Introduction

Composite materials with nanoparticles or fiber additives are widely utilized across industries such as aerospace, healthcare, and consumer products. These additives not only enhance the mechanical properties but also provide exciting opportunities to introduce additional functionalities, which broaden the applicability and impact of composites.

Conventional composite manufacturing methods, from hand lay-up to filament winding, typically involve a trade-off between the degree of automation and design freedom. These processes require expensive molds to shape the resin and fiber architecture. High-volume production is required to offset tooling and equipment costs. In contrast, additive manufacturing (AM), or 3D printing, has emerged as a transformative approach for composite fabrication, offering unprecedented freedom in material composition and geometric complexity1. AM eliminates the need for molds and significantly reduces the cost per part even at low production volumes. As a result, it is ideally suited for rapid prototyping and product development in composite manufacturing.

Over the past decade, the field of composite 3D printing has rapidly advanced with expanded printable materials, including various mechanical and functional reinforcements, thermoplastics, and thermally or UV-curable resins2. New printing methods continue to emerge, which aim to enhance the quality, scalability, and versatility of composite 3D printing. By precisely controlling the shape of printed composites and the distribution of reinforcements, AM enables the customization of composite products previously unattainable using conventional manufacturing methods.

Realizing the full potential of AM for composite fabrication demands advancements across multiple disciplines. Among them, design optimization is increasingly being recognized as a critical component3. Optimization tools enable engineers to predict material behaviors and develop effective composition placement strategies. Recent advances in design methodologies, ranging from density-based4 to level-set topology optimization approaches5, have facilitated the fabrication of manufacturable prototypes6,7. Furthermore, design objectives in existing studies have expanded beyond maximizing stiffness to achieving diverse multi-physical and functional properties8,9,10,11.

This review highlights very recent advancements in design methodologies and AM techniques of composites, such as thermoset composites with continuous fibers. Unlike existing reviews12,13, this article simultaneously emphasizes the mechanical properties and contemporary functionalities of printed composites, as well as the development of design theories that support these applications. Such integration is essential yet notably absent in the current review literature, as it exploits the exceptional design freedom offered by AM to achieve optimal material distributions and targeted composite properties. Ultimately, this review aims to inform researchers about emerging trends, attract broader research interest, and inspire interdisciplinary collaborations to drive continued innovation in this rapidly evolving field.

Additive manufacturing methods for composites

According to ISO/ASTM 52900:2021, AM technology is divided into seven major techniques: vat photopolymerization (VPP), material extrusion (MEX), powder bed fusion (PBF), binder jetting (BJT), material jetting (MJT), directed energy deposition (DED), and sheet lamination (SHL). Among them, four techniques are commonly employed in composite 3D printing. Figure 1 provides a schematic overview of the principles underlying these methods.

Fig. 1: Schematic view of existing printing methods applicable to polymeric composites.
Fig. 1: Schematic view of existing printing methods applicable to polymeric composites.
Full size image

a Fused deposition modeling. b Direct ink writing. c Stereolithography. d Digital light processing. e Material jetting. f Powder bed fusion.

MEX techniques mainly include fused filament fabrication (FFF, Fig. 1a) and direct ink writing (DIW, Fig. 1b). FFF deposits thermoplastic filaments through a heated nozzle, while DIW uses viscoelastic inks. Reinforcing particles or fibers can be incorporated into both methods, either by embedding them within thermoplastic filaments before FFF printing or by mixing them with the printable ink for DIW.

VPP methods, such as stereolithography (SLA, Fig. 1c) and digital light processing (DLP, Fig. 1d), employ photopolymerization to fabricate parts from liquid resin. These techniques are highly valued for their exceptional resolution and smooth surface finish. SLA and DLP are often employed to produce functional composites containing particles or milled fibers.

MJT deposits droplets of photopolymer in precise patterns, followed by layer-by-layer UV curing (Fig. 1e). While directly incorporating fibers is challenging due to the small nozzle sizes, this method excels at fabricating heterogeneous material structures or nanoparticle-reinforced composites with intricate geometries and localized functional properties.

PBF methods, such as selective laser sintering (SLS, Fig. 1f), use a high-powered laser to fuse powdered materials into solid layers. This approach is compatible with a wide range of thermoplastic composites embedded with particle reinforcements.

