Large-format additive manufacturing of polymers: a review of fabrication processes, materials, and design

ABSTRACT

Large-format 3D printing for polymers enables cost-effective mass customisation and production of structurally robust, large-scale components for industries like aerospace and automotive. This review analyses additive manufacturing scalability, including throughput, volume, and essential criteria for 3D printing techniques. Challenges in large-scale polymer additive manufacturing are explored, including material selection, interlayer bonding, surface quality versus production speed, recyclability of materials, and post-processing. Materials development is found to be crucial for addressing thermal shrinkage issues, with solutions involving process control and fibre reinforcement while considering rheological properties and nozzle clogging. Balancing production speed and surface finishing in material extrusion 3D printing involves factors like print speed, nozzle size, and innovative designs to optimise throughput and surface quality. In large-format 3D printing, meticulous process control and quality assurance are vital to ensure the expected printing outcomes and defect-free parts, given the substantial material and energy investment.

KEYWORDS:

• Pellet-based printing
• large-scale 3D printing
• composites
• robotics
• advanced manufacturing

1. Introduction

As we stand on the cusp of a fourth industrial revolution, additive manufacturing (AM), commonly known as 3D printing, has taken a central role in driving the transformation. Originally developed for prototyping, AM has quickly expanded its application range, particularly with polymers, and is now a viable alternative for conventional manufacturing processes. According to market reports, the market for polymer 3D printing reached $11.7 billion in revenue in 2020, and the demand for the production of functional parts is expected to grow [1,2]. Yet, while the potential benefits of AM are vast and varied, there remains a significant obstacle to overcome: scaling the technology to produce large-format parts.

The need for large-format additive manufacturing (LFAM) has been established by numerous industries. The aerospace and automotive sectors, for example, demand substantial, single-piece components that can only be produced through LFAM [3]. Producing these large parts in one piece not only simplifies the assembly process, but also significantly enhances the structural integrity of the components. This advancement in manufacturing techniques has the potential to bring significant changes to various sectors.

One of the pivotal advantages of LFAM is its potential to contribute to sustainability. AM inherently minimises material waste by building parts layer-by-layer, only adding material where necessary. LFAM extends this benefit, offering the possibility of constructing larger structures with minimal waste and energy use, a key consideration as we strive for more sustainable manufacturing processes. It was estimated that by 2025, AM has the ability to save expenses by 170 to 593 billion U.S. dollars, primary energy consumption by 2.54 to 9.30 EJ, and CO₂ emissions by 130.5 to 525.5 Mt [4]. Furthermore, LFAM offers unprecedented possibilities for design innovation and customisation. The constraints of traditional manufacturing methods, such as the need for moulds or tooling, do not apply to AM. This freedom opens up a vast design space, allowing for the creation of complex, intricate structures that could not be produced by any other method.

Despite these advantages, there are considerable concerns to be addressed in the implementation of LFAM. The quality of the finished product is paramount, as is the speed of manufacturing. Balancing these two factors can be a delicate task; rushing the process may lead to an inferior product, while taking too long may negate the efficiency advantages of AM. Similarly, cost-effectiveness is a crucial consideration. While AM can offer significant savings in terms of reduced material waste and simplified assembly, these benefits need to be weighed against the investment required for large-scale 3D printers and their operation. The versatility of AM, in terms of the range of employable materials, is another major consideration. While small-scale AM for polymers has proven successful, there is a question of whether these results can be replicated when manufacturing larger components, potentially involving different, more advanced materials. Lastly, the scalability, reliability, and repeatability of the process are essential considerations when evaluating LFAM's viability. As the operation's scale increases, it is imperative to ensure consistent quality control, process stability, and the ability to reliably reproduce results.

The field of large-format 3D printing has been extensively explored in recent literature, highlighting advancements and challenges across various applications and materials. The review by Ngo et al. [5] provides a comprehensive examination of materials, methods, and applications within additive manufacturing, setting a foundational understanding of the 3D printing landscape. Similarly, contributions by Alghamdi et al. [6], Najmon et al. [7], and Salifu et al. [8] delve into material progress, aerospace applications, and automotive structures respectively, indicating the vast reach of 3D printing.

Despite these general reviews, specialised works such as those by Tiwary et al. [9], Love et al. [10], Moreno Nieto et al. [11], and Vicente et al. [12] focus explicitly on large-format additive manufacturing (LFAM), discussing post-processing techniques like welding to join 3D printed parts, barriers in polymer additive manufacturing, and recent developments in pellet extrusion and polymer composites. These works are critical in addressing specific challenges such as build volume limitations and the integration of large-format printing within industrial applications.

Additionally, Pignatelli et al. [13] present an application-and market-oriented review focusing on polymer pellet-based 3D printing and provide a list of equipment for LFAM, while Urhal et al. [14] emphasise the role of robot-assisted additive manufacturing, underscoring the interplay between robotics and large-scale 3D printing. These reviews collectively enhance the understanding of large-format 3D printing, a rapidly growing domain with significant potential for innovation and industrial application. Despite the abundance of review articles available, the pace of technological advancements in large-format 3D printing is so rapid that there exists a need for a comprehensive review that encapsulates the recent and significant developments in this field.

This review seeks to comprehensively covers the various types of LFAM and delves into some concerns and challenges regarding LFAM. In this review, we will adopt the definition of LFAM as proposed by Moreno Nieto and Molina, which characterises LFAM systems as those 3D printers with a print volume exceeding one cubic meter [11]. The literature survey for this review was conducted through Google Scholar, utilising a variety of keywords in different combinations, including ‘large-format’, ‘large-scale’, ‘additive manufacturing’, ‘3D printing’, along with the complete names of several AM processes. The review also contains a critical discussion on the specific challenges faced in various AM techniques including vat photopolymerisation, powder bed fusion, and material extrusion. A section on the application of large-scale 3D printing is also included to give the readers an understanding of the transformative impact this technology has on industries such as energy, automotive, and aerospace. This includes a comprehensive exploration of the latest advancements, real-world examples, as well as the challenges and solutions in scaling up 3D printing processes. Furthermore, this review also discusses potential future research directions in realising the deployment of LFAM technologies.

2. Various AM techniques for large-scale fabrication

In this section, 3 common categories of AM techniques for large-format polymer printing will be discussed. They are vat photopolymerisation, powder bed fusion, and material extrusion. The advantages and disadvantages of the AM techniques in terms of the scalability, throughput, surface finishing, material choices and cost will be touched on.

2.1. Vat photopolymerisation

Vat photopolymerisation refers to a group of AM techniques that involves the use of a liquid resin that contains photoactive polymers stored in a vat or container [15]. A build platform is immersed in the resin, and it gradually lifts the object being printed out of the liquid. There are typically two variants of vat photopolymerisation, namely stereolithography (SLA) and Digital light processing (DLP). They differ in terms of the method used to cure the resin. In SLA, the resin is then exposed to a specific light source, such as a UV laser, which selectively cures or solidifies the resin layer by layer, following the desired shape of the object. Once a layer is cured, the build platform moves up, and the process is repeated until the complete object is formed. In DLP, instead of UV laser, a UV projection system is used.

One of the primary advantages of vat photopolymerisation is its high precision. The technology offers exceptional accuracy and detail resolution, making it suitable for applications that require intricate geometries and fine features. The layer-by-layer curing process enables the creation of complex shapes with smooth surfaces, resulting in objects that closely match the intended design. SLA printed parts exhibit excellent surface quality, with surface roughness of sub-microns level [16], reducing the need for post-processing or additional finishing steps. This advantage is particularly valuable in industries like automotive, aerospace, and consumer goods where aesthetics and texture matter.

Currently, there are several commercially available large-format vat photopolymerisation printers in the market. For instance, CMET developed Rapid Mister ATOMm-8000 [17], which has a build volume of 800 × 600 × 400 mm. ProtoFab developed SLA2400, which offers a build volume of 2400 × 800 × 800 mm [18].

DLP technique is commonly used in large-format printing. DLP is one version of vat photopolymerisation that employs a projection system rather than a scanning laser beam [19]. It uses a digital micromirror device (DMD), consisting of thousands of microscopic mirrors arranged in a matrix on a semiconductor chip, to project a complete digital image onto the resin vat. As the DLP polymerises the entire layer at once, DLP technique tends to be faster than the scanning laser beam. However, the resolution and build space of DLP printers depend on factors such as the number and size of micromirrors, as well as the lenses used for focusing the projected light. In normal projection, larger magnification would result in lower resolution due to the fixed pixel size to image projection ratio. Apart from that, the intensity of the UV light reaching the print bed would be significantly lesser as the magnification becomes larger, thus increasing the resin photopolymerisation time.

One of the challenges in the DLP technique is the need for a peel step to separate the transparent bottom of the vat from the print, which is the slowest part of SLA/DLP technique. To solve this, CLIP (Continuous Liquid Interface Production) was invented which uses an oxygen-permeable membrane which creates a ‘dead zone’ at the bottom of the resin pool [20,21]. This prevents the resin from attaching to the bottom and allows for the continuous exposure process, unlike DLP where each layer must be separated from the resin tank floor. Nonetheless, it still has got its limitation. For instance, the print speed is still limited by the curing time and the time needed for the resin to flow to fill the gap created by raising the bed for the new layer. It is imperative to maintain a consistent resin supply from the substrate level to avoid resin depletion. Insufficient resin flow can lead to the formation of voids within the printed structure, compromising the integrity and homogeneity of the final product. This limitation necessitates careful consideration of the printing parameters and apparatus design to ensure that the resin dynamics are conducive to the continuous and uniform construction of the object, thereby preventing any discontinuities or structural deficiencies.

Efforts have been made to enlarge the print area in vat photopolymerisation technique. One of them is known as scanning projection stereolithography (SPSL) (Figure 1(a)), which combines scanning and projection methods simultaneously, reducing printing time and polymerisation overlapping [22]. BMF has introduced 3D printers with similar concept that can move either the projector while maintaining a stationary vat or move the vat while keeping the projector stationary [23]. In another work, Tripteron parallel mechanisms with orthogonal three-DoF have been used to move the projection module to different positions, offering increased accuracy [24]. By shifting the printing area, these printers enable larger parts to be built with good surface finishes. However, the stitching of images after reaching a certain size may result in poorer mechanical properties. The stability of these mechanical moving systems is critical to achieve high dimensional accuracy compared to standard DLP processes. To address the problems associated with the movement of the optical projection unit, a multi-projector DLP equipped with energy homogenisation was employed, expanding the construction area without shifting the projection unit [25]. However, merging the projected images from two projectors can be problematic when trying to maintain specific pixel intensities. Another system, known as the double mask projection stereolithography (DMPSL), was introduced to enlarge the build dimension, integrating a large envelope projection (LEP) module for scale and a high-accuracy projection (HAP) module for detail [26]. Techniques like straight line and sawtooth splicing have been implemented to reinforce the bond strength at the splice point.

Figure 1. Various setup to increase the build volume in SLA. (a) Projector system attached to the translational stage (Adapted from [26,28]) (b) a dead layer–free approach to rapid SLA printing, HARP (high-area rapid printing) (Adapted from [27]).

