1 Introduction

The now well-known technique of manufacturing by adding material layer-by-layer is called Additive Manufacturing (AM). This term was officially approved in 2009 by ASTM F42 and has been used worldwide since then [1]. AM is currently present in almost all industries and a wide range of materials, even from multiple materials, can be used to produce components, creating interesting composites that were previously impossible to produce [2,3,4]. In recent years, experiments have even begun with combinations of metallic materials that can combine, for example, the excellent mechanical properties of high-strength steel with the conductive properties of copper [5].

Among the most advanced AM techniques from the perspective of metal production is certainly the powder bed fusion (PBF) technique according to the latest ISO/ASTM 52900:2021 standard [6]. The technique is also well known as two trademark names, that is, Selective Laser Melting (SLM) or Selective Laser Sintering (SLS) depending on the degree of heat that is employed to either sinter or melt the input metallic powder. In general, the technology employs a laser heat source that selectively sinters or melts the powdered material layer by layer to produce a 3D object [7, 8]. The process begins with a 3D model sliced ​​into thin digital layers, which represent the individual powder layers to be melted. The file prepared in this way is then sent to the printer. The parts are then printed on a preheated substrate that has a similar chemical composition to the powdered metal that is being printed. At the beginning of the printing process, a thin layer of fine metal powder is evenly spread over a build platform inside a closed chamber, which is usually filled with an inert gas (Ar or N). Once the platform layer with the substrate is successfully melted, the substrate is lowered in the Z-axis by the height of the layer and a recoater (wiper or blade) spreads the powder in the desired layer thickness and the laser selectively sinters or melts the next layer. This process is repeated until the entire model is created.

During the printing process, the powder is not mechanically compressed, but a typical Gaussian single-component distribution with an fp factor between 40 and 60% can be observed when spreading it out [9]. The entire process takes place in an inert atmosphere to minimize the effect of oxidation and contamination from the external environment, which can cause failure of the structure [10]. Argon is typically used to fill the build chamber with an inert gas, which drives the oxygen concentration below 0.1% without creating a vacuum. During printing, an inert gas is continuously supplied, flowing over the substrate with the powder and constantly displacing residual oxygen to prevent buoyancy. In addition, the gas flow also carries away unwanted sintered material from the melt pool, which is created by process instability [10]. Tolochko et al. [11] reported in their publication that it is practically impossible to completely remove air trapped between the pores in the powder bed. This effect also negatively affects the laser absorption by the powder. Due to the unique characteristics of the PBF manufacturing process, parts produced using this method exhibit significantly different properties compared to their cast counterparts [12,13,14], in terms of mechanical properties, microstructure, defect types, etc.

Nowadays, PBF method has been widely used in many industrial fields, not only for its stable mechanical properties but also for its dimensional accuracy, which is the highest among the available metal AM technologies. However, the largest number of PBF-produced parts can be found in the aerospace and medical sectors, mainly because it can produce functional parts that are design-free, structurally complex, and in a relatively short time [15,16,17,18]. However, there are some technical problems that still remain unsolved. One of the biggest technical problems is the production of horizontal overhanging structures or bridging. It has been proven that without a substrate or a previous anchoring layer, undesirable balling and dross formation effects or deformation effects can occur, leading to failure of the entire print. The solution to the creation of horizontal overhanging structures has so far been to build support elements under the first layer [19].

In view of this, support structures play an important role in PBF printing, even though their presence is both beneficial and detrimental. The primary advantage of supports lies in facilitating heat transfer from the part to the base plate, providing effective cooling and thereby reducing residual stress, as noted by Zeng et al. [20]. Another clear benefit is the improved stability of the build during printing, as the part remains firmly anchored to the baseplate, preventing them against warping and distortion. In addition, printing parameters can be tailored separately for supports and the solid part, which can partially ease both the manufacturing process and subsequent support removal. Supports also serve the essential function of sustaining loose powder in the chamber, forming a base layer for printing before subsequent powder layers are applied, thus preventing collapse.