Additive manufacturing of composites: enhanced mechanical properties

Nanoparticles, milled fiber, and short fiber reinforced composites

Various carbon, ceramic, and metal-based nanoparticles can be readily incorporated into AM processes due to their small sizes (1–100 nm). However, a notable challenge is achieving and maintaining uniform dispersion. Due to the strong interparticle attractions, these small fillers tend to aggregate, leading to inconsistent distribution within printable inks. Magnetic stirring, shear mixing, and ultrasonic sonication are increasingly being employed to promote their homogeneous dispersion before composite printing14. In addition, high filler loadings can significantly alter the ink’s rheological properties and impair printability. This issue is particularly critical for vat photopolymerization-based printing, where precise viscosity control is crucial for proper layer formation and curing. Existing studies reveal that the mechanical reinforcement provided by these nanoparticles is often moderate due to the low particle content to avoid agglomerates. Instead, their primary roles lie in introducing various electrical, magnetic, and thermal properties to 3D-printed composites15.

Milled fibers (50–200 µm in length) and chopped fibers (0.2–10 mm in length) are effective in enhancing the mechanical strength of composites. Most research in this area focuses on extrusion and vat polymerization-based printing methods. In FFF, fiber and thermoplastic pellets are processed using a twin-screw extruder to produce continuous filaments, which typically contain 10–40 wt% short fibers16. The inherent line-by-line manufacturing process of FFF results in mechanical properties that are notably lower than those achieved using conventional fabrication methods with the same fiber content (Fig. 2a). Printing with higher fiber content poses challenges such as increased void formation, weak filament-filament and fiber-matrix bonding17. Current research trends focus on optimizing printing parameters (e.g., nozzle temperature, print speed, and layer height) to enhance layer fusion and minimize voids18 and incorporating nanoparticles or sizing agents to improve fiber-matrix bonding19.

Fig. 2: Additive manufacturing of composites with milled fibers or short fibers.
Fig. 2: Additive manufacturing of composites with milled fibers or short fibers.
Full size image

a Tensile strength as a function of fiber volume ratio for parts manufactured using conventional and additive manufacturing techniques (Image reproduced with permission from ref. 16, Copyright John Wiley and Sons). b Mechanical vibration-integrated anti-clogging DIW printhead enabling composite printing with up to 45% short carbon fiber content (Image reproduced with permission from ref. 23, Copyright Elsevier). c 3D printing of short fiber composites using a frontal polymerization resin, which facilitates uniform matrix curing and supports free-standing composite printing without additional support materials (Image reproduced with permission from ref. 24, Copyright American Chemical Society). d DLP printing utilizing ultrasound to align short fibers in specific directions and patterns within each printed layer (Image reproduced with permission from ref. 27, Copyright Elsevier). e Electrically assisted DLP printing with graphene nanoplatelets aligned along the electric field direction (Image reproduced with permission from ref. 28, Copyright Elsevier).

Recently, a pellet-based 3D printing method has emerged for processing thermoplastics20,21. Instead of using pre-extruded filaments, this technique directly feeds thermoplastic pellets into a screw extruder, which mixes and extrudes the material along the deposition path. It also supports the use of pellets with short or milled fibers to print thermoplastic composites with enhanced mechanical performance and improved interfacial adhesion. A major advantage of this method is its capability to print multimaterial structures by dynamically adjusting the pellet mixing ratio during printing. Unlike conventional filament-based FFF, which produces sharp, discrete interfaces between different materials, pellet-based printing enables continuous modulation of material composition, allowing for smooth transitions and graded properties within and across printing layers.

DIW is a viable method for printing thermal or UV-curable composites with short fiber additives, but it requires careful consideration of the ink rheological requirements. High fiber content significantly reduces ink flowability and can lead to nozzle clogging. To address these issues, researchers are developing new shear-thinning inks with tunable rheological properties22, as well as advanced nozzle designs with anti-clogging features23 (Fig. 2b). In addition, to mitigate the issue of opaque fibers hindering light penetration and curing of UV-curable composites, innovative resin formulations, such as frontal polymerization24 (Fig. 2c) and dual-cure systems25, have been recently utilized to ensure homogenous matrix curing and enhance the mechanical properties of printed composites.