The heat generated during the polymerisation process is another major challenge in vat photopolymerisation, which limits printing speeds and the overall size of the printed object. To address this issue, Walker et al. have developed a novel approach using a nonreactive fluorinated oil as a mobile liquid interface for the DLP technique [27] (Figure 1(b)). This interface reduces adhesive forces between the interface and the printed object, enabling a continuous and rapid printing process regardless of the type of polymer used. Unlike traditional thermal limitations, the flowing oil facilitates direct cooling across the entire print area, allowing for larger bed areas. The researchers achieved impressive continuous vertical print rates exceeding 430 millimetres per hour and a volumetric throughput of 100 litres per hour. Furthermore, they successfully printed proof-of-concept structures using various materials such as hard plastics, ceramic precursors, and elastomers. In fact, they have successfully printed a 1.2 m tall prototype with a 30 × 30 cm base using the HARP technology in 3 h. This innovative technique holds promise for advancing large-scale and high-speed 3D printing.

2.2. Powder bed fusion

Powder Bed Fusion (PBF) refers to a group of additive manufacturing techniques that includes Selective Laser Sintering (SLS) [29], Selective Laser Melting (SLM) [30], and Electron Beam Melting (EBM) [31]. Among these techniques, SLS is commonly used to fabricate polymer parts. These techniques operate based on a shared working principle. A thin layer of powdered material, which can be plastic, metal, ceramic, and so on, is evenly spread over the build platform. Then, a heat source, usually a laser or an electron beam, scans the surface according to the cross-section of the part being printed. This process selectively melts and fuses the powder particles together. The build platform is then lowered, another layer of powder is spread, and the process repeats layer by layer until the part is fully constructed. Upon completion, the unfused powder is removed to reveal the final product.

The PBF technique offers several advantages. It produces parts with high resolution and good surface quality, making it suitable for many functional, aesthetic, and precision applications [32]. It works with a broad range of materials, further increasing its versatility in use cases [33]. One of its unique strengths is the ability to create complex geometries and internal features, without the need for support structures, reducing the need for post-processing step to enhance the surface finish. Moreover, the parts produced by PBF techniques are usually strong and fully dense, making them suitable for various demanding applications.

There are currently limited numbers of large-format PBF printers with build volume exceeding 1 m³ in the market. In 2010, EOS launched EOS P 800, which offers a build size of 700 × 380 × 580 mm [34]. GE Additive came up with the in 2015, which has a build size of 800 × 400 × 500 mm [35]. In 2019, Trumpf introduced the TruPrint 5000, providing a build size of 300 × 400 mm [36]. Recently, in 2020, SLM Solutions released the NXG XII 600 with a build size of 600 × 600 × 600 mm [37]. Nonetheless, the Binhu’s SLS machine, developed in 2014, was still one of the largest SLS printer, boasting a build area of 1.4 m x 1.4 m. It's worth noting that the throughput for these printers highly depends on the specific parameters of the print job, such as the part geometry, layer thickness, laser speed, and material.

However, scaling up PBF, particularly SLS, poses significant challenges that warrant further research. Achieving uniformity in the laser beam is a critical challenge. The incident angle of the laser beam has been found to impact the energy density, leading to variations in the shape of the laser spot [38]. Consequently, it affects the geometry of the melt pool, where wider melt pools and varying depths are observed based on the scan direction relative to the elongated direction. The study showed that an elongated laser beam at larger incident angle (18°) resulted in an increased interaction cross-section area by up to 6.7%, causing a decrease in energy density of up to 6%. This decrease in energy density can lead to lack of fusion defects and increased surface roughness. Therefore, there is a practical limit to the angle at which the laser can be incident. As the print bed size increases, maintaining the desired incidence angle would require placing the laser source further away from the bed. However, this approach becomes impractical due to the need for additional vertical space and more precise control of Galvano mirrors.

An alternative approach involves implementing a multi-laser system, with each laser covering a specific region of the print bed (Figure 2) [39]. While this can enhance printing speed, ensuring uniform energy intensity across all lasers becomes crucial to maintain consistent printing. Any variations in energy intensity among lasers can introduce inconsistencies in the printed parts. Moreover, multi-laser powder bed fusion (ML-PBF) faces additional challenges, such as deteriorating densification with an increasing number of laser beams, the formation of directional pores along overlap lines due to laser switching, resulting in decreased microhardness in PBF processed parts [39]. Further research on laser scan path management and thermal loads is required to address these challenges and advance the capabilities of PBF technology.

Figure 2. Diagram illustrating: (a) the individual scanning area of a single laser, and the overlapping areas for dual- and quad-laser in ML-PBF; additionally, alternating x/y raster scan strategy, (c) single-, dual- and quadruple-L-PBF locations on the platform. (d) Density comparison among various scan strategies [39] (reproduced with permission).

Large-scale SLS faces challenges in terms of material usage, particularly regarding the amount of powder required and the limitations of powder reusability [40,41]. In SLS, the entire layer needs to be filled with powder, even if only a small region within that layer requires sintering. This leads to a high amount of material usage and results in a significant portion of unused powder after each build. This inefficiency in material utilisation poses challenges in terms of cost and sustainability. The excessive amount of unused powder contributes to material waste and increases production costs. This is because the polymer would undergo thermal degradation which limits how many times the powder can be reused in SLS. Repeated usage of powder can lead to changes in its physical properties, such as particle size distribution, shape, and surface characteristics. These changes can negatively impact the sintering process and the quality of the printed parts.

Another challenge associated with scaling up SLS is that the energy management becomes complex with the need to pre-heat and maintain the temperature of a larger powder bed significant amount of unused powder during the printing process. This unused powder needs to be heated and maintained at the required temperature throughout the printing process, which demands additional energy. Apart from that, as the build volume gets larger, more inert gas would be required to prevent the materials from oxidation. Special attention may be required when working with large amount of inert gases. Managing the energy consumption and temperature control of a larger powder bed becomes more complex, requiring sophisticated heating and temperature regulation systems. The Fraunhofer Institute for Laser Technology ILT has developed an innovative concept for a PBF machine, as part of the Fraunhofer lighthouse project futureAM [42]. This machine is designed to produce large components, with dimensions measuring 1000 × 800 × 350 mm. The concept incorporates a movable processing head that utilises linear axes for its movement. The processing head is equipped with a localised inert gas system, which allows for precise control over the process conditions. In its latest stage of development, the processing head now features a five-scanner system to ensure high productivity during the additive manufacturing process. Rolls Royce Ltd. provided an aircraft engine oil transfer coupling as a demonstrator to assess the concept's viability. This component was specially redesigned for additive manufacturing. Utilising the new machine concept, the demonstrator was successfully produced with a maximum size of Ø620 × H280 mm, replacing a complex assembly of 44 individual parts.

Despite these challenges, PBF remains one of the leading techniques in the field of additive manufacturing due to its ability to produce high-quality parts from a variety of materials. Continuous research and technological advancements are being pursued to improve the scalability and efficiency of PBF.

2.3. Material extrusion

Material extrusion is a widely used 3D printing technique, also known as fused deposition modelling (FDM ^TM) or fused filament fabrication (FFF). It involves the process of creating three-dimensional objects by extruding and depositing successive layers of molten material, typically thermoplastics, through a nozzle. In material extrusion 3D printing, a solid filament of material is fed into the printer, which then passes through a heated nozzle [43]. The nozzle melts the filament, allowing it to flow and be deposited onto a build platform or previously printed layers. The extruded material quickly cools and solidifies, creating a strong bond between the layers. This layer-by-layer approach enables the construction of complex geometries and intricate designs.

One of the advantages of material extrusion is its versatility. A wide range of thermoplastic materials can be used, offering different properties such as strength, flexibility, and heat resistance. This flexibility makes material extrusion suitable for various applications, including prototyping, product development, and even end-use part production. Moreover, material extrusion 3D printers are relatively affordable and widely available, making them accessible to individuals, small businesses, and large-scale manufacturers. The simplicity of the process and the ability to use different materials make it a popular choice for many applications.

2.3.1. Extrusion technology: FFF vs fused granules fabrication (FGF)

The FFF technique is widely recognised and mature, finding applications in various industries due to its relatively low capital and material costs. In recent years, there have been developments in FFF 3D printers specifically designed for large-scale printing. One notable example is the WorkCenter 500 jointly developed by Oak Ridge National Laboratory (ORNL) and 3D Platform (3DP), which boasts a substantial build volume of 1400 × 2800 × 700 mm [44]. This allows for the production of parts on a larger scale.

However, a limitation of the conventional FFF process lies in the speed of material deposition. Factors such as the head speed, acceleration, and extrusion diameter, typically ranging from 0.1 to 0.5 mm, constrain the deposition rates. As a result, the rate of material deposition is typically limited to less than 0.5 kg/h on printers with small build volumes, typically less than 1 m³. ORNL’s researchers have developed a filament-based extruder with high-throughput, using 6-mm filament and is able to achieve an output rate of 3 kg/hr. To overcome issues of early filament melting and bridging, they employed a larger infeed, longer barrel, and air-cooling system [44]. However, higher printing speeds exposed an over-torque problem that requires further investigation.

To overcome this limitation and achieve higher throughput, the use of pellet-based extruders has become more common in large-format material extrusion techniques. By utilising pelleted materials, the deposition process can be accelerated, allowing for faster production rates in large-scale printing. This addresses the challenges posed by the speed limitations of the conventional FFF process, making it more suitable for efficiently producing parts on a larger scale.

FGF has emerged as a promising solution in the field of large-format 3D printing, offering a range of advantages over filament-based extrusion. One notable advantage is the high deposition rate they enable, which can significantly reduce production times by up to 200 times compared to traditional filament-based methods.

Various research endeavours have focused on harnessing plastic pellets for 3D printing purposes. Venkataraman et al. devised a 3D printer equipped with a screw extruder, enabling the utilisation of polypropylene (PP), polyethylene (PE), and polystyrene (PS) materials [45]. Volpato et al. developed a plunger-type 3D printer specifically designed for polymer pellets, and conducted a comprehensive analysis of the extrusion system's aspects such as material degradation, fuse continuity, product scale, and surface precision [46].

Table 1 shows some of the commercially available extruders with their respective specifications on material throughput.

Table 1. List of commercially available extruders for large-format polymer AM.

	Model	Weight (kg)	Max temperature (°C)	Available nozzle sizes (mm)	Max flow (kg/h)
Dyze	Pulsar	7	500	1.0, 3.0, 5.0	3
Rev3rd	RD-M10	18	500	2.5–10.0	12
	RD-M25	40	500	3.0–12.0	25
	RD-M40	90	500	2.0–20.0	50
Massive Dimension	MDPE 2	8.4	450	0.4–4.0	0.9
	MDPE 10	14	450	1.5	5.7
	MDPE 50	59	400	–	22.7
Extrudinaire		1.2	500	0.2–4.2	0.3
Weber	AE 16	–	450	–	2
	AE 20	–	450	–	4.5
	AE 30D	–	450	–	20
CEAD	E25	<29	450	2–18	12
	E40	<70	450	4–24	40
	E50	165	450	8–20	84

It can be observed that a wide variety of extruders have been developed to cater for different production scales. The extruders are typically single-screw extruders, mainly due to size and weight consideration as the extruder needs to be agile for printing. Although twin-screw extruders offer several benefits such as better compounding and mixing, especially useful for in-situ mixing of fibre-reinforced polymers, there are several drawbacks that restricted its application in 3D printing. Researchers from ORNL noted that twin-screws are typically designed to process powdered thermoplastic material instead of pellets, which could increase operating costs and complicate the conveying process [47]. Additionally, the use of a twin-screw extruder necessitates the addition of a melt pump to precisely control extrusion output, adding weight, cost, and complexity to the system. Mounting a twin-screw extruder horizontally, as required, poses a challenge as it may require sacrificing significant X/Y printing area in some AM systems. Lastly, concerns arose regarding the vibration and dynamic conditions of printing, which could potentially cause misalignment and damage to the tightly matched screws of a twin-screw extruder.

