comparing Figure 5(a) and (c).

By the results presented above, it appears clear that two energies should be considered: the energy required to perform the printing operations and the energy embodied within the material used, this latter contribution is strongly influenced by the supporting strategy adopted. It seems that the part orientation plays an important role in both these energies: let’s consider a part with an elongated shape (which is the case of the surgical guide). If the major axis is perpendicular to the building plate a higher number of layers will be required but fewer supports need to be used, on the other hand, if the major axis is parallel to the building plate a lower number of layers will be required while more material will be used in printing the supports. Considering only the energy needed for the printing process it can be seen, looking at Table 4 reported in the results part, energy increases with the increasing number of layers. Figure 6 reports the ratio between the energy consumption and the number of layers versus the number of layers, and it can be noted that this ratio does not have a clear trend but slightly increases with the increase of the number of printed layers.

A more interesting diagram, reporting the correlation between printing energy and printed mass, is reported in Figure 7.

It can be noted that the amount of energy required to print a unit mass of material linearly increases (almost) with the increase in number of layers. This further diagram confirms that the number of layers should be limited as much as possible to have an energy-efficient process. The embodied energy of the material increases the complexity of the evaluation. The histograms in Figure 8 report the total energy required (printing energy plus embodied energy of the material used) to print the surgical guides. The embodied energy was calculated considering the embodied energy of the material used and the mass of both the printed parts and the supports [Figure 8(c)], as reported in Table 4. Different scenarios, varying the embodied energy of the material, have been modeled and depicted in Figure 8.

It can be noted that a slight inclination of the parts can reduce the total energy required when materials with embodied energy higher than 100 MJ/Kg are used, while the “full vertical” configuration can become interesting only when materials with very high embodied energy (higher than 300 MJ/Kg) are used. The part orientation should be generally chosen aiming to minimize the number of layers, except when materials with high embodied energy are used, in that case a balance between the number of layers and the use of supports should be achieved.

As mentioned above, the purpose of this study is to examine and identify the most efficient way to print medical surgical guides manufactured by SLA while maintaining efficiency in energy consumption and product quality. In this regard, mechanical properties, accuracy and energy consumption were investigated by varying the material, Biomed Amber and Biomed White, and the locations. The preliminary study showed that orientation 2 (horizontal) required the minimum amount of energy consumption for printing among all the orientations. Thus, it was chosen as the best case to carry out the main study. By comparing the two locations under investigation, it was found that NTW required more energy than RschC to print the same job by using the same 3D printer, Formlabs 3B, and the same material, Biomed Amber. Power consumption was crucial in this case, which was higher for NTW. Moreover, by comparing the two materials used in NTW, there was an increase of up to 30% in terms of energy consumption by selecting Biomed White material rather than Biomed Amber as well as an increase up 27% in terms of printing duration. Thus, the higher energy consumption is due to the higher time required to complete the printed job. In this regard, NTW is more energy efficient than RschC as well as Biomed Amber helps to save energy consumption in 3D printing, rather than Biomed White.

In Figure 9, histograms summarizing the percentual variation for all the measured outputs with respect to the production environment (i.e. the different buildings were printers are located) and across the various jobs are depicted. The first interesting result is that the percentual variation for the dimensional accuracy is near zero [Figure 9(d)], while the variation of mechanical properties can be relatively high [Figure 9(a) and (b), ]. These results can be explained by taking into account that the geometrical accuracy is only influenced by the printing machine because relies on the precision of both the control system and the mechanical components of the machine. On the other hand, mechanical properties are strictly related to temperature. In particular, exposure time depends on the reactivity of the resin, and room temperature, thus affecting curing speed significantly. The warmer resin will cure faster, while the colder ambient temperature will slow down the curing speed. In general, the recommended room temperature is around 25 °C. If room temperature drops below 20 °C, issues with 3D prints can be noted (Steyrer et al., 2018). Also, mechanical properties are affected by process conditions whose stability is strongly influenced by the effectiveness and reliability of the electric supply system of the production environment. This consideration is further strengthened by the fact that a huge difference in variability has been observed between the two production environments, underlining the influence of the production environment on the mechanical properties of parts printed with the same equipment and under the same processing conditions.