However, in PBF technology, the disadvantages of support structures often outweigh the benefits. In many cases, the volume of supports exceeds that of the part itself. Guoqing et al. [21] investigated the optimization of support structures with a focus on thermal gradients and stress, from which suitable geometries that reduced support volume were identified, though complete elimination was not possible. Greater support volume inevitably increases build time, which in turn drives up overall production costs. The geometry of supports also contributes to another drawback which is powder entrapment. Powder trapped within support structures can only be removed after the part is cut out off the base plate. During cutting, this trapped powder is exposed to the surrounding environment, often cooling medium, causing powder degradation that makes it unsuitable for reuse. Another significant disadvantage is the surface quality left behind after support removal. In a study on Ti6Al4V, Wadea et al. [22] found that manual removal of supports resulted in a surface roughness Ra of 27.25 µm, whereas removal via EDM produced an Ra of 4.01 µm. These findings highlight that the surface left by supports remains considerably rough and requires further post-processing. The ultimate goal, therefore, is to reduce or ideally eliminate the need for supports despite their functional role.

Zhang et al. explored in general different approaches to reduce or remove supports for metal printing [23]. In general, supports can be from the same material as the print (produced by PBF) or are from different materials as the print [produced by Laser Engineering Net Shaping (LENS), and Direct Energy Deposition (DED)]. Interestingly, for dissimilar materials, the metallic supports can be dissolved in 41 wt% nitric acids with the cost of long removal time of more than 10 h [24]. For same material supports, the only effective ways are manual with pliers or with power tools, such as mechanical saw, WEDM, CNC machines, etc. The supports are usually in the shape of cones or blocks that are easy to collapse under milling force. Thus, Cao et al. attempted to fill them with liquid resin and cured to increase the stiffness for easier machining [25].

As the reviewed literature shows, minimizing or eliminating support structures has been a critical requirement in metal AM since its early development. However, this challenge has not been widely investigated. This paper therefore reviews recent studies that attempted to achieve support-free metal printing with PBF. In addition, the underlying mechanisms behind related defects that involved in printing without support are discussed, and emerging strategies aimed at enabling truly support-free fabrication are highlighted.

This review is organized as follows. First of all, Sect. 1 introduces the necessary background toward the topic. Section 2 reports a bibliometric review about the research landscape into the support-free printing with metal materials using PBF technology. Guided by this literature review, Sect. 3 discusses different failing mechanisms of printing without support. Continuously, Sect. 4 presents the possible approaches to achieve support-free printing, complemented by case studies in Sect. 5. Finally, the review is concluded, and future suggestions are discussed in Sect. 5.

2 Bibliometric review

To gain an overview of the current state of the art, a literature search was conducted in the Web of Science database on April 9, 2025. The aim was to identify published articles related to the topic up to the date of the search. The search term used was TS = (("support free" OR "overhang" OR "dross" OR "downskin") AND ("Laser powder bed fusion" OR "Selective Laser Melting")). This resulted in a relatively limited total of 188 articles, with the composition shown in Fig. 1.

Fig. 1
Fig. 1
Full size image

Articles related to the topic and their composition

As in Fig. 1, the number of articles began to gain traction in 2016, increasing from a modest total of 4 to 40 articles in 2024, noting a decline to 27 in 2023. As of the date of the search, in 2025, 5 articles have been recorded. The majority of these (81%) are journal articles, while a smaller portion (13%) is conference proceedings, and the remainder are early access, reviews, or letters.

Based on the search results, VOSviewer software was used to generate network map for the keywords and the countries where the articles originate from. It should be noted that only keywords with a minimum of four occurrences were included in the analysis. In addition, the nearer the keywords, the more often they appear together in one article. The total link strength signifies the sum of its co-occurrence counted with all other keywords. The keywork network together with its statistics are reported in Fig. 2 and Table 1.

Fig. 2
Fig. 2
Full size image

Keyword network of the collected articles

Table 1 Keyword count
Full size table

It is worth noting that the keywords were subject to be cleaned before the creation of the network maps. Specifically, for terminology unification and clearer mapping, “selective laser melting” was changed to “laser powder bed fusion”, all the plural forms of the keywords, such as, “… structures” were returned to it singular form “…structure”. The “mechanical properties” remained to represent a variety of properties of the printed materials, such as hardness, tensile strength, etc.