SLA and DLP techniques enable the fabrication of UV-curable fiber composites with intricate and high-resolution geometries. Traditionally, short fibers or other additives are uniformly dispersed within the resin and then subjected to printing. Due to the small layer thickness, the additives do not significantly block light penetration. Recently, there has been growing interest in manipulating the organization and alignment of fibers within the resin during printing26,27,28, which dramatically enhances composite performance and expands the design space. For instance, ultrasound generated by piezoelectric transducers was adopted to align carbon fibers along the pressure nodes (Fig. 2d)27. In another study, an electric field was applied during SLA to align polarized graphene nanoplatelets along the field direction28 (Fig. 2e). The composites produced in these studies demonstrated improved mechanical strength and electrical conductivity along the fiber alignment direction.

Continuous fiber reinforced composites

Among various additives, continuous fibers enable the highest mechanical strength of printed composites and extensive design possibilities. Extrusion-based printing methods are often the only viable approach for printing continuous fiber-reinforced polymer (CFRP) composites. For instance, FFF has been employed to print thermoplastic composites using in-nozzle impregnation29 or pre-impregnated filaments30. Recent studies have focused on optimizing printing conditions to enhance fiber/matrix bonding and interlayer adhesion, employing techniques such as microwave heating31 (Fig. 3a), hot-compaction32 (Fig. 3b), and laser-assisted heating33.

Fig. 3: Additive manufacturing of composites with continuous fibers.
Fig. 3: Additive manufacturing of composites with continuous fibers.
Full size image

a Microwave-assisted high-speed FFF printing for thermoplastic composites with continuous carbon fiber (Image reproduced with permission from ref. 31, Copyright Elsevier). b Hot-compaction roller technique for 3D printing of thermoplastic composites to reduce voids and enhance adhesion between printing layers (Image reproduced with permission from ref. 32, Copyright Elsevier). c A DIW printhead for thermally curable composites with continuous fiber (Image reproduced with permission from ref. 34, Copyright Elsevier). Embedded 3D printing of continuous fiber composites (Image reproduced with permission from ref. 36, Copyright Royal Society of Chemistry): d the schematic working mechanism, e composite printed with variable fiber volume fraction and matrix materials, f composite printed with free-standing filaments. 3D printing of continuous fiber composites with two-stage UV-curable resin (Image reproduced with permission from ref. 40, Copyright Royal Society of Chemistry): g the matrix curing mechanism, h shape reshaping, i recyclable 3D printing.

3D printing of thermosetting composites has recently been demonstrated using thermally curable or UV-curable resins. He et al.34 developed a novel DIW printhead, which relies on viscous ink flow within the deposition nozzle to apply shear stress to the fiber and drive its flow through the nozzle (Fig. 3c). In UV-curable CFRP printing, Continuous Composite Inc. (Coeur d’Alene, ID) patented a DIW printhead design that solidified the matrix resin under UV light immediately after the deposition of composite filaments35. Subsequent studies adopted similar mechanisms to enhance process control and application potential.

Extrusion-based CFRP printing faces challenges such as fiber spreading and void formation. Additionally, it lacks flexibility for adjusting fiber volume fraction or changing matrix materials during printing. Printing complex structures with hollow or overhanging features also remains challenging without support materials. Ding et al.36,37 recently addressed these limitations by developing an embedded 3D printing technique that utilizes a deposition nozzle to write continuous fibers below the resin (Fig. 3d). A laser beam is directed onto the resin surface and cures the resin around the fiber bundle. The printing method demonstrated its advantages in producing high-quality composite samples with well-aligned fibers, minimized void density, and outstanding mechanical properties. It also enabled dynamic control of fiber volume fractions, matrix material changes during printing (Fig. 3e), and the fabrication of filaments along overhanging pathways without support materials (Fig. 3f).