The flow rate of the single screw extruders is in the range of 1–80 kg/h, highlighting the advantage of pellet-based printing technique. Most of the extruders have a maximum temperature of at least 450°C, and are capable of processing the high-performance materials such as polyetheretherketone (PEEK). This increased deposition rates and high nozzle temperature make them particularly well-suited for large-scale printing projects and deal with wide range of thermoplastics, where efficiency and productivity are paramount.

Another key benefit of pellet-based extruders is the cost reduction they offer. Unlike filament-extruding processes, pellet-based extruders eliminate the need for filament production. Successful attempts of printing shredded recycled thermoplastics (PLA, ABS, PP and PET) of a wide range of sizes (< 22 mm²) have been reported with no significant drop in the mechanical properties [48]. As a result, the costs of raw materials can be reduced by more than tenfold compared to filament-based FFF printing. This cost-effectiveness opens up new possibilities for exploring a wide range of materials, including the recycled ones, as most of the industrial polymers are available in pellet form. Additionally, the versatility of pellet-based extrusion allows for the creation of composite materials by incorporating elements such as fibres and metal particles, further expanding the potential applications and material properties.

The pellet-based printing offers the advantage of material mixing during the printing process, enabling the creation of functionally graded materials (FGM). FGM allows for the incorporation of a less expensive material with sub-optimal mechanical properties for most of the part, while employing a higher-performance material in specific regions. This approach aims to optimise cost, weight, and mechanical performance. Sudbury et al. conducted a study on the transition between materials within the extruder, using Thermogravimetric Analysis (TGA) to determine that it takes approximately 32 s or 2.43 m of material for an extruder rotating at 100 RPM to initiate material extrusion, and an additional 1.52 m of deposited material to achieve a steady state condition for the second material [49]. Brackett et al. introduced a dual-hopper system designed for multimaterial printing, which utilises a rocker mechanism to facilitate smooth transitions between materials (Figure 3) [50,51]. This mechanism effectively closes off one hopper's feeding tub while simultaneously opening and positioning the other hopper. The researchers observed that despite the system's ability to consistently produce uniform material compositions during the transition in the initial layers, the transition location between CF/Acrylonitrile butadiene styrene (ABS) and Neat ABS exhibited a notable disparity between Layers 1 and 20. Additionally, they highlighted that the transitions were irreversible, suggesting that the direction of a transition should be considered as a contributing factor in predicting the behaviour of material pairs during transitions. More research is required to understand the mixing of the pellets to enable an accurate control of the composition during the transition phase. Table 2 provides a list of materials that have been used in FGF LFAM.

Figure 3. (a) The dual-hopper setup on the big area additive manufacturing (BAAM) system. (b) An illustration showing a bead consisting of four layers, where transitions are restricted to a single direction per layer; grey represents Material A and white symbolises Material B. This also demonstrates the variation in material composition based on the distance covered in the initial two layers. (c) Representation of the sample positions for assessing the consistency of the process, with the central number in each rod indicating the distance from the hopper switch [51] (reproduced with permission).

Table 2. List of materials for FGF LFAM.

Materials	Suppliers	Refs.
CF/ ABS	Techmer Engineered Solutions	[50,51]
PLA, PP, ABS, and PET	Nature Works LLC, McDonnough Plastics, Northwest Polymers and CiorC	[48]
PEEK 90G and PEEK 450G	Victrex®	[52]
TPU (Shore hardness 25A-65A)	Kraiburg TPE (Waldkraiburg, Germany)	[53]

Despite the advantages, the pellet 3D printing technique has encountered several drawbacks, including unstable material transfer resulting in phenomena like ‘bridging’, susceptibility to air entrapment during the plasticising process, and uncontrolled oozing at the nozzle leading to uneven product quality. For instance, when printing with shredded recycled plastics, Woern et al. noticed that a decrease in the mass flow rate associated with increased printing speeds, and the extruder experienced interruptions and stall at lower temperature, causing failures in the printing process [48]. To address these limitations, Liu et al. implemented a double-stage-screw extrusion mechanism for the pellet extrusion system [54]. This design optimises power distribution, with a larger first-stage screw responsible for melting, conveying, and applying pressure to the materials, while a smaller second-stage screw aids in further plasticising and melt measurement. Real-time pressure control is achieved by monitoring and adjusting the rotating speed of the first-stage screw using a pressure sensor, and a vent hole is incorporated to release trapped gas and prevent bubble entrapment. The overall printing process involves a melt generation unit, a 3D model information unit, and a process control unit, where plastic pellets are transformed into melt, subjected to pressure, and extruded through the nozzle for printing, facilitated by the rotation of the screws driven by dedicated motors. Nonetheless, the horizontal placement of the first extruder means that a significant amount of print area is sacrificed to accommodate the extruder.

The issue of inability to retract the material, which can result in oozing or stringing of plastic during the printing process is solved by incorporating the extruder with the retraction capability. For example, Extrudinaire has designed a mechanism that allows the entire screw to be lifted quickly, enabling retraction of the material [55]. When needed for subsequent printing, the screw can be lowered rapidly. This retraction mechanism helps prevent oozing and ensures more precise and controlled printing.

Overall, pellet-based extruders offer higher deposition rates, cost reduction, material versatility, improved processability of soft thermoplastics, and the opportunity to utilise plastic waste as a raw material. These advantages make them an attractive option for enhancing throughput, enabling the use of diverse materials, and promoting sustainable practices in the field of 3D printing.

2.3.2. Hardware level: gantry vs robot arm

Traditional gantry-based 3D printers are typically limited to three degrees of freedom, which restricts their manufacturing flexibility. One common requirement in gantry-based printing is the need for support structures when printing objects with overhanging features. These supports must be manually removed after printing, adding extra post-processing steps. Moreover, not all materials have compatible soluble support materials, further complicating the support removal process.

Robot arm 3D printing, on the other hand, offers the potential to overcome these limitations. Robots have greater degrees of freedom, allowing for more intricate and complex movements during the printing process. For example, the helical pipe with a small pitch has a shallow slope which makes it an overhanging feature and hard to print without using sacrificial materials in the conventional gantry-based printer (Figure 4(a)). This increased freedom of motion enables the robot to navigate around overhangs and print without the need for additional support structures [56,57]. As a result, the need for post-processing steps to remove supports is eliminated, saving time and effort.

Figure 4. (a) Sequence of images (a to h) showing how robotic extruder can overcome the need of support structure to print overhanging features, which normally requires support structures 3-axes gantry-based system [57], (b) Flexibility of robotic extruder to lay fibre along the stress lines [58], (c) Robotic printer on a mobile base [59] (reproduced with permissions).

In addition to increased flexibility and reduced reliance on supports, robot arm 3D printing offers several other advantages. One advantage is the ability to print on non-planar surfaces. Robots can adapt to curved or irregular geometries, enabling the printing of complex shapes that would be challenging or impossible with gantry-based printers (Figure 4(b)). Kubalak et al. introduced an algorithm that automates the generation of multi-degree-of-freedom (DoF) tool paths for applying a reinforcing skin onto AM parts [60]. The efficacy of this skinning approach was demonstrated by comparing skinned tensile bars to similar bars without the skin. The results showed that a single layer of skin led to a 9% increase in tensile modulus and a significant 59% increase in yield strength compared to the unskinned bars. This highlights the positive impact of the skinning strategy, taking advantage of the multiple degrees of freedom of the robotic printing, on enhancing the mechanical properties of the parts.

Taking inspiration from the intricate geometry of spider silk and with the objective of enhancing the structural performance of architectural structures, a team of researchers from Tongji University in China utilised a printing system consisting and extruder and a 6-axis KUKA robotic arm [61]. They devised a novel approach that deviates from the traditional layer-by-layer deposition method by combining primary curve structures with auxiliary curves that make contact with the main curve. To facilitate the simultaneous printing of the primary and auxiliary curves, the researchers designed one main nozzle and three secondary nozzles specifically for printing ABS material. This strategy enables the creation of complex architectural elements with improved structural capabilities.

It is possible to print larger-scale products with theoretically infinite build volumes by fusing a robotic printer with a mobile base [62,63]. A large-scale part printed by a mobile-base mounted robot arm is shown in Figure 4(d). However, because of the mobile base's presence, using mobile bases for AM may present difficulties in terms of accuracy and repeatability. To address this drawback, additional sensors for registration, a precise motion tracking system, dynamic feedback planning, and control should be implemented to overcome this drawback. It is possible to account for vibrations brought on by the mobile base by using feedback control loops to modify the location and orientation of the end-effector [64]. Additionally, improvements in mobile platforms’ positional accuracy and manoeuvrability have increased the potential of mobile manipulators for additive manufacturing in the future. Orbital Composite has introduced an innovative concept of a large-scale continuous fibre additive manufacturing involving containerised 3D printing robots [65]. These robots are positioned on a linear rail system to facilitate the construction of wind blades exceeding 100 metres in length [66].

Overall, robot arm 3D printing offers increased manufacturing flexibility, the potential to eliminate support structures, and the ability to print on non-planar surfaces. These advantages make robot arm 3D printing an attractive option for various applications, including complex and customised manufacturing processes.

2.3.3. Print orientation: conventional vertical printing vs lateral printing

Vertical layer-by-layer printing, the traditional approach in large-format material extrusion printing, has certain limitations. Firstly, it is constrained by the build height of the printer, restricting the size of objects that can be produced. This often necessitates dividing larger objects into smaller parts and assembling them later, introducing complexities and potential weak points.

One advantage of material extrusion is that instead of stacking the layers vertically up, it can be modified to print in a lateral fashion, where the layers are vertically oriented or any other angles (Figure 5). Lateral printing, also known as continuous or infinite build printing, offers several advantages over vertical layer-by-layer printing. The most significant advantage is the potential for infinite build length through a conveyor system. By enabling continuous printing along the length of the build platform, lateral printing eliminates the need to divide large objects and simplifies the manufacturing process. This reduces the risk of weak joints or seams and allows for the production of objects with theoretically unlimited length.

Figure 5. (a) Angle layer print and (b) vertical layer print to overcome limitation on the vertical height, allowing long objects to be fabricated [69] (reproduced with permission).