As can be assessed together in Fig. 2 and Table 1, “laser powder bed fusion” and “additive manufacturing” appear in the center of the map. This is because they are terminologically related. Other keywords associated to the search topic that worth attention are “overhang structure” and “support structure”. When discussing the challenges of support-free metal printing, the typical properties assessed include “surface roughness”, “porosity”, “microstructure”, and “microhardness”, with “surface roughness” standing out as the most critical. This aligns with empirical knowledge that support structures play a vital role in ensuring the surface quality of overhang structures where they are most required. Specifically, without support structures, overhangs may fail to print successfully, or the surface roughness of the down-facing regions may deteriorate significantly, because the layers are built directly on powder without solid support.

The search results also show that the two most common metal alloys used in the reviewed studies are “alsi10mg” and “ti6al4v”, both of which are primarily applied in high strength-to-weight ratio applications such as aerospace. Among these, “alsi10mg” has attracted greater research interest, likely due to its lower material cost. This highlights a gap in research involving the more expensive and difficult-to-machine “ti6al4v”. Support structures for “ti6al4v” are particularly important, as they not only anchor the part to the powder bed but also facilitate sufficient cooling. Thus, removing support structures can be critical, because Ti6Al4V tends to develop significant residual stress from excessive heating and its high strength, making it prone to cracking and, ultimately, print failure.

Another notable group of keywords contains “topology optimization”, “finite element analysis”, “distortion”, “residual stress”, and “deformation”. It shows that there have been attempts to use numerical methods to study and understand the nature of printing with and without support, as well as to optimize the design for support-free printing. In addition, “machine learning” appears closely with “design for additive manufacturing” indicating that it has been explored as a means to improve the design for support-free printing. These advanced numerical tools are essential in reducing the R&D cost with designs and simulations before printing, especially for metal printing, which is expensive by nature.

The keyword network suggests that, to achieve support-free printing, recent research has focused on redesigning or minimizing the supports and structures of printed parts, with particular attention to overhang features. These strategies must ensure that the surface roughness of printed parts meets the standard requirements as of their conventionally manufactured counterparts, either in their as-built condition or with minimal post-processing. In summary, the keywords offer valuable insights for further assessment of support-free printing and help form the backbone of this study.

In addition, the map of the countries where the articles originated from can be observed in Fig. 3 and Table 2.

Fig. 3
Fig. 3
Full size image

Country network of the collected articles

Table 2 Country count
Full size table

The three countries that ranked the first are “usa”, “peoples r china”, and “england”. It is worth noting that the number of documents does not necessarily correspond to the number of citations the articles have received. Since support-free printing is a relatively advanced topic, it is typically pursued only after institutions from these countries have established a solid foundation in metal printing. This can partially reflect the level of funding and support that these countries have received for advancing research in the field of AM. Information on the affiliated countries of the reviewed papers can therefore serve as a useful indicator for identifying strong research teams and guiding future investigations.

3 Failure mechanisms of prints without supports

Summarized from the reviewed papers, this section describes the failure mechanisms of prints when no support is used. To achieve the goal of support-free printing, it is essential to understand the worst-case scenarios in their absence, so that proper measures to mitigate these effects can be conducted.

3.1 Warpage and deformation phenomena

The warpage phenomenon is closely related to the printing process and can be localized in the overhang region. It can result in an interruption of the printing or a total failure of the production, which has a negative impact on the final price and material consumption. This defect arises due to extreme differences in temperature gradients (up to 20 K/µm) and their subsequent cooling rate (up to 40 K/µs) [26]. These high-temperature gradients and fast cooling rates cause the material to develop large thermal residual stresses, which arise during the laser melting process of the layers and begin to develop and manifest themselves during the cooling phase of the material. It is necessary to take into account the powder metal itself (viscosity, thermal conductivity, and specific heat), which complicates the thermal gradients of AM.