Conventional thermal and UV-curable resins have distinct advantages and limitations. Thermally curable resins provide high mechanical strength and strong bonding but face challenges with unsupported structures. In contrast, UV-curable resins enable high fiber content and rapid fabrication of complex shapes but exhibit lower mechanical strength and weak filament bonding. Opaque fibers also impede light penetration, causing inhomogeneous matrix curing. To overcome these challenges, recent studies have introduced innovative resins, including those using frontal polymerization38 and resins with conductive rheological modifiers39. Jiang et al.40 utilized a two-stage UV-curable resin. After the first stage UV polymerization, a second stage post-heating is applied to trigger dynamic covalent reactions and form an interpenetrating network (Fig. 3g). This process uniformly increases the matrix crosslinking density and composite stiffness by 11 times. It also enabled the repairability and reshaping ability of printed composites (Fig. 3h). Additionally, the printed composites were fully recyclable, with embedded fibers being reclaimed for subsequent printing (Fig. 3i).

Another promising research trend involves integrating robotic arms with composite 3D printing to enhance motion control and expand printing capabilities. Baur et al.41 investigated the mechanical properties of CFRPs produced using the robotic Continuous Fiber 3D Printing (CF3D®) process, which combines in situ resin impregnation, extrusion, roller compression, and UV curing (Fig. 4a). The resulting composites exhibited properties comparable to aerospace-grade epoxy composites. Rahman et al.42 developed robotic 3D printing processes for UV-curable composites with continuous glass fibers and systematically analyzed the effects of various parameters (Fig. 4b). Abdullah et al.43,44 integrated a six-axis robotic arm with CFRP 3D printing and established a digital workflow. This approach enabled the fabrication of large-scale composite structures on planar (Fig. 4c) and curved substrates (Fig. 4d), including substrates whose shape is determined during the printing process by laser-based 3D scanning (Fig. 4e).

Fig. 4: Robotic arm-driven 3D printing of composites.
Fig. 4: Robotic arm-driven 3D printing of composites.
Full size image

a 3D printing of UV-curable composites with continuous carbon fiber using a robotic arm. The process involves in situ resin impregnation of dry fiber followed by extrusion from a nozzle (Image reproduced with permission from ref. 41, Copyright Sage Publication). b 3D printing of UV-curable composites with continuous glass fiber using a robotic arm (Image reproduced with permission from ref. 42, Copyright Sage Publication). c Robotic DIW printing of CFRPs on large scale. d Composite printing on curved substrates and (e) substrates whose shape is determined during the printing process by laser-based 3D scanning (Image reproduced with permission from ref. 43, Copyright John Wiley and Sons).

While this review focuses primarily on 3D printed composites reinforced with synthetic fibers, there is growing interest in incorporating plant-derived natural fibers, driven by their renewability, biodegradability, low density, and significantly reduced environmental footprint. The sustainability of these composites can be further enhanced by printing natural fibers with bio-based polymer matrices, such as polylactide and poly(3-hydroxybutyrate), to create fully biodegradable material systems. Early research has primarily explored the use of short natural fibers, including flax, jute, ramie, and hemp, in processes such as FFF and DIW45,46. More recently, considerable progress has been made in adopting continuous natural fibers using in-situ impregnation and prepreg filament extrusion methods47. These fibers enable the design of lightweight, bio-based structures with improved performance and tunable mechanical behavior. Although factors such as mechanical strength and interfacial compatibility are still present for future optimization, natural fiber composites have already demonstrated strong potential in applications that value material sustainability and design flexibility, such as in automotive interiors, consumer products, packaging, and building components.

Additive manufacturing of composites: innovative functionalities

The embedded additives not only enhance the mechanical properties of printed composites but also offer promising opportunities to introduce additional functionalities. To date, these functionalities have primarily focused on the following three categories.

4D printing of shape-changing composites

Early studies in this field enabled shape-changing behaviors using conventional shape memory polymers (SMPs). Incorporating reinforcement fillers into SMPs enhanced their stiffness and actuation force and also introduced multi-stimuli responsiveness. For instance, conductive particles enabled Joule heating, while carbon black facilitated light-induced heating, both of which allowed the material to reach the necessary temperature for shape recovery.