Two notable examples of lateral printing technologies are the Stratasys Infinite-Build system and Thermwood's vertical layer printing. The Stratasys Infinite-Build system features a conveyor belt-style build platform and a pellet-based robot arm 3D printer that facilitates continuous lateral printing [67]. This technology enables the production of large-scale, high-quality parts with exceptional build length. On the other hand, Thermwood Corporation has introduced vertical and angle layer printing, a technique that enables the fabrication of large-scale objects in a continuous, lateral manner (Figure 5) [68]. The LSAM 1540 printer incorporates a unique approach to lateral printing by incorporating a second moving table perpendicular to the main fixed horizontal table. This design, implemented by Thermwood, allows for the printing of large-scale objects with an extended length. Thermwood emphasises that their ‘controlled cooling’ print technology effectively reduces sag, a common issue encountered in thermoplastic 3D printing. With the implementation of the ‘Vertical Layer Print’ configuration, the LSAM 1540 printer has the capability to produce parts with a maximum length of 12 m. This approach eliminates the limitations of build height, providing greater design freedom and the potential for infinite build length.

2.3.4. Innovative additive manufacturing processes

Largix, an Israeli company, has developed an innovative 3D printing process called ‘APS’ (Additive Production System) for creating large-scale objects with simple geometry (Figure 6(a)) [70]. The APS process combines elements of Fused Filament Fabrication (FFF) and Welding Additive Manufacturing (WAM).

Figure 6. (a) Innovative multi material feeding mechanism to improve throughput [70] (Adapted from [70]) (b) hangprinter that consists of printhead hoisted with cables [71] (reproduced with permission).

In this process, multiple materials are fed into the extruder, and a laser is used to selectively melt the outer ‘skin’ of each filament strand. The regional melting minimises the internal stress generated due to the thermal shrinkage, resulting in lower tendency of warpage. This partial melting allows the filament to attach to previous and adjacent layers, ensuring strong layer adhesion and promoting excellent z-axis strength. The resulting structure exhibits z-axis strength exceeding 90%, indicating its exceptional integrity along that axis.

The key advantage of APS is its ability to produce large objects with enhanced strength by incorporating multiple filaments. By using the laser to melt and fuse the filaments, a solid and interconnected structure is formed. This approach provides a significant advantage for industries requiring custom large parts with simple geometries, such as storage tanks. These objects can be produced cost-effectively with APS, resulting in a potential 50% reduction in production costs.

Despite its novelty, APS has certain limitations. The process primarily involves melting the outer layer of the filament, resulting in stiff and less flexible filaments. This rigidity imposes constraints on the turning radius of the printhead, thereby limiting its ability to print parts with intricate features requiring sharp turns.

Furthermore, the feedstock filament used in this technique has an unconventional square cross-section, which introduces additional challenges. The production of such filaments requires an extra step compared to the pellet-based printing technique, increasing both cost and energy requirements. This additional step involves shaping the feedstock material into the desired square cross-section before it can be utilised in the printing process.

Hybrid manufacturing systems integrate different manufacturing processes, such as additive and subtractive methods, into a single process or separate yet interconnected environments [72]. Combining these processes offers several advantages, including improved surface quality, reduced tool wear, and shorter production times. Furthermore, hybrid processes enable the production of parts that may not be economically feasible using standalone processes [73]. While there are limited examples of multi-axis robotic applications in this field, most of them involve the combination of additive manufacturing and multi-axis CNC machining techniques [74].

Keating and Oxman devised a system where a KUKA robotic arm was used to alternately move a construction platform between an extrusion-based printer and a milling setup [75]. They first 3D printed large parts using polyurethane foam, followed by a milling and sanding process to improve the surface finish of the component. In a related development, researchers at the University of Illinois (U.S.A.) developed a platform that combined an extrusion printer and a milling system using a 6-DOF robotic arm [74]. The objective of this system was to eliminate the necessity for support structures, thereby saving on manufacturing time and minimising material wastage. Here, instead of moving the construction platform, they swapped the manufacturing tool.

In addressing the scalability challenges of traditional gantry-type 3D printers, Swedish innovator Torbjørn Ludvigsen introduced the Hangprinter (Figure 6(b)) [76]. More innovations have been carried out in this technology since the inception [77,78]. This groundbreaking design deviates from the conventional frame-based model; instead, it employs tensioned lines anchored to external structures. This not only allows for a vastly superior print volume relative to its footprint but also offers unparalleled adaptability to different settings. The Hangprinter's signature scalability, a limitation in standard printers, opens the door to expansive possibilities. Its ability to create oversized pieces positions it as a game-changer for large-scale endeavours, from art installations to architectural prototypes, heralding a transformative era in large-format 3D printing.

3. Applications

The advent of large-format polymer 3D printing has sparked transformative innovations across a spectrum of industries, redefining traditional manufacturing approaches. This section delves into compelling examples where this cutting-edge technology has disrupted conventional norms, from aerospace and marine engineering to automotive and beyond. These real-world applications underscore the versatility, cost-efficiency, and sustainability that large-format polymer 3D printing brings to industries striving for enhanced efficiency and groundbreaking solutions.

3.1. Aerospace and space

The aerospace industry has one of the longest supply chains of any industry. Having spare parts available when needed causes companies to store large quantities of parts in warehouses at great cost. The speed, flexibility, and efficiency of 3D printing for the aerospace industry enables manufacturers to produce replacement parts on demand. This is significantly faster and cheaper than ordering through standard supply channels, with the added benefit of 3D printing custom parts. This helps to optimise inventory levels and eliminates the need for storage facility maintenance. Furthermore, engineers can take hard to obtain or obsolete components and redesign them to be 3D printed, leading to time, cost, and labour savings. For instance, CNE Engineering printed moulds for engine exhaust covers for Scandinavian Airlines (SAS) during the COVID pandemic [79]. With the grounding of passenger planes and supply chain disruptions, it was impossible for SAS to order more equipment. Using large-format 3D printing, CNE was able to meet SAS's timing requirements as tooling was printed in a few days, and castings required only hours. In another work, by redesigning the restrain cradles for the helicopters of the Royal Navy and using 3D printing process to fabricate, SFM Technology was able to increase the production yield by 4 times as compared to the traditional manufacturing methods [80].

Another area of production where large-format 3D printing has proven particularly beneficial is the manufacture of inexpensive rapid tooling, jigs, and fixtures. For instance, the Chair of Carbon Composites at the Technical University of Munich employed LFAM to produce a mould for a composite flaperon, reducing production time to less than 8 h [81]. Ingersoll Machine Tools printed a 22-foot-long tool used for producing helicopter rotor blades (Figure 7(a)) [82]. Using its 3D printer, the MasterPrint, the tool was made with 521 kg of ABS with 20% carbon fibre fill in 75 h.

Figure 7. (a) 3D printed tool, machined and vacuumed sealed, for fabrication of helicopter’s rotor blade [82] (b) 3D printed concept car's monocoque [83] (reproduced with permission).

LFAM also shows promise in space applications. For instance, attempts have been made to 3D print rocket fuel grains in hybrid rockets, which combine aspects of solid and liquid rocket technologies, offering advantages like simplicity, safety, and throttle-ability. This involves the fabrication of thermoplastic fuel grains using Material Extrusion (ME) techniques [84]. The process allows for the creation of complex internal geometries in the fuel grains, which can lead to improved performance in hybrid rockets. This approach offers potential advantages in terms of customisation and efficiency in the manufacturing process, providing a significant innovation in the field of rocketry.

LFAM in space is a pivotal technology for future space missions. Its advantages include the ability to construct large-scale structures directly in space, reducing the reliance on Earth-based resources and significantly cutting transportation costs. This technology enables the creation of habitats and infrastructure using local materials, like lunar regolith or Martian soil, which is crucial for long-term space colonisation [85]. However, the challenges are substantial: 3D printers must operate in harsh space conditions, including extreme temperatures, high radiation levels, and microgravity, and there is a need for materials that can withstand these environments [85]. For instance, a study has focused on enhancing the thermal properties of 3D printed polymer composites for space applications by studying the effects of composite geometry and print direction on their thermal anisotropy, as well as assessing these composites’ impact on print quality and mechanical properties [86]. Furthermore, precision and reliability in remote operations are critical. Successfully overcoming these challenges could lead to a new era in space exploration, where building off-planet becomes a practical reality.

3.2. Automotive

AM provides the ability to reduce weight or volume with the freedom of a more optimal design. By applying topology optimisation and working with lattice structures, part weight and cost can be reduced. The automotive industry is constantly trying to reduce the overall weight of its components and design. 3D printers enhance this effort by using lightweight materials such as engineering plastics and composites. This significantly reduces vehicle weight, improves fuel efficiency and reduces emissions, while maintaining an edge over competitors that may not be able to match these performance improvements.

In the pursuit of sustainable manufacturing, students from Eindhoven University of Technology and Delft University of Technology harnessed large-format polymer 3D printing for groundbreaking automotive innovations. Eindhoven's student team fabricated a concept car's monocoque and body panels using Mitsubishi Chemical Europe’s Carbon-P, a recycled PETG with carbon fibres, achieving circularity by producing virtually zero waste and enabling the reuse of materials (Figure 7(b)) [83]. Meanwhile, Delft's students 3D printed a lightweight bodywork mould for a hydrogen-fuelled city car, minimising manual post-processing and securing victory at the Eco-Marathon 2022 [87].

Further illustrating the versatility of large-format polymer 3D printing, AREVO partnered with Kimoa to craft custom-sized bicycles featuring 3D printed Nylon/CF frames [88]. By printing these frames in a single pass, the need for part assembly was eliminated, resulting in stronger and more efficient bicycles.

In the realm of electric mobility, the Advanced Propulsion Centre of Canada (APMA) embraced large-scale 3D printing to fabricate an EV chassis under Project Arrow [89]. This innovation bypassed the requirement for costly moulds and fixtures, reducing both production costs and the environmental impact associated with traditional manufacturing processes.

The Nera electric motorcycle, a creation of BigRep’s Nowlab, stands as a testament to the prowess of large-scale 3D printing in the automotive sector [90]. Fully 3D printed except for its electrical components, the Nera motorcycle was crafted using BigRep’s FFF printers, showcasing the feasibility of producing intricate components like tyres, forks, and frames in 15 printed pieces. This not only expedited manufacturing but also improved the motorcycle’s fuel efficiency owing to its lightweight design.

These projects collectively underline the potential of large-format polymer 3D printing to reshape the automotive industry through circularity, reduced waste, customisability, and lightweight design.

3.3. Maritime

Within the maritime industry, the expenditure of around 13 billion USD annually on spare parts has spurred a search for more efficient solutions AM has emerged as a game-changer, particularly in low-volume production, offering accelerated lead times and the production of near-net-shape components. Large-format polymer 3D printing has significantly impacted by maritime sector by addressing spare part availability. For instance, Shell partnered with Poly Products to reverse engineer and 3D print a mooring buoy seal cover [91]. This effort drastically reduced lead time from 16 weeks to two, accompanied by a noteworthy 90% cost reduction. By converting scan data into a 3D printable file and utilising a CEAD Prime printer with fibre-reinforced Polyethylene terephthalate glycol (PETG), the half-metre long seal cover plates were effectively produced.

Meanwhile, JAMADE made strides with a 3D printed underwater scooter, with 75% of its components additively manufactured [92]. Leveraging three BigRep ONE large-format printers and the Pro HT filament, which has good resistant to warping and exceptional softening temperature of 115°C, this approach yielded remarkable cost-efficiency, accuracy, and quality. The utilisation of large-format printing also enhanced the scooter’s water-resistance, mitigating potential leak concerns that could arise from assembling smaller parts.