Residual stresses mainly stem from three different sources: mechanical, structural, and thermal. During printing with PBF process, the laser beam melts the powder and creates a melt pool. In the vicinity of the melt pool, there is an intense localized shrinkage of the liquid metal as the material cools down. The phase transition of the material from liquid to solid as the molten metal rapidly cools down and solidifies. This is accompanied by a relatively significant volume change that leads to structural stress. Finally, when the metal in the solid phase cools to ambient temperature, the localized transient rapid heating and cooling during AM induces a strong thermal gradient that culminates in thermal stress. The thermal stress is accumulated as the new layers are melted upon this layers. The total stresses can exceed the yield strength of the material, leading to severe plastic deformation of the printed part during the printing process or after the part is removed from the powder (insulating layers) or from its supports.

These defects typically occur in regions where there are overhangs or bridging. Warpage is one of the most common defects that can occur during printing, the basic prevention lies in preheating the base plate (the preheating temperature is related to the material used). Another way to minimize defects is to understand the process parameters and their subsequent influence on the melt pool. Thus, optimizing the process parameters can be realized. Both of these methods aim to reduce residual stresses that subsequently cause the part deformation. Type I residual stresses, or macrostresses, in PBF parts in general is explained by two mechanisms: the temperature gradient mechanism (TGM) and the cooling phase model (CDPM) [27,28,29]. TGM creates residual stress in a single melt trace, because the molten material tries to expand due to the high temperature, but the surrounding cooler material prevents it from doing so. This causes compressive stress in the heated region. Once these stresses exceed the yield strength of the material, downward deflection occurs (in the opposite direction to the product growth), because there are no mechanical obstacles or support to prevent this. As the laser moves further and the region begins to cool, the material begins to shrink. However, this shrinkage is limited by the previous compressive stress, so tensile residual stresses are created, which cause deformations in parts manufactured by the PBF method. In contrast, CDPM describes the behavior of molten powder differently. The newly applied layer is warmer than the previous one. As it solidifies and cools, it begins to shrink. However, the previous layer partially inhibits this shrinkage—this leads to tensile stress in the new (top) layer and compressive stress in the layer below it. As additional layers are applied, tensile stress accumulates in the upper layers and compressive stress in the lower ones. As a result, the interior of the part has predominantly compressive stress, while the surface has tensile stresses. The reason for residual stress can be explained by the following equation for thermal expansion [30]:

$$\lambda =\alpha \bullet \Delta T\bullet L,$$
(1)

where \(\lambda\) ( −) is the thermal strain, α (1/K or 1/°C) is the coefficient of thermal expansion of the material, ΔT (K or °C) is the temperature difference (gradient), and L (mm) is the diameter (length) of the material. The equation shows that the thermal strain of the material is in direct relation with the coefficient of thermal expansion and the temperature gradient, which is relatively severe due to the rapid heating–cooling of the PBF process.

It can be drawn from Eq. 1 that the greater the temperature differences and cooling rate, the greater the residual stress that arises. It is important to note that the support structures not only serve as physical support under the overhang but also serve to transfer heat from the molten pool to the base plate, thereby reducing thermal gradients and RS [31]. Therefore, to achieve support-free printing, it is necessary to understand and mitigate this thermal effect. Table 3 contains updated published studies addressing the warping problem.

Table 3 Selected case studies on the methods to facilitate support-free printing against warping
Full size table

From the analysis of available articles, it follows that warping defect is a critical factor when trying to print without support and is directly linked to thermal changes during printing. The basic measure is preheating the substrate to reduce the thermal gradient that is caused during the printing process. Several different approaches have been published to reduce warping, for example by changing the scanning strategy, or changing the geometry of the part or optimizing the support structure; see Table 3. However, no research has been able to find a way to completely eliminate warping, and therefore, it still remains one of the biggest obstacles in printing without support.