Recently, liquid crystal elastomers (LCEs) have gained attention for their large, reversible actuation capabilities. Functional fillers have been incorporated into LCEs to enable electrothermal and photothermal heating. For example, Roach et al.48 demonstrated electrothermal actuation by embedding conductive silver wires into LCE-based soft grippers. Liu et al.49 incorporated carbon nanotube (CNT)-filled LCE filaments that exhibited reversible deformation triggered by infrared light or electric fields (Fig. 5a). Ambulo et al.50 introduced a DIW-printable LCE composite with liquid metal droplets for electrothermal and photothermal responses. Kotikian et al.51 employed coaxial DIW printing to create LCE shells around liquid metal cores, which electrically activated LCE composites for use as soft actuators (Fig. 5b).

Fig. 5: 4D printing of shape changing composites.
Fig. 5: 4D printing of shape changing composites.
Full size image

a Photothermal actuation of LCE filaments with carbon nanotubes (Image reproduced with permission from ref. 49, Copyright John Wiley and Sons). b Core–shell 3D printing of LCE filament with a liquid metal core and the reversible actuation upon Joule heating (scale bar = 5 mm) (Image reproduced with permission from ref. 51, Copyright John Wiley and Sons). c 4D printing of LCE composites with continuous fiber and their reversible shape changes upon heating (Image reproduced with permission from ref. 52, Copyright Springer Nature). d DIW printing of ferromagnetic particles reinforced composite, where particles are aligned during printing by a magnetic field around the nozzle, enabling complex shape changes in the composite structures. (Image reproduced with permission from ref. 53, Copyright Springer Nature). e Shear-induced alignment of cellulose fibrils during DIW. Bilayer flowers were demonstrated during the swelling process (scale bars = 5 mm, inset = 2.5 mm) (Image reproduced with permission from ref. 54, Copyright Springer Nature). f Folding deformation of composites due to mismatching thermal expansion coefficients between continuous fiber and polymer matrix (Image reproduced with permission from ref. 55, Copyright Elsevier). g Coaxial extrusion of hollow LCE fibers and their electrically driven reversible actuation when incorporating liquid metal (Image reproduced with permission from ref. 56, Copyright John Wiley and Sons).

Jiang et al.52 further advanced 4D printing by incorporating continuous fibers into LCEs (Fig. 5c). During DIW printing, the relative motion between the fiber and resin within the deposition nozzle generates shear forces that align the microscale mesogens. This alignment enables the monodomain state and actuation behavior in the matrix materials. The inclusion of continuous fibers dramatically enhances the stiffness, actuation force, and speed of the materials. It also expands the frontiers of 4D composites by allowing fibers to be strategically positioned at various locations and orientations to achieve diverse shape-changing patterns, such as twisting, curling, and folding. Additionally, incorporating conductive fibers enables electrically activated LCE composites for wearable electronics and soft robotic applications.

Beyond matrix materials, the shape-changing behaviors of printed composites can also be driven by the motion of embedded particles. As a notable example, Kim et al.53 demonstrated DIW printing of ferromagnetic particle-reinforced composite (Fig. 5d). During printing, the particles are aligned by an applied magnetic field around the dispensing nozzle. When magnetic fields are later applied to the printed composites, the particles tend to reorient, which induces shape changes in the composite material.

Hydrogel swelling provides another mechanism for shape changes. Gladman et al.54 printed hydrogel precursors with nanofibers using DIW (Fig. 5e). These nanofillers undergo shear-induced alignment within the nozzle. The printed filaments exhibit a smaller swelling ratio and higher elastic modulus in the longitudinal direction compared to the transverse direction. This principle enables programmable plant-inspired architecture, such as folding 4D-printed flower structures in water.

Different thermal expansion between fibers and the polymer matrix is another strategy for achieving shape changes. Wang et al.55 designed and fabricated bilayer composite structures with continuous fibers using the FFF process (Fig. 5f). Because the fibers exhibit a much smaller thermal expansion coefficient, bending deformation occurs in the bilayer structure upon heating. The value and direction of the principal curvature were shown to depend on the fiber orientation, which enabled predictable deformation.

In nearly all existing studies, the shape-changing capabilities of printed composites are pre-determined by the choice of matrix materials or additives prior to printing. Recently, Li et al.56 introduced novel hollow LCE fibers for multifunctional composites, which were fabricated using a coaxial spinning technique (Fig. 5g). These hollow fibers can be filled with various functional liquids during operation, introducing entirely new dimensions of functionality into already printed composites. For instance, this approach enabled rapid-response actuators driven by fluid flow, electrically driven actuators utilizing liquid metal for Joule heating, and light-guiding systems incorporating optical fibers.