The University of Maine achieved a remarkable Guinness World Record through the 3D printing of a boat measuring 7.62 m in length and weighing 2268 kg (Figure 8(a)) [93]. Named 3Dirigo, the boat was fabricated within a mere three days, employing a large-scale 3D printer from Ingersoll Machine Tools. The usage of a plastic and wood cellulose blend showcased the feasibility of scaling up additive manufacturing for substantial applications. Simultaneously, Al Seer Marine unveiled the groundbreaking HYDRA, the world's inaugural 3D printed drone boat [94]. Fabricated through a 36-metre-long composite-based pellet extrusion system from CEAD, this innovation showcased a drone boat spanning five metres in length and weighing 345 kg, achieved in a span of five days. In tandem, Caracol employed robotic 3D printing to craft a sailboat hull from recycled polymer. Realised in a single piece through Caracol's large-scale 3D printing system, this accomplishment exemplified the integration of recycled polymers and circular economy principles to forge advanced, high-performance components [95].

Figure 8. (a) 3D printed boat printed in University of Maine, Advanced Structures and Composites Center [93]. (b) Interior cabin parts being printed for the prototype. (c,d) External view of the completed PLA 3D 850 prototype [96] (reproduced with permission).

In the naval industry, large-format 3D printing is emerging as a transformative tool, offering innovative solutions for design and manufacturing challenges. For instance, a notable application can be seen in the production of cabin prototypes. Instead of the traditional approach of using glass fibres or steel sheets paired with rock wool insulation, researchers leveraged the potentials of 3D printing using flame retardant ABS Toyolac© and Ingeo Polylactic acid (PLA) 3D850 materials(Figure 8(b)) [96]. By adopting this technique, they could ensure enhanced fire resistance, water resistance, and UV protection. Impressively, the weight of the traditionally constructed cabin, which stood at 450 kg, was significantly reduced to 250 kg for the ABS prototype and 290 kg for the PLA one. Additionally, this method streamlined the manufacturing process, with both prototypes being printed, post-processed, and assembled efficiently. This example underscores the promise large-format 3D printing holds for the naval industry, emphasising weight reduction, structural integrity, and efficient material use.

3.4. Renewable energy

AM has the potential to improve wind turbine performance and bring benefits to the wind energy sector, which is growing at a record pace. In 2020, the U.S. has seen a 24% year-on-year increase in its offshore wind pipeline and the EU had 17.4GW of new wind installations, held back by permit bottlenecks and global supply chain issues. Currently, the moulds and equipment for blade production can cost more than $10 million, and it can take 16–20 months before they are market-ready. The University of Maine will tackle this issue by 3D printing moulds using bio-based feedstock [97]. Utilising advanced 3D printing techniques with bio-derived materials can decrease blade development expenses by 25% to 50% and speed up the process by a minimum of 6 months. Moulds made from these substances can be recycled by grinding and repurposing them for other moulds, enhancing their sustainability. Nevertheless, the material's mechanical characteristics remain comparable to that of aluminium.

Orbital Composites on the other hand, are developing on-site, high-throughput manufacturing of wind blades with large-scale continuous fibre additive manufacturing using its containerised 3D printing robots for wind blade manufacture [66]. This reduces supply chain problems as shipping containers are the most versatile and lowest-cost method for global relocation. Mould transportation costs can be greatly reduced if 3D printing feedstock-loaded shipping containers are being transported instead of mould surfaces.

These examples illustrate the diverse applications of large-format polymer 3D printing across various industries, from aerospace and automotive to maritime, architecture, furniture, and renewable energy. Each application showcases the technology's capacity for customisation, rapid prototyping, and sustainable production methods.

3.5. Building and construction

Leading the charge in innovative architectural applications, Nowlab introduced the BANYAN ECO WALL – a pioneering irrigated green wall, entirely 3D printed using BigRep’s large-scale FFF 3D printers [98]. This structure, inspired by plant systems, merges as both a plant support and a water supply system. In a breakthrough, advanced CAD software and AM facilitated the intricate design of this wall, showcasing the power of fully digitised processes to achieve functional complexity that traditional methods cannot replicate.

In the realm of public art, Branch Technology has made its mark with an exceptional project at the Durham Main County Library [99]. Utilising six-axis Kuka robots on a rail, the company employed carbon-filled ABS filament to craft a captivating 8 × 4 × 5 m artwork. This endeavour, made possible by the advances in 3D printing, underscores the technology's capability to actualise intricate geometries within a remarkably short timeframe, producing a stunning structural masterpiece formed by interconnected space frame cells.

Fusing innovation with sustainability, DSM and Royal HaskoningDHV have embarked on a significant venture – a 3D printed circular composite bridge poised for a Rotterdam park [100]. Crafted from Arnite, a blend of PET and PBT materials, this circular composite bridge embodies a move towards circular and sustainable bridge designs, minimising wear and tear while offering a novel solution for urban infrastructure.

These endeavours collectively illuminate the transformative potential of large-format polymer 3D printing in architecture and infrastructure, from functional green walls and intricate public art to sustainable circular bridges.

3.6. Furniture and interior design

AM enables product designs and dimensions that are difficult to create through traditional manufacturing, thus bypassing existing design and manufacturing constraints. In traditional fabrication, some topology-optimised designs cannot be fabricated due to their complex shapes and designs. There is also a growing demand for bespoke and personalised products. Consumers are looking for products that reflect their unique styles and preferences, and manufacturers can use AM to provide customisation and personalisation options.

For instance, Caracol undertook the design and 3D printing of furnishings for a historic Italian tower dating back to 1161 [101]. This project incorporated numerous 3D printed furniture and decor pieces, carefully contrasting the mediaeval charm of the Capitolare Tower. These interior and exterior components were meticulously designed, engineered, and produced using Caracol's robotic material extrusion additive manufacturing technology, employing eco-friendly 3D printing materials. Leveraging the flexibility of the 6+ axes movements, the pellet-based extruder attached to the robotic arms can manufacture parts with very complex geometries, including non-planar tool paths, and unconventional slicing at different angles such as 45° or 60°.

In a parallel vein of innovative design, Delft University of Technology engineered a lounge chair capable of seamless adaptation to user preferences, allowing users to switch between sitting and reclining positions by leaning against the chair's back structure [102]. This mechanism, activated by the user's weight, supports an average-sized human and was realised through robotic material extrusion 3D printing by 3D Robot Printing BV in Rotterdam. This shape change is realised by combining variation in material distribution and the use of thermoplastic elastomers. Multi-colour 3D printed chairs have also been fabricated using scrap and waste plastics, highlighting the advantage of 3D printing that contributes to the sustainability of the environment [103].

Concurrently, Aectual demonstrated sustainability-focused ingenuity by 3D printing interior furnishings from recycled Tetra Pak drink carton materials [104]. Through the utilisation of pellet extrusion technology executed by a robotic arm, discarded cartons were transformed into fully recyclable architectural elements. This innovative approach not only curbs waste and fosters circular processes but also addresses the cost and environmental challenges typically associated with custom architectural designs.

These endeavours collectively underline the transformative capabilities of large-format polymer 3D printing in interior architecture, spanning from historic preservation and adaptable furniture design to sustainable material use and circular design philosophies.

4. Challenges and potential

This section evaluates the challenges and potential aspects across five key areas: material selection, innovative manufacturing processes, enhancement of interlayer bonding, the compromise between surface quality and production speed, and quality assurance with process control. The objective is to provide an encompassing view of the current landscape and potential progression of large-format additive manufacturing.

4.1. Materials

Materials development plays a crucial role in addressing the challenges associated with large-format printing. One significant issue that arises when working with polymers is their tendency to shrink upon cooling. This phenomenon, known as thermal shrinkage, can have a significant impact on the dimensional accuracy and structural integrity of printed parts.

Pure polymers exhibit varying degrees of thermal shrinkage, with coefficients of thermal expansion (CTE) typically ranging from 25 to 200 µm/m°C [105]. For example, common polymers like ABS and PC have CTE values of around 60–70 µm/m°C and 70–110 µm/m°C, respectively [106]. These shrinkage tendencies become more pronounced when scaling up printing to larger formats due to the higher volume of material involved.

In addition to conventional polymers, there is a growing interest in using high-performance materials like PEEK for large-format printing. PEEK offers exceptional mechanical properties and high-temperature resistance, making it suitable for demanding applications. However, printing with PEEK presents more stringent requirements due to its high melting point and increased risk of thermal stresses during cooling, which could lead to warping [107]. Hu et al. noted that the warping rate could be as high as 20.3% when printing PEEK, leading to serious delamination [108].

To mitigate the shrinkage-related challenges, various approaches have been explored. One approach involves controlling the print environment, including factors such as temperature, humidity, and cooling rates. By carefully managing these parameters, it is possible to minimise the impact of thermal shrinkage and reduce warping in printed parts. Alsoufi and Elsayed investigated the effect of nozzle temperature and print speed on warping deformation and found that nozzle temperature of 220°C and print speed of 15 mm/s were found to be the optimum for printing PLA to achieve minimum warping [109]. Hu et al. introduced a heat collector near the nozzle to trap and reflect the heat radiated from the nozzle to preheat the layer right before depositing the filament [108]. The method has been shown to reduce the warping rate from 20.3% to 5% for PEEK printing. Another strategy involves the use of additives to modify the CTE of polymers. These additives are incorporated into the polymer matrix to restrict shrinkage and minimise the associated stresses [110]. The mechanism behind this is typically related to the inclusion of fillers or reinforcements, such as fibres or particles, which act as physical barriers to restrict polymer chain movement and reduce overall shrinkage. By carefully selecting and incorporating these additives, it is possible to tailor the CTE of the material and improve dimensional stability during the printing process.

Another advantage of adding additives, such as fibres, into polymers is that it offers opportunities to functionalise the printed parts beyond their mechanical properties. Additives can enhance the electrical, magnetic, or other specific properties of the materials, expanding their potential applications. Farahani et al. have comprehensively reviewed the advancement of the use of composite material in 3D printing to achieve multifunctionality [111]. However, the addition of these additives also introduces challenges related to the rheological properties and processability of the polymers.

When additives are introduced into polymers, they can affect the viscosity, flow behaviour, and melt strength of the material. This, in turn, can influence the processability and printability of the filament. In the case of fibre-reinforced polymers, the high aspect ratio and alignment of the fibres can significantly impact the flow characteristics of the material during extrusion. It may require adjustments in the extrusion parameters, such as temperature, speed, and nozzle design, to ensure proper filament formation and prevent nozzle clogging. For the case of vat-photopolymerisation techniques, sedimentation of the reinforcement could be an issue which would affect the homogeneity of the composite material [112].

One limitation in the use of high fibre loading for filament-based 3D printing techniques, such as FFF, is the occurrence of nozzle clogging. The highest fibre loading typically achievable without frequent clogging is around 40–50% [113]. To enable the successful printing of higher fibre loading filaments, further research and development are needed to improve the printability and overcome the challenges associated with nozzle clogging. In vat-photopolymerisation techniques, it has been observed that increasing the fibre content typically results in less effective dispersion of the fibres [114]. Furthermore, higher fibre loading can lead to the entanglement of fibres, which in turn diminishes the bonding between the fibres and the matrix material.