Postprocessing of SLA parts includes mesh cleaning. Power and energy consumption were also measured during this stage after printing the job by using the Biomed Amber material in NTW. Table 7 contains the results in this regard. It can be observed that very little contribution to the overall energy consumption is given by the required mesh cleaning stage.

Figure 10(a) and (b), depict the current and tension over time, respectively, of the jobs printed by varying materials and locations. For both the materials used, Biomed Amber and White, a current of around 1200 mA has been found during the printing in NTW whereas a current of 1100 mA was found in RschC. Concerning the tension, deriving from the specific electrical system of the location considered, a voltage of 125V was found constantly in RschC while a voltage of 121V was found in NTW. These results justify the power consumption over time for each material and location selected. Moreover, these results underline the relevant contribution of the printing duration to the total energy consumption.

Thus, Formlabs settings are fine-tuned to each material. According to our results, Biomed White took longer to print, likely due to differences in proprietary print settings such as time between layers, specific power of the laser for each material, or laser exposure time/speed of the laser.

For a better visualization of the best strategy to adopt in terms of both mechanical properties and energy consumption, an energy index score for each material/print location was created by dividing the average tensile strength by the average energy consumed during the printing of three jobs for each location. The choice of this index derives from the fact that UTS can be seen as the objective to be maximized whereas the energy consumption is the cost to spend to be obtained. Table 8 contains the results of the energy index scores. As Biomed Amber and NTW had the highest tensile strength and lowest energy consumed while printing, it received the highest energy index score.

Thus, for our experimental campaign, Biomed Amber printed in NTW is the best combination of material used and location chosen to maximize the UTS while minimizing energy consumption.

Hence, compared to other 3D printing methods such as FDM, SLA is still a good choice to minimize the printing time while obtaining a good resolution especially for biomedical applications (Shahrubudin et al., 2019). In particular, 3D-printed surgical guides offer a lot of advantages such as an increase in reconstruction accuracy and a reduction in the time spent on the operation (Hanasono and Skoracki, 2013). Figure 11(a) and (b), show the model and the 3D-printed prototypes of both the surgical guide and the anatomic model, respectively. Figure 11(c) and (d), display pictures of the surgical guide implanted in a person in the operating room. It was printed by SLA using a Form 3B by using Biomed White as material. Biomed White is a resin recently produced by Formlabs that is certified for the same application as Biomed Amber, but it additionally can be used for short-term contact with tissues, bone and dentin. Thus, according to our results, it can be said that Biomed Amber resin is the best choice for applications where good mechanical properties are required. Otherwise, Biomed White seems to be now the best solution for applications having its material properties.

While mechanical properties are important for personalized surgical guides, there are other design requirements for surgical guides. Depending on intended use, these can include a matte finish to minimize reflection from lights in the operating room (OR) or radiopacity so that a surgeon can see the guide inside the surgery site when using real-time x-rays in the OR. Thus, further investigation of this study may include both additional characteristics to be investigated and optimized from a sustainable perspective.

The results of this work are valuable since they can be generalized to other objects printed by using the same materials. Printed just one object (maybe different from the one under exploration in this paper) by using one of two materials investigated in this work, all the other results can be proportionally estimated. What is more, 3D-printed models of surgical guides are also used not only by surgeons but also by medical students to enhance training in the operating room. Thus, this work highlights that 3D-printed models are not limited to operating room applications. They can also be used for disease modeling as well as medical device prototyping. These parts are produced in companies, hospitals, schools, clinics or small farms that work close to doctors. For these reasons, from a large-scale perspective, the energy consumption of 3D printing of surgical guides is not negligible and needs to be taken into account, as shown by this work. Hence, this study fills gaps of knowledge existing in the literature on the impact of the production environment and the repeatability of the results when the same procedures are implemented in two or more machines of the same model. Also, this work represents a starting point to optimize the productivity of big companies, where 3D printing workflow is automated and is becoming even more widespread. In fact, applications using biocompatible materials such as the creation of tissue without any damage to living cells, blood vessel production, dental implants and special medical prostheses are just some of the main contributions offered by 3D printing to the biomedical field. On these premises, it is clear that this phenomenon will become increasingly widespread and recurrent, so making 3D printing processes more robust, repeatable and energy efficient represents an imperative.