Composites with tailored thermal and electrical properties

By incorporating various fillers and carefully designing their distribution, 3D-printed composites exhibit unique thermal and electrical properties suitable for applications such as heat dissipation, energy storage, and electromagnetic wave absorption.

Recent studies have highlighted the anisotropic thermal and electrical conductivity of 3D-printed composites containing nonspherical conductive fillers, such as carbon fibers and graphite flakes. These fillers tend to align during extrusion-based 3D printing, which forms efficient conductive pathways and results in physical anisotropy. For instance, Shemelya et al.57 demonstrated that FFF-printed composites with graphite flakes exhibit significantly higher in-plane thermal conductivity compared to the thickness direction. Huang et al.58 showed the alignment of short carbon fibers along the DIW printing direction. The resulting anisotropic electrical properties were further utilized to design a sandwich strain sensor that responds to bending deformation (Fig. 6a). Phatharapeetranun et al.59 investigated anisotropic dielectric properties of barium titanate nanofiber-reinforced polyvinylidene fluoride composites, which provided a pathway to develop ferroelectric composites for embedded capacitors and energy storage devices.

Fig. 6: Additive manufacturing of functional composites with unique thermal or electrical properties.
Fig. 6: Additive manufacturing of functional composites with unique thermal or electrical properties.
Full size image

a A sandwich strain sensor mounted on human wrist and the resistance change during cyclic wrist bending and releasing. (Image reproduced with permission from ref. 58, Copyright Elsevier). b A 3D printed lattice truss structure using continuous carbon fiber as a sensing element to achieve self-monitoring (Image reproduced with permission from ref. 60, Copyright Elsevier). c A carbon fiber-embedded 3D printed artificial hand with self-monitoring capabilities (Image reproduced with permission from ref. 63, Copyright Elsevier). d Photograph of the carbon fiber composite shells fabricated by the 3D printing process with outstanding shielding effectiveness (Image reproduced with permission from ref. 64, Copyright Springer Nature). e Schematics of 3D continuous carbon fiber structural battery structure and a printed functional, full cylindrical structural battery with same structure (Image reproduced with permission from ref. 66, Copyright Elsevier).

Embedding conductive fibers in 3D-printed composites enables sensing and structural health monitoring capabilities. These functions are realized by measuring and analyzing changes in electrical conductivity under external loading. For example, Wang et al.60 designed a smart thermoplastic composite lattice using continuous carbon fiber as a sensing element to achieve self-monitoring (Fig. 6b). Other studies have demonstrated sensing networks with copper and nichrome wires61, smart honeycomb structures62, and artificial self-monitoring hands63 utilizing continuous carbon fibers (Fig. 6e).

Conductive fillers also enable electromagnetic wave manipulation in 3D-printed composites. Yin et al.64 investigated the shielding effectiveness of FFF-printed polylactic acid (PLA) composites with carbon fibers. Conformal composite shells were designed and printed with a shielding effectiveness of 38.5 dB. This corresponded to a 7079-fold attenuation of electromagnetic waves, making them suitable for moderate EMI shielding applications (Fig. 6f). In a separate study, Yin et al.65 developed dual-layer metamaterial PLA composites embedded with graphene and continuous carbon fiber. The structure exhibited an ultra-wide absorption bandwidth of 32 GHz and a superior shielding effectiveness exceeding 63 dB.

Continuous fiber composites can be used as structural batteries for energy storage. In a representative study66, carbon fiber reinforcement served as the anode and current collector, while a doped polymer matrix with high electrical and ionic conductivity acted as the cathode (Fig. 6g). A solid polymer electrolyte coated onto the carbon fiber functioned as both an electrolyte and a separator. Both photopolymer and PLA were employed as polymer matrix materials. The study demonstrated that polymer-coated carbon fibers not only enabled energy storage but also improved the mechanical performance of the composites.