Several aspects can be focused on to enhance the printability of high-fibre loading filaments. One area of interest is to study the impact of fibre characteristics, such as length, diameter, and aspect ratio, on the rheological properties and printability of the composite filaments can provide insights into optimising the formulation of these materials [115]. Understanding the flow behaviour and how the fibres interact with the polymer matrix during the printing process can aid in designing filaments with improved extrusion behaviour and reduced nozzle clogging risks. Furthermore, investigating the effect of various additives, such as coupling agents or compatibilisers, can help tailor the rheological properties of the composite filaments and mitigate processing challenges. These additives can enhance the interfacial interactions, reduce agglomeration, and improve the dispersion of the fibres within the polymer matrix, leading to enhanced printability.

3D printing's flexibility enables selective deposition of different materials, thereby creating functional parts. Goh et al. highlighted the advantage of multi-material 3D printing in producing multifunctional parts, specifically in the fabrication of soft robotic grippers [116]. This expansion in multi-material 3D printing capability has sparked interest in a new research domain: the interfacial properties of heterogeneous polymers. Utilising FFF, Freund et al. explored the factors impacting the interface strength of multi-material components [117]. They conducted peel resistance tests on various combinations of rigid and flexible materials. Their findings underscored the critical role of material selection, particularly polarity and mechanical interlocking, in determining the interface strength of FFF-produced parts. In a similar manner, Hasanov et al. investigated different material transition strategies like mechanical interlocking and functionally graded designs to enhance interfacial properties [118]. Their research demonstrated that a gradient-based design could yield approximately double the strength of an interlocking design, indicating its superior effectiveness. However, it is worth noting that most research to date has been conducted on a small scale. To determine the applicability of these findings to larger scale production, further investigation is warranted.

4.2. Interlayer bonding

One of the fundamental issues with 3D printing is the unequal mechanical properties resulting from the layer-by-layer printing technique. In conventional 3D printing, the material is deposited layer upon layer in the x-y plane, aligned with the build plate, and the following layers are introduced in the z-direction. As a result, the bonding between layers is typically weaker compared to the continuous material within the x-y plane. This leads to significantly higher strength in the x- and y-directions than along the z-axis.

Several studies have demonstrated this directional dependence of mechanical strength in small-scale FFF. For example, tensile tests conducted on ABS and polylactic acid (PLA) samples printed in various orientations revealed a strength reduction of up to 78% in the z-axis compared to the x-axis [119]. Similarly, tests on carbon fibre-reinforced ABS samples displayed an even more pronounced orientation dependence of around 8 times [120]. A study has also shown that for continuous carbon fibre-reinforced nylon, the along-the-fibre strength can be 150 times more than that of the across-the-layer [121].

In powder bed fusion and vat photopolymerisation processes, the mechanical properties of printed parts would also vary with their orientation, but to a smaller extent. In a review of relevant research, Liaw and colleagues observed that for parts fabricated using Selective Laser Sintering (SLS), the tensile strength along the vertical axis (z-axis) is approximately 10% less than that along the horizontal axis (x-axis) [122]. In contrast, parts produced through Stereolithography (SLA) displayed a negligible difference in strength, with only about a 1% decrease in the z-axis compared to the x-axis. Nonetheless, there are also reports showing the anisotropic properties of DLP-printed parts. Steyrer et al. [123] and Keβler et al. [124] conducted research showing that the stiffness of DLP-printed materials, as gauged by Dynamic Mechanical Analysis (DMA) and tensile testing, differs based on the print direction. Aznarte et al. found that samples printed in the vertical Z direction exhibited lower tensile modulus and ultimate tensile strength compared to those oriented in the tensile testing direction [125]. This variation is attributed to the directional characteristics imparted by the printing process and the tensile testing alignment.

Enhancing the bond between layers during the printing process is crucial for reducing anisotropy. The thermal history of the interface between the newly added material and the layer beneath it greatly influences the strength of the bond. Maintaining the interfacial temperature above the glass transition temperature (T_g) of the material is essential for promoting molecular chain interfusion. Studies have indicated that substantial molecular chain motion occurs between the T_g and a critical temperature associated with neck formation, suggesting the importance of maintaining temperatures above T_g during printing [126,127]. For instance, in a study about improving interlaminar fracture toughness, it was found that raising nozzle and bed temperatures enhanced the mode I interlaminar fracture toughness [128]. Another study showed that printing at a lower speed (20 mm/s as compared to 80 mm/s) improved the interlayer bonding (18 MPa vs 6 MPa), which could be due to slower cooling as a result of the slower printhead movement, leading to a longer time for molecular chain interfusion [129].

To address this issue, researchers have explored various approaches to mitigate the anisotropy of printed components. One approach involved exposing printed parts to ionising radiation, which initiated a crosslinking reaction between layers of PLA, resulting in a 20–50% reduction in anisotropy [130]. A different research employed targeted microwave heating on polymer composites with carbon nanotubes, resulting in a remarkable 275% improvement in bond strength [131].

Researchers have explored various methods to elevate the bonding temperature at the interface. Attempts have been made to actively heat the interface using forced air, mixed results were obtained. Prajapati et al. observed a 19% increase in tensile strength and 145% increase in tensile toughness as a result of using heated jet [132] but Partain noted that the disturbance caused by the heated jet compromised the deposition geometry and did not yield improved strength [133]. Laser-based systems have shown promise by increasing the bonding interface and resulting in a 195% increase in tensile strength [134] and 50% increase in inter-layer bond strength [135].

Infrared (IR) preheating has been proven effective in improving the interlayer mechanical properties of 3D printed structures, especially for larger components with longer layer times [136,137]. Regardless of the initial substrate temperature, IR preheating consistently raised the pre-deposition temperature to the target value of 150°C and significantly enhanced the mechanical performance of the printed parts. At lower substrate temperatures, IR preheating resulted in over a 500% increase in fracture energy and an 80% improvement in tensile strength. However, at the highest IR preheating settings (TH = 220°C), there was a slight negative impact on bond strength, likely attributed to thermal degradation of the polymer under intense infrared energy. Further research is necessary to optimise the thermal energy transfer of infrared radiation in reinforced polymers commonly used in large-scale additive manufacturing.

ORNL researchers have developed a tamper, a rapidly reciprocating platen positioned around the extruder nozzle, to enhance the compressive force and improve the Z-strength of printed parts (Figure 9(b)) [47]. The tamper aims to maximise material properties by promoting better layer-to-layer bonding. Intriguing data was obtained from the tamper experiments, revealing a remarkable 52% increase in strength compared to untampered prints. However, the data showed an unexpected outcome when the tamper was positioned below the nozzle to provide additional compression. In this case, there was a loss of strength compared to using a tamper flush with the nozzle. The leading hypothesis suggests that excessive cooling caused by the tamper may hinder effective layer bonding. Further investigation is required to fully comprehend and optimise the tamper with a platen positioned below the nozzle.

Figure 9. Z-tamping mechanism and IR heating studies conducted by ORNL researchers. (a) Print head with IR lamp and z-tamping mechanism [136], (b) detailed description of the z-tamping mechanism [47] (c) chart showing improvement in tensile strength due to IR heating (reproduced with permissions).

By addressing the anisotropy issue in 3D printed components, these research efforts contribute to the advancement of additive manufacturing technologies, enabling the production of more structurally robust and reliable parts.

4.3. Process control and quality assurance

Process control and quality assurance play a critical role in large-format 3D printing due to the significant amounts of materials and energy involved in the process. It is essential to ensure that the printing proceeds as expected and that the printed parts are free from defects.

In material extrusion, maintaining a consistent shape of the material being deposited can be challenging. This is especially true during transient operations such as starting, stopping, and navigating corners. The delayed response of the material flow during these operations can lead to issues like inadequate material deposition at the beginning and end, resulting in underextrusion and oozing respectively. These problems contribute to the formation of seam defects, which are more noticeable when using larger nozzle sizes.

To address this problem, researchers at ORNL have developed a device called the ‘Posiverter’ that diverts the material flow and eliminates the delayed response, thereby improving the quality of the seams (Figure 10) [138]. Additionally, the delay in material flow also affects the consistency of the material at corners. When the extruder slows down at a corner, the material flow does not decelerate as quickly due to the delay, leading to a bulge at the corner. As the extruder accelerates away from the corner, the material flow fails to catch up promptly, resulting in a narrower width than expected.

Figure 10. (a) Design of the posiverter and (b) the flow profiles before and after the posiverter and the difference between (c) the extruded filament with and (d) without feedforward control [138] (reproduced with permission).

To enhance the consistency of material flow at corners, a feedforward control mechanism has been implemented. This control mechanism anticipates the variations in material flow and compensates for them, thereby reducing the irregularities in the shape of the material deposition at corners and improving the overall print quality.

In a study of temperature profile of the large-scale material extrusion technique, Borish et al. noted additional waiting time is required for the deposited material to cool down sufficiently before the next layer is deposited, especially on layers with short toolpaths [139]. This was due to the slower cooling rates (∼ 0.5°C/s) because of the larger mass of the extruded filament. The wait time allowed the material to have enough time to solidify so that the extrudate can support the weight of the subsequent layers. This highlights the importance of process control to achieve good-quality prints.

In vat-photopolymerisation, shrinkage of 3D printed parts is known to deform the geometry and adversely affects function. Through a series of experiments, it was found that several process parameters could effectively minimise shrinkage in the printing process [140]. The investigation revealed that increasing the laser exposure time, raising the concentration of the photoinitiator during the post-print UV-curing step, and extending the duration of UV exposure were all effective measures for reducing shrinkage. Additionally, it was observed that shrinkage became more pronounced as the layer spacing and structure height increased. However, the application of UV-curing post-processing techniques proved to be effective in mitigating these effects. Remarkably, when employing post-processing, the areal shrinkage rate could be decreased to as low as 1%.

One way to know whether the printed parts are defect free is to use monitoring systems to inspect and track the printing process. However, manually inspecting the information layer by layer can be a laborious and time-consuming task. Additionally, establishing a relationship between the sensor signals and potential defects can be challenging.

In order to address these issues, it is necessary to develop automated methods for analysing the sensor data and identifying any anomalies or defects in the printed parts. Machine learning techniques are normally being applied to automatically detect potential defects during the printing process [141–143]. This approach represents a significant step towards achieving in-process closed-loop feedback, which is crucial for effective quality assurance. By leveraging machine learning algorithms, the monitoring system can analyse sensor data in real-time and identify deviations or anomalies that may indicate the presence of defects [144]. This automated defect detection not only reduces the reliance on manual inspection but also enables proactive measures to address issues during the printing process.

Implementing robust process control and quality assurance measures in large-format 3D printing is paramount. By continuously monitoring the process and leveraging machine learning techniques, manufacturers can ensure that the printed parts meet the desired specifications and quality standards. This approach not only minimises material waste and energy consumption but also enhances productivity and reduces the likelihood of post-processing or rework. Ultimately, it contributes to the overall efficiency and reliability of large-format 3D printing, making it a viable and cost-effective manufacturing solution for various industries.