It can be drawn from Table 3 that the most effective approach is to combine simulations and experiments, which thereby can provide quantitative design guidelines for printing parameters, such as hatch distance, time delay, overhang length, etc., in relation to warpage and deformation. On the other hand, the main drawback lies in transferring these methods to different printing setups with other material, geometry, and machine. Therefore, its application is more suitable for research purposes or small-scale productions in light industry.

3.2 Dross formation phenomena

Dross formation is a typical manifestation of a defect in the PBF process, which can be corrected to some extent by creating support structures. It is also manifested by a deterioration in surface roughness. Dross tends to form in overhang areas when the molten pool expands, leading to its downward displacement along the build direction. The extent of this sinking is influenced by the overhang angle and the energy density of the laser [40]. Dross formation is directly influenced by the Volumetric Energy Density (VED) factor, where more energy input leads to higher energy accumulation, which does not only melt the material but also subsequently creates a larger dross effect. Charles et al. [41] investigated the effect of dross formation on dimensional accuracy via numerical and experimental investigations and verified that the thermodynamic equation

$${E}_{V}=\frac{P}{vht},$$
(2)

where \({E}_{V}\) (J/mm3) stands for the amount of energy input into a unit volume of material, P (W) denotes the power of the laser, v (mm/s) is the scan speed, h (mm) is the hatch spacing, and t (mm) is the layer thickness. It can be used to regulate the dross effect, and the optimal setting can mitigate this effect. It was also experimentally verified that the most influential parameter is the laser power and the scanning speed. Similar research was carried out by Lin et al. [35], who concluded that the variation in thermal conductivity between the powder bed and the solid support region results in differences in molten pool size, which is the primary factor contributing to dross formation. In the powder-supported regions, or the regions without any support structures, the molten pool tends to sink into the underlying powder due to the combined effects of gravity and capillary forces, leading to the adhesion of numerous powder particles to its bottom surface. Moreover, the repeated thermal cycling inherent to the PBF process further complicates the behavior, as previously solidified layers influence the dynamics of subsequent melting and solidification. The schematic of dross formation can be observed in Fig. 4.

Fig. 4
Fig. 4
Full size image

Dross generation formation, schematic diagram [35]

Chen et al. [40] demonstrated that regions with high thermal gradients are typically associated with elevated melt pool flow velocities. These high-gradient zones are predominantly located beneath the tips of overhang structures, which promotes the formation of dross. Through a combination of experimental investigations and numerical simulations, the authors established that increased VED leads to the development of larger melt pools characterized by steep thermal gradients. This condition significantly contributes to excessive dross accumulation. Generally speaking, the dross formation is associated with the amount of energy input, commonly characterized by VED measure in PBF system, and heat transfer between individual layers. This type of defect has been mitigated in the existing literature using several experiments and simulations.

4 Mechanics to achieve support-free printing

After understanding the failure mechanisms, this section examines the measures used to counteract them to enable support-free printing. A number of approaches are presented, accompanied by a detailed discussion of their underlying principles.

4.1 High-energy downskin

A general rule of thumb for BPF technology, specifically metal PBF printing, is that an angle lower than 35° cannot be printed without the help of support structures [42,43,44]. The size and length of the overhang is related to the layer thickness and also depends on the material [19]. The extent of the overhanging section of the structure is determined by the following formula [45]:

$$a=h\bullet \text{cot}\theta ,$$
(3)

where \(a\) (mm) is the length of the overhang, h (mm) is the layer thickness, and θ (°) is the angle between the contour of the layer and the overhanging surface. To overcome the need for support construction and without the necessary post-processing step, some methods have been investigated and presented with promising results, even the ones that can deliver “support-free” printing. Yuan et al. [46] investigated the defocused printing method with a new strategy of scanning laser. By reducing the scan speed and increasing the laser spot size, the balling effect was suppressed, and the so-called horizontal overhang was created. The laser beam was defocused following Gaussian distribution in Fig. 5, which allowed the foundation of the first layer to be established on the cuboid supports and subsequent extension of the printing was realized by refocusing the laser as the printing continued under optimal conditions. Other authors have also addressed the optimization of process parameters to achieve almost support-free printing [43, 47,48,49].