In this paper, the repeatability of the SLA process has been investigated. In particular, two different machines have been used in different production environments, that are NTW and RschC buildings, and the surgical guide for cases involving mandibular resection and reconstruction was printed by using two different resins, Biomed Amber and Biomed White. Energy consumption, mechanical properties and dimensional accuracy have been measured as reference output to assess the repeatability of the process. The results obtained allowed to release some guidelines to limit the energy consumption and to give some insights regarding the repeatability of the process. The main conclusions are the following:

• Biomed Amber resin is better than Biomed White in terms of mechanical properties. In particular, the UTS ranged for 60–70 MPa for Biomed Amber whereas NTW Biomed showed to have the lowest UTS, around 40 MPa.

• By comparing the two materials used in NTW, there was found an increase of up to 30% in terms of energy consumption by selecting Biomed White material rather than Biomed Amber as well as an increase up 27% in terms of printing duration. Thus, the higher energy consumption is due to the higher time required to complete the printed job.

• The NTW building required more energy than the RschC building to print the same job by using the same 3D printer and the same material, that is Biomed Amber. In this case, the time required to do it was very similar, so this difference is due to the power consumption, which was slightly higher for NTW. Many power peaks over time and many oscillations can be observed for the printed job in NTW; that explains the higher energy consumption in NTW.

• By fixing all the process conditions of the SLA process, the orientation of the part rules the energy consumption of the 3D printer. Generally, energy consumption decreases by minimizing the number of layers, thus reducing the time for printing. However, by choosing the horizontal orientation, by reducing the number of layers, the mass of supports increases. Thus, if also the embodied energy is considered in the energy consumption analysis, the support mass is not negligible. In particular, while using materials having a lower embodied energy, it is suggested to print the part by minimizing the number of layers, and so horizontally. Conversely, whether materials having higher embodied energy are used, it is better to print the parts vertically, even if the number of layers is the highest one because in this way the quantity and the mass of the supports are minimized. For other conditions, a trade-off is needed.

• Repeatability is kept quite stable and reliable by using the same machine in the same production environment. In particular, dimensional accuracy showed a low variability by varying machine and production environment, whereas mechanical properties showed higher variability because they are affected by resin temperature, room temperature, and electrical system of the building. Whether the electrical system is not in the best condition, a variation in terms of mechanical properties is usually observable, and this should be taken into account since the design stage as much as possible: they should be lower than expected. Thus, the production environment plays a non-negligible role in the output of the manufacturing processes, and it is necessary to be considered especially for scaling up 3D printing.

The authors would like to gratefully acknowledge the Center for Design and Manufacturing Excellence (Ohio State University) and the Fulbright Visiting Scholar Program.

Funding: This research did not receive any specific grant from funding agencies in the public, commercial or not-for-profit sectors.

Conflict of interest: The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Ethics statement: The authors declare no conflict of ethics.

Conflicts of interest: The authors declare no conflict of interest.

Author contribution: Ersilia Cozzolino: conceptualization, data curation, formal analysis, methodology, investigation, software, validation, visualization, writing–original draft and writing–review and editing. Gabriela Del Risco: writing–review and editing. data curation. Natalia von Windheim: conceptualization and writing–review and editing. Cameron Gygi: methodology and validation. Antonello Astarita: conceptualization, funding acquisition, supervision, and project administration. Nathan Ames: conceptualization, funding acquisition, supervision, and project administration.