Composites with self-healing properties

Integrating self-healing capabilities into 3D-printed composites substantially enhances reliability and extends the service life of printed components. In thermoplastic composites, healing can be achieved by heating the material to melt the substrate and induce interpenetration of polymer chains across the interface. Recent studies have introduced alternative heating strategies that are more convenient for practical engineering applications. For example, Liu et al.67 recently demonstrated microwave-assisted welding of FFF-printed composites reinforced with short and continuous carbon fibers. The embedded fiber system functions as a susceptor that enables remote heating to effectively heal damaged regions.

Another strategy for repairing 3D-printed composites involves embedding microcapsules or microvascular networks containing healing agents to fill and repair cracks after damage. These systems can be integrated into composites using different methods. One approach involves coating thermoplastic composite filaments with microcapsules prior to FFF printing68. This method has been demonstrated in thermoplastic composites reinforced with CNTs. Alternatively, microcapsules can be directly embedded within the thermoplastic matrix during printing, such as in composites reinforced with continuous carbon fibers69. The recovery ratios of mechanical strength in the repaired composites were reported to exceed 80%.

Self-healing in printed thermosetting composites can be achieved by incorporating dynamic covalent bonds into matrix materials, such as those present in vitrimer networks. Most current studies in this area have focused on composites incorporating conductive70,71 or magnetic nanoparticles72,73. Due to their small size, these nanoparticles allow composites to be printed using high-resolution vat photopolymerization-based 3D printing techniques. Embedded nanoparticles in these composites not only function as mechanical reinforcements but also introduce additional functionalities, such as indirect Joule heating for self-healing or magnetic responses to assist crack closure. However, the application of this strategy in continuous fiber composites remains limited, which highlights a promising research direction. Recently, Huan et al.40 incorporated continuous fibers into a UV-curable vitrimer matrix with transesterification reactions. This approach not only imparted self-healing capabilities but also significantly enhanced inter-filament bonding in printed continuous fiber composites. Compared to extrinsic healing strategies, composites employing dynamic covalent bonds offer the advantage of repeated healing cycles and mechanical strengths approaching those of undamaged composites.

Optimization design for additive manufacturing of composites

Integration of design and 3D printing enables a tight workflow to create innovative solutions for composites with optimized material distributions, enhanced mechanical performance, and practical manufacturability. This approach has the potential to disrupt conventional composite product development by improving efficiency and productivity while unlocking new capabilities and functionalities.

Optimization design for enhancement of mechanical properties

Continuous fiber reinforcements significantly expand the design space of 3D-printed composites to enable the development of tailored mechanical properties and various functional capabilities. Early studies emphasized optimizing fiber patterns to maximize composite stiffness. To simultaneously design the shape of composite structures and fiber distributions, topology optimization has emerged as a versatile method3. Various design methodologies differ in their representation of composite topology and the parameterization of fibers.

Composite topology commonly uses the density method, where the local material distribution is defined by a volume fraction, or density, which ranges from 0 (void) to 1 (solid). Most density approaches rely on the Solid Isotropic Material with Penalization (SIMP) method to interpolate material properties4. Fiber orientations during optimization can be parameterized using either discrete or continuous methods. In discrete parameterization, fiber orientations are treated as a set of predefined angles74, whereas continuous parameterization allows smooth variations in fiber orientation. For example, fibers can align with the local principal stress or strain directions to optimize load transfer75. Boddeti et al.6,76 developed a multiscale topology optimization workflow for continuous fiber composites that used Finite Element Analysis (FEA) to determine the optimal composite shape and fiber orientation (Fig. 7a). Their framework included a material compilation procedure to translate mathematical optimization results into physically realizable 3D material layouts and machine code for printing. This workflow was validated through the design, fabrication, and testing of 2D cantilever beams and 3D components. Other continuous parameterization techniques include treating fiber angles as a continuous variable77 and using iso-parametric transformation to convert fiber angles into Cartesian vectors78. To enable a smooth fiber layout, methods such as aligning fibers along iso-contours of level-set functions79 and streamlines of vector fields80, adjusting filter radius, and constraining design variable derivatives81 have been used.

Fig. 7: Optimization design of composites for enhanced mechanical properties or functionalities.
Fig. 7: Optimization design of composites for enhanced mechanical properties or functionalities.
Full size image