Fig. 5
Fig. 5
Full size image

Schematic of energy distribution of defocused laser beam

This method is specifically suitable for printing of components with enclosed geometries, such as impellers. It simplifies manufacturing with fewer use of supports, while maintaining the printing speed and quality. This is because larger and more stable melt pool is formed that results in better layer adhesion in critical overhang regions. A notable feature of the method is the downskin surface it produces which is considerably rougher than that of the ones with supports. However, it is only limited to the downskin area. Therefore, for practical scenarios, it is recommended to plan the part orientation and dimension carefully, so that the downskin area is well accessible for subsequent machining or that they are not the functional surfaces. This should be accompanied by further optimization of the printing parameters, and proper post-processing with technology such as abrasive flow machining (AFM).

4.2 Heat-support method

Another approach is based on the use of “heat support”, which also provides horizontal overhangs. Several authors [50, 51] including Cooper et al. [52] investigated the implementation and effects of contact-free support structures focusing on overhang surfaces with varying inclination angles and gap distances between the supports and the part. Unlike conventional supports that connect to the overhangs, heat-supports maintain a controlled powder gap (see Fig. 6), providing thermal conduction and melt pool stabilization without direct contact, thereby reducing surface roughness and minimizing thermal-induced distortions.

Fig. 6
Fig. 6
Full size image

Demonstration of adding a heat support beneath the printed part with controlled powder gap: a from left to right, gap of 0.49 mm, 0.63 mm, and 0.77 mm; b from left to right: the heat support with thickness of 10 mm and 5 mm [52]

The results demonstrated that with an optimal gap of 0.63 mm (Fig. 6a) and optimal thickness of 10 mm (Fig. 6b), the parts can be successfully produced without being fused to the heat support or not supported at all. The research demonstrated that optimizing the gap distance and heat support’s thickness is critical to prevent powder sintering or loss of thermal conductivity benefits, resulting in improved surface quality across all tested overhang angles. This novel support strategy also reduces post-processing time and risk of component damage associated with support removal.

The main advantage of this approach is the improved surface roughness of the downskin, because the supports are not adhered directly but separated by a thin and compact powder layer. The in-between powder layer also helps to improve heat distribution. As a result, less effort is required for post-processing on support removal and surface treatment. However, this method requires precise control of the gap size. If it is too large, the heat transfer efficiency decreases, causing the part to distorted upward. On the other hand, if it is too small, unwanted sintering of powder can happen that complicate the post-processing. In addition, the effectiveness of this method can be reduced in extreme cases, such as printing a long overhang or highly complex geometry. The industrial use of this method is recommended for cases where surface quality and post-processing speed are critical, and where support removal is challenging, for example, in orthopedics, aerospace, etc.

4.3 Adding surface-active elements

A completely different approach to support-free printing was used by Chen et al. [40], where they study the addition of small amounts of surfactants, such as boron and niobium, to standard powder alloys. Surfactants have the function of reducing the surface tension of the melt pool during PBF. This is of great help in increasing the stability of the melt pool and improving the quality of processing overhanging structures using PBF. In other words, adding these reinforcement elements will affect the Marangoni flows, reduce the surface tension of the melt pool, and lengthen it, which may lead to the elimination of the need for support structures. This topic is also addressed in studies [53, 54].

However, research in this area is still relatively limited, rendering its wide adoption in industry. The method is also possibly limited by uneven distribution of chemical elements, which can result in defects and undesirable part properties. In summary, the method shows great potential in achieving support-free printing, but more experiments and careful process control are needed for broader practical applications.

4.4 Optimization of support or component shapes

Another approach, which is often found in publications [55,56,57,58], approaches the solution of support-free printing by changing or optimizing the shape of the printed part or the support structure itself. Barroqueiro et al. [59] use in their study a computational methodology that allows the design of self-supporting structures using topological optimization (TO) with respect to AM constraints, such as minimum feature size and overhang restrictions. By combining TO with a simple manufacturing model, it is possible to generate initial designs that are better adapted to the capabilities of PBF and minimize the need for support structures. The authors compare different numerical approaches (Solid Isotropic Material with Penalization, Rational Approximation of Material Properties) and show that the improved model (using the so-called softmax operator) increases the accuracy and capability of the algorithm. In addition, Li et al. [60] also approached the design of parts without support structures with TO, but using a different methodology. It includes analysis of the original part, optimization of material distribution, interpretation of TO results into an editable 3D model, adjustments for self-supporting, and final verification using FEM and real production. In a specific application of a truss beam, the authors managed to achieve a 31% volume saving compared to the original model and produce it without support; see Fig. 7.