4.4. Tradeoff between production speed and surface finishing

In material extrusion, the production speed is a crucial factor determining the throughput of the printing process. The maximum volumetric speed, which governs the throughput, can be calculated as the product of print speed, layer height, and extrusion width. Increasing the maximum volumetric speed leads to higher throughput. To improve throughput, one can increase the print speed, layer height, and extrusion width. However, there are limitations to how high the print speed of the nozzle can achieve, often due to constraints in the moving mechanism. Furthermore, higher print speeds generally result in poorer surface finishing, leading to rougher surfaces on printed objects.

The layer height and extrusion width are closely related to the nozzle size. Using a larger nozzle size can increase the throughput, as it allows for more material to be deposited in a shorter period. However, this approach significantly reduces surface finishing quality. When using a larger nozzle, the layer thickness increases, resulting in a more obvious ‘stair-stepping’ effect on the printed object's surface.

To overcome this trade-off between production speed and surface finishing, researchers have developed innovative nozzle designs (Figure 11) [138]. These designs are capable of transforming from a large nozzle size to a smaller one when necessary, particularly when printing angled surfaces. By using a smaller nozzle for finer details and critical surfaces, these designs can improve surface finishing without significantly sacrificing overall throughput.

Figure 11. Variable size nozzle to improve surface quality without significantly sacrificing print time. (a) Mechanism of variable size nozzle (Adapted from [138]) and (b) the difference in surface finish using different nozzle size [138] (reproduced with permission).

Batt et al. have developed an innovative approach for printing large composite parts using a robotic printer equipped with two 6 DOF robotic manipulators [145]. This setup enables multi-resolution printing, where one manipulator utilises a large diameter nozzle to rapidly print the interior regions of the part, while the second manipulator employs a small diameter nozzle to ensure high-quality surface finish for the exterior regions. This multi-resolution additive manufacturing technique significantly reduces the build time of large parts without compromising surface quality. Moreover, the flexibility of the 6 DOF manipulators allows for printing on non-planar surfaces, expanding the possibilities of deposition paths beyond horizontal planes. This approach holds promising potential for further time savings in the fabrication of larger parts.

The speed of the material extrusion technique can be improved by multi-nozzle system. For instance, Mhatre et al. developed a print process strategy for multi-nozzle printhead to increase the deposition rate [146]. This printhead, however, has an additional rotational axis which requires a specialised 4-axis multi-nozzle toolpath generator.

Bacciaglia et al. introduce a novel toolpath planning approach for gantry-based systems with multiple independent extrusion heads [147]. By implementing a static priority strategy, where one extruder takes precedence over the others, the method significantly reduces printing time while maintaining part quality. The provided an efficient solution for planning extrusion head trajectories in collaborative material extrusion architectures, ultimately enhancing the printing efficiency by a factor equal to the number of extrusion heads.

In attempt to speed up the deposition rate in robotic 3D printing, Shen et al. introduce a novel large-scale 3D printing system utilising collaborative robots [148]. They introduced an optimisation algorithm for printer task scheduling based on efficiency egalitarianism and devised a strategy for robot interference avoidance by partitioning printing layers into secure and interference zones. Collaborative printing experiments conducted on the multi-robot platform validate the efficacy of this approach, showcasing a remarkable efficiency improvement of over 73% compared to conventional printing methods. Nonetheless, they highlighted that to make this collaborative printing approach universally applicable, it is imperative to develop a more comprehensive segmentation algorithm capable of handling intricate shapes in printing models.

Poudel et al. introduced a generative method capable of autonomously generating diverse printing schedules for segmented objects by exploring an extensive solution space, surpassing the limits of human cognitive capacity [149]. They showcased the effectiveness of the generative approach through two practical examples: one involving a large, straightforward rectangular bar and another featuring a smaller, intricately designed folding sport utility vehicle (SUV). This research highlights that the generative approach can create numerous distinct print schedules for collision-free Composite 3D Printing (C3DP), a task that extends beyond the capabilities of human heuristics alone. Nonetheless, as this approach predominantly operates as an exploratory algorithm for navigating the complete design space, a significant limitation lies in the fact that the generated schedules are confirmed as valid but do not guarantee optimality. Subsequent efforts should entail the development of an optimisation layer to augment this approach, with the aim of pinpointing the optimal collision-free print schedule while minimising printing time.

These efforts to optimise the trade-off between production speed and surface finishing are crucial for achieving desired results in material extrusion. Balancing the need for high throughput with the requirement for smooth and visually appealing surface finishes is essential in various applications, ranging from rapid prototyping to functional part production. By employing advanced nozzle designs and optimising print parameters, it is possible to strike a balance between production speed and surface finishing in material extrusion, enabling efficient and high-quality 3D printing processes.

4.5. Complexity in toolpath generation for higher degree of freedom 3D printing

Despite the fact that more robot arm systems have been used in 3D printing due to its multi-axes printing that enhanced manufacturing flexibility, there are limitations related to the robot arm systems which requires more research and development to overcome. Specifically, achieving sharp corners in 3D printing presents a challenge due to the need for abrupt changes in the nozzle's velocity, which results in infinite acceleration. However, the majority of robotic systems have inherent constraints on their maximum acceleration, which are determined by the type of actuators they employ. While performing the printing operation, the robot may approach a configuration called a ‘singular configuration’, characterised by exceptionally high joint oscillations that can significantly degrade the quality of the printed layer. To address this issue, Apis Cor introduced an algorithm grounded in the pseudo-inverse of the Jacobian matrix [150]. This algorithm ensures the generation of smooth trajectories, especially when the robot approaches singular configurations.

Another approach to large-scale 3D printing involves the use of articulated robot systems, which consist of robotic arms. These arms offer the advantage of requiring less space than gantry systems and can be mounted on transportable platforms, facilitating on-site construction [151]. However, it's important to note that articulated robot systems typically have more limited workspaces compared to gantry robots. This limitation arises from the substantial moments generated at the base when the robot reaches its maximum extension. One such example is a ‘cylindrical robot’, featuring a first joint for vertical translation, a second joint with revolute motion, and a third joint with telescopic translational movement. This design, while space-efficient, poses challenges in creating sharp corners, often resulting in circular paths [150]. In this configuration, only three DoF are employed. However, for more intricate geometric shapes, additional DoFs become necessary. These extra DoFs serve to alter the orientation of the nozzle, especially when executing complex 2D movements, such as those required for sharp corners. Importantly, since 3D printing occurs layer by layer, there is generally no need for more than four DoFs, where the fourth DoF is utilised for rotating the printing head around the vertical axis.

The additional DoF requires specialised slicer to generate toolpath for the robot arm. Although there is already commercially available software for robot arm 3D printing, they are still at the early stage of adoption and there is a lot of room for improvement. For instance, Lim et al. used Grasshopper – a plugin of Rhinoceros® to create Curved Layered FDM toolpath in a single scripting environment [152]. Nonetheless, the generated toolpath will need to be checked for potential collision using a post-processor for the robotic 3D printer, which then involved multiple software. Future software should ensure interoperability of the data and integration of the necessary functions for robotic 3D printing. This ensures that different teams can work seamlessly with the same digital representation of the design. In the context of 3D printing, where all processes are digitised, achieving interoperability among architectural design, structural analysis, and the printing process is essential. Also, to streamline and optimise the building process with minimal manual intervention, it's crucial to translate the digital model and confirm its compatibility with the 3D printing process. Ideally, this verification should occur automatically. Additionally, the creation of digital models for 3D printing is a time-intensive task.

4.6. Recyclability of materials

Most AM parts are not in a ready-to-use state, but rather require additional machining to smooth surfaces, which results in a large amount of waste. This waste plastic can contain carbon fibres or other filler used for stabilising the print, which can make it difficult to recycle. Recycling composite materials is a complex process fraught with challenges, yet it also harbours considerable potential. The fundamental challenge in recycling these materials lies in the separation process; the distinct components must be isolated into pure forms suitable for effective recycling.

The recycling process for composites necessitates specialised techniques. Conventional recycling methods are not fully equipped to manage the intricacies of composites. For example, thermoplastics can be readily melted and reshaped, but the high temperatures required can degrade both the thermoplastic and the reinforcing fibres. Consequently, advanced methods such as chemical recycling including acid digestion, which breaks down the composite to its chemical base, or mechanical recycling, involving shredding and remelting, must be further developed [153]. Chemical recycling is important for photopolymer resins, which are typically thermosets. This is because once cured, they cannot be melted down for reuse like thermoplastics. This characteristic makes recycling photosensitive resins more challenging. Chemical recycling breaks down the cured resin into its chemical constituents using various chemical treatments [154]. These constituents can potentially be reused to synthesise new resin, although this process is complex and requires further development.

A significant concern in recycling composites is the quality of the recycled output. Often, the recycled material is not as robust as the original due to the degradation of properties. For instance, it was found that reprocessing the thermoplastic would cause a drop in molar mass, which in turned resulted in a drop in tensile strength [155]. Therefore, ongoing research into enhancing the quality of recycled composites through better additives and processing techniques is crucial.

The economic aspect of recycling composite materials is yet another challenge. The entire process, including collection, transportation, and processing, can prove costly, and the market for recycled composites is still evolving [156]. Technological advancements that reduce recycling costs and create higher-value recycled materials could make the process more economically feasible.

Despite the challenges, recycling composite materials presents an excellent environmental opportunity. Minimising waste from prevalent industries like 3D printing can significantly reduce environmental impacts. Innovative recycling methods could transform waste into valuable resources, thereby promoting a more circular economy where materials are continuously reused instead of discarded.

Research and development play a pivotal role in the future of composite material recycling. There are ongoing studies into new recycling methods, including ‘design for recyclability’, which focuses on creating composites that are easier to recycle from the outset [157–159]. Another area of interest is bio-based composites, which may be simpler to decompose and recycle [160].

Finally, the establishment of industry standards and legislation can stimulate the recycling of composite materials. Regulatory policies mandating responsible end-of-life product management can incentivise manufacturers to consider the recyclability of their products more seriously.

In summary, while the recycling of composite materials presents significant challenges, the potential benefits are equally substantial. Through technological innovation, economic incentives, and regulatory frameworks, the recycling of composites can be transformed into a viable, sustainable practice, contributing to environmental conservation and resource efficiency.

4.7. Integration of automated fibre placement with material extrusion AM

Automated Fibre Placement (AFP) is renowned for its ability to construct large-scale fibre-reinforced polymer composites. Its capability to lay down fibre-polymer tapes in a predetermined path allows for the creation of components with a high strength-to-weight ratio, which is particularly advantageous in the aerospace industry for manufacturing parts like airplane fuselage and wing skins [161]. However, the AFP process is generally not suited for creating parts with intricate details due to its lower resolution compared to other AM techniques [162,163].

On the other hand, FFF is an AM process that excels in precision and detail, and it can create parts with complex geometries and fine features. By merging AFP and FFF, the composite manufacturing industry can overcome the trade-off between scale and detail. The proposed hybrid approach would utilise AFP to create the core structure of a large part, providing the necessary mechanical strength and structural integrity. Once the core is established, FFF can be employed to add detailed features and complex geometries to the surface.