Fig. 7
Fig. 7
Full size image

Support-free design proposal, where supports are part of the model [44]

In general, optimization of support and shape combines material savings with improved quality and possibly enhanced mechanical properties of the part. Because it is a numerical method, the targeted effect can already be assessed before the actual production. The drawbacks, however, lie in the demands for manufacturing and programming expertise, software, and computational capacity. Moreover, it is essential to obtain accurate data input of material properties, printing process, and parameters. This method is recommended for a wide range of industrial applications, taking into account the aforementioned requirements.

5 Summary and outlook

The studies mentioned in this section show that the problem can be approached using various techniques. The most promising results are shown by optimizing the process parameters of the lower layers, or the "high energy down-skin method", which is also the most developed because of its simplicity. This method focuses on individual printing parameters and is usually accompanied with substrate preheating. It is also necessary to realize that for each type of metallic powder, different parameters need to be optimized. In addition, it is necessary for the parts to meet the requirements for dimensional accuracy and optimal mechanical properties. Table 4 lists and summarizes other mechanisms.

Table 4 Summary of approaches to achieve support-free printing in literatures
Full size table

Based on the literature review and the aforementioned assessments, the criteria that designers need to take into account while attempting to control the quality of the overhang features to achieve the support-free printing are summarized in Fig. 8.

Fig. 8
Fig. 8
Full size image

Mindmap for consideration for support-free printing. Adapted from [45]

To control the overhang quality, it is important to consider two primary groups of factors that are process-based (energy input, scanning region, and residual stress) and design-based (support structure). For printer operators, the most accessible approach is to examine the process-based factors during build preparation in the slicer. A more advanced approach, where programming and designing expertise are required, involves optimization of the supports and printed components. It is worth noting that these approaches can be combined to achieve optimal support-free printing, depending on the operator’s objectives and capabilities.

6 Case studies

Among the reviewed scientific articles, the majority focused solely on explaining the support-free printing mechanisms, with limited representation of real-world case studies. Therefore, in this section, some typical case studies are selected for presentation. It is worth noting that each case study uses a different approach to eliminating or reducing the number of supports.

One typical example is the impeller, whose production by a method other than AM is very complicated. Two examples of a closed and open impeller profile can be seen in Fig. 9. For the closed impeller, it is necessary to eliminate supports mainly inside for smoother surface finish and simplicity of the support removal. While for the open impeller, the problem lies in supporting the curved blades where downskin and upskin of the profile needs to be with sufficient surface roughness for functional purposes.

Fig. 9
Fig. 9
Full size image

An example of a typical geometry of an impeller

In view of this, Kumar et al. [64] managed to produce support-free closed impellers with diameters of 10.25 and 20.5 cm from IN625, including overhangs as shallow as 15° and up to 60 mm in length without support structures. By inspecting the inner geometry and orienting the impeller with an appropriate angle, their goal of eliminating the internal supports was achieved.

Optimization of downskin and upskin parameters was also addressed in the article by Tian et al. [65], which mainly addressed surface roughness parameters on parts without support, that is, an open impeller part made of Ti-6Al-4V material. The primary concern with such a geometry is the open curved blades with functional surfaces. In the work, it was found that downskin surfaces exhibit significantly higher roughness than upskin surfaces, especially at lower overhang angles, due to powder adhesion and insufficient support. By employing an RSM using the biharmonic spline interpolation method, the process parameters were optimized. This resulted in simultaneous improvement of the surface quality of both skin types, achieving up to a 50% reduction in roughness compared to standard parameter settings.