The integration process would require a synchronised manufacturing platform where the AFP lays down the broad, structural layers while the FFF head adds details simultaneously or in a secondary process. This not only saves time but also enables the production of parts that are both robust and intricately detailed, which is not feasible with either process alone. The FFF process could also be used to form the mould structure for the AFP process, significantly shortens the lead time of fabricating large composite parts.

The synergy between AFP and FFF could open up new possibilities in AM, such as the production of large moulds with fine surface textures or lightweight structural parts with integrated small-scale functional elements. This combined method holds the potential to revolutionise the manufacturing of large, complex parts across industries, including aerospace, automotive, and beyond.

To effectively implement this hybrid method, it is essential to develop a comprehensive understanding of the material properties and process parameters of both AFP and FFF. Research and development efforts should focus on optimising the transition between the two methods to ensure a seamless interface and mechanical bonding between the large-scale and detailed features. Additionally, software tools need to be developed to manage the complex path planning required to execute this integrated approach effectively.

In summary, the combination of AFP and FFF could provide a revolutionary step forward in the field of additive manufacturing, enabling the creation of large parts with the desired complexity and detail that meet the industry's demanding standards.

4.8. Design considerations

In the realm of large-format additive manufacturing (LFAM), design considerations encompass a range of challenges and opportunities that demand a meticulous scientific approach to optimise the technology's potential. One of the primary challenges lies in adapting design methodologies to fully exploit LFAM capabilities, particularly the integration of complex geometries and multifunctional components that traditional manufacturing methods cannot achieve. This necessitates a profound understanding of the material properties unique to LFAM, including the behaviour of thermoplastic composites under varying printing conditions, such as temperature gradients and layer adhesion, which significantly impact the mechanical integrity and performance of the final product. Moreover, the scalability of designs to large formats introduces issues related to structural stability, precision, and the mitigation of thermal distortion, requiring innovative computational models and simulation tools for accurate prediction and control. Roschli et al. detailed the design considerations for big area additive manufacturing, highlighting that in addition to the design considerations for small-scale AM, new rules and considerations are also required for LFAM due to the change in the thermo-mechanical behaviours of materials at larger scale [164].

The advancement in technology has also led to innovation in printing strategies such as angled printing and nonplanar printing for LFAM, which has broadened the print path planning approaches but also brought along their own set of considerations which designers need to keep in mind. Alireza Badiee from Savonia University of Applied Science published two articles offering practical insights and methodologies for overcoming specific challenges associated with angled printing [165,166]. Through detailed experimentation and analysis, the author provided valuable guidelines that not only address the complexities of LFAM but also pave the way for more efficient and effective use of this transformative manufacturing technology. Future research directions are likely to focus on the machine learning or optimisation algorithms for efficient and reliable design recommendation and path planning for LFAM.

4.9. Post-processing

Post-processing in 3D printing is crucial because it enhances the mechanical properties, surface finish, and dimensional accuracy of printed parts, particularly in large-format printing. Kumar and Velmurugan reviewed various surface treatment and surface modification techniques for 3D printed materials in general, touching on processes such as machining, blasting, polishing, texturing, oxidation, coating and deposition [167]. However, post-processing in LFAM presents unique challenges: managing microstructural heterogeneity that affects mechanical properties, and navigating complex surface finishing and metrology for large and intricate parts. These challenges require innovative solutions to ensure the quality and efficiency of large-format 3D printed components.

4.9.1. Managing microstructural heterogeneity that affects mechanical properties

In polymer 3D printing, managing microstructural heterogeneity is vital to ensure consistent mechanical properties across the printed object. This heterogeneity can arise due to variations in the printing process, such as temperature fluctuations, print speed, and layer adhesion, leading to inconsistencies in the material's density and strength. These variations can significantly impact the final product's durability, flexibility, and stress resistance. Addressing these challenges requires precise control over the printing parameters and often post-printing treatments to homogenise the microstructure, ensuring that the printed parts meet the required mechanical specifications.

Scaling up 3D printing processes complicates post-processing due to the increased volume and surface area of the printed parts. Uniformity in post-processing becomes a formidable challenge, as any inconsistency can lead to compromised structural integrity and aesthetic appeal.

Heat treatment is vital in enhancing the mechanical properties of polymer 3D printed parts. For instance, Kishore et al. explored isothermal annealing as a post-process technique to enhance the mechanical properties of carbon fibre-reinforced polyphenylene sulphide (PPS) components printed using the Big Area Additive Manufacturing (BAAM) system [168]. The annealing process, conducted at 250°C for 18 h, significantly improved the storage modulus of all PPS grades (both neat and reinforced) above their glass transition temperature (T_g). The treatment resulted in increased crystallinity due to secondary crystallisation, and though oxidative structural changes occurred mainly on the surface, they did not notably impact the degree of crystallinity within different regions of the sample. This indicates that thermal treatment through isothermal annealing effectively enhances the mechanical performance of PPS components for tooling applications, particularly in autoclave environments. Uniform heat application on large parts is challenging, requiring innovative furnace designs or localised heating methods [137]. The development of new techniques for effective heat distribution is a growing research area, with potential solutions including segmented heating and infrared treatments.

4.9.2. Navigating complex surface finishing and metrology for large and intricate parts

In large-format 3D printing, navigating complex surface finishing and metrology for intricate parts is challenging. This involves ensuring a high-quality surface finish and accurate dimensional measurements on large, often complex geometries. The process becomes more intricate as the size and complexity of the parts increase, requiring advanced techniques for surface smoothing, polishing, and precision measurement. These steps are essential to meet specific functional and aesthetic requirements, especially for applications in industries like aerospace, automotive, and medical, where precision and reliability are paramount. One promising solution is the integration of robotic machining in post-processing. Robots can provide precise and efficient surface finishing, ensuring uniformity over large areas. However, the complexity lies in programming the robots to adapt to varied geometries and in maintaining the delicate balance between efficiency and precision. The complexity of these tasks underscores the need for specialised equipment and expertise in large-format 3D printing.

Common coordinate measuring systems for geometrical metrology in additive manufacturing include tactile coordinate measurement machines (CMM), optical methods, and X-ray computed tomography (XCT) [169]. While all these systems can measure external shapes and dimensions of moderately complex components, XCT stands out for its ability to measure internal and hard-to-access features in highly complex components. XCT has become one of the most widely used methods for geometrical measurement and quality inspection in additive manufacturing due to this unique capability.

Scaling up the manufacturing of large parts in additive manufacturing (AM) poses challenges in geometrical measurements due to the limitations of X-ray computed tomography (XCT). The complexity and size of the components make it difficult to effectively utilise X-rays, especially considering the material thickness limits observed by Moroni et al., where maximum penetration thicknesses were noted as 150 mm for aluminium and 40 mm for steel [170]. This implies that the dimensions and complexity of large AM components may be constrained when measured using XCT. Moreover, the requirement for an enclosed chamber in XCT restricts the maximum size of the component that can be scanned, with one of the largest XCT devices being Baker Hughes’ speed-scan, accommodating a maximum size of 600 mm in diameter and 900 mm in length, and a weight limit of 50 kg.

A summary of the challenges and potential in LFAM is shown in Table 3.

Table 3. Summary of the challenges and potential in LFAM.

Study focuses	Challenges	Existing approaches	Potential research directions
Material development	Polymer shrinkage leading to warping Sedimentation of fibres in vat-photopolymerisation [112,114] Nozzle clogging at high fibre loading in material extrusion technique [113]	Addressing shrinkage problem through print process optimisation [108,109] Addressing shrinkage problem through modification of CTE of polymers [110]	Improving the printability of composite materials with high fibre loading Functionalising LFAM parts through the use of composite materials Functionally graded multimaterial Interfacial study in LFAM
Interlayer bonding	Weak z-direction strength [119] Poorer z-direction strength in composites [120,121]	print process optimisation [129] microwave heating [131] laser-based heating [134] Infrared heating [136,137] Tampering device [47]	Tampering device with heating to overcome excessive cooling
Process Control and Quality Assurance	Inconsistency of extrudate shape during transient operation shrinkage of 3D printed parts in vat-photopolymerisation	‘Posiverter’ that diverts the material flow and eliminates the delayed response	Defects detection using machine learning techniques
Optimising production speed and surface finishing	Poor surface finish for LFAM	Adaptable nozzle design [138] Non-planar printing with 6 DOF robotic manipulators [145] multi-nozzle system [146] collaborative robots [148]	more advanced nozzle design toolpath planning for collaborative robots
Toolpath generation	motions of individual axes of robotic arms are coupled	Curved Layered FDM toolpath in a single scripting environment [152]	interoperability of the data and integration of the necessary functions for robotic 3D printing
Recyclability of materials	Composites materials are hard to be separated into their constituent materials Degradation of thermoplastics during recycling [155]	chemical recycling [153,154] development of bio-based composites that are simpler to decompose and recycle [160]	establishment of industry standards optimising the transition between the two methods to ensure a seamless interface software tools need to be developed to manage the complex path planning
Design Considerations	change in the thermo-mechanical behaviours of materials at larger scale multiple print strategies available	develop design guideline for each print strategy	machine learning or optimisation algorithms for efficient and reliable design recommendation and path planning for LFAM.
Post-processing	Managing microstructural heterogeneity that affects mechanical properties Navigating complex surface finishing and metrology for large and intricate parts	isothermal annealing[168] localised heating methods [137] tactile coordinate measurement machines (CMM), optical methods, and X-ray computed tomography (XCT) [169]	development of new techniques for effective heat distribution Development of suitable measurement method of large 3D printed parts

5. Concluding remarks

This review seeks to provide a comprehensive exploration of large-format additive manufacturing (LFAM) for polymers, with a focus on its scalability, challenges, and potential solutions. By assessing the scalability of various 3D printing techniques such as the vat photopolymerisation, powder bed fusion, and material extrusion techniques and delving into key challenges such as material selection, interlayer bonding, surface quality versus production speed, challenges in robotic arm 3D printing, quality assurance, and recyclability of material this review aims to pave the way for advancements in LFAM.

Key findings indicate that material development, including additives and fibre reinforcement, can significantly improve dimensional stability and multifunctionality while requiring careful consideration of rheological properties and nozzle clogging and sedimentation. Furthermore, optimising production speed versus surface finishing in material extrusion 3D printing can be achieved through innovative nozzle designs and parameter adjustments.

In light of these findings, the future trajectory of LFAM research and development should concentrate on multiple fronts. First and foremost, the creation of novel materials tailor-made for LFAM applications is essential. These materials should not only address thermal shrinkage but also unlock new dimensions of multifunctionality, enabling the fabrication of parts with enhanced properties. The advancement of recyclable materials in LFAM is equally essential as it contributes significantly to the sustainable manufacturing practices.

Simultaneously, advanced process control techniques and quality assurance methodologies demand our attention. These aspects are pivotal for ensuring the reliability, repeatability, and overall success of LFAM processes across a diverse range of industrial applications.

In conclusion, this review is a comprehensive compass for navigating the ermering field of LFAM for polymers. It provides valuable insights into the present state of the technology, highlighting its promises and challenges, while also charting a course towards a future where LFAM can revolutionise the way we manufacture complex and large-scale components across industries.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Data sharing is not applicable to this article as no new data were created or analysed in this study.