A different approach was taken by Biedermann et al. [66], who addressed the printing of fluid flow channels such as hydraulic manifolds, by changing the shape of the channels following automatic shape generation to comply with the overhang constraint in AM. They designed and experimentally verified several shapes, of which three proved to be successful, namely the “droplet, diamond and house”; see Fig. 10. The study applied the procedures to generate production-ready 3D part designs for different orientations to the build direction and modified pathways of flow channels.

Fig. 10
Fig. 10
Full size image

Automated generated shapes of the channels [66]

The optimization of the shape of printed parts was also addressed by Mikki [67], who examined self-support TO considering distortion for metal AM. Specifically, the study proposed a TO method that integrates self-supporting constraints along with thermal and distortion considerations to enable support-free, high-precision AM. A novel overhang constraint based on a Helmholtz-type PDE and thermal and mechanical models addressing overheating and distortion were formulated and validated through 2D and 3D numerical examples. The author demonstrated their algorithm by designing a heat conduction component as can be seen in Fig. 11. The results demonstrate that the proposed method effectively suppresses dripping and distortion, achieving manufacturable designs with minimal impact on structural performance.

Fig. 11
Fig. 11
Full size image

Distortion results of the heat conduction model inherited from the printing process: a without any constraint, b with the self-support constraint, and c with the self-support and distortion constraint [67]

7 Conclusions

Laser beam technology for metal printing using the PBF process is based on local energy input followed by layering of metal alloy powders, making it a very advanced AM technology. The scale and processability of alloys is limited only by the weldability of the given alloy. However, production is limited by specific geometries, especially low-angle overhangs as the printed components have to be anchored to the baseplate with support structures for cooling and structural stability during printing. Due to the nature of the process, the extreme thermal gradients and rapid cooling arisen due to localized laser melting lead to the formation of significant residual stresses. These stresses are particularly problematic in unsupported areas, such as low-angle overhangs. In addition, the low thermal conductivity of unsintered powder areas impairs heat accumulation, leading to keyhole porosity and balling phenomena. Furthermore, this phenomenon leads to dross formation, staircase effect, warpage, and other surface defects in areas where there are no supports.

The article critically reviews scientific articles dealing with support-free printing and their underlining mechanisms. Most of the reviewed articles focus on optimizing the printing parameters of the downskin section on the part, but there are also publications that try to use other approaches. According to the research, no comprehensive and systematic solution has been proposed yet that would be able to print satisfactorily without support. In the future, it is therefore necessary to develop a reliable and repeatable method that allows printing of overhanging structures with a minimum angle, which would also ensure sufficient internal density and a high-quality surface finish.

Although there are several publications that discuss the theoretical foundations of layered PBF printing technology, few focus on the support-free printing and the information is scattered across different sources, which complicates their systematic summary. This review presents a comprehensive theoretical explanation of PBF technology with an emphasis on overhang structures and extends it to specific case studies. Its goal was to create an overview of selected studies that offer specific methodological guidelines for solving the issue of overhang printability with different approaches.

Research wise, advanced physics-based deep learning (DL) has been investigated in AM modeling and optimization [68]. Therefore, a promising path of development could be real-time adjustment of printing process parameters with machine learning (ML) algorithms [69]. In addition, recent studies also discussed the possibility to change the laser beam profile from Gaussian to doughnut shape for wider and shallower melt pool, which can potentially reduce the residual stress and the need for supports [70, 71].

In addition, hybrid solutions can be considered, for example, printing strategies that combine process optimization with alloy design can yield benefits from both material and process perspectives. It is worth mentioning as well that TO can be employed as well for optimal designs. Another approach to support-free printing is to shift the focus, that is, instead of eliminating supports entirely, they can be printed as usual, but with improved methodologies for efficient removal. This would avoid increased post-processing costs and preserve surface quality. Ultimately, research in this field should aim not only to improve printed surface quality but also to enhance the sustainability of the AM process itself, through reduced material usage, lower energy consumption during printing, and significantly decreased post-processing time, which directly translates into cost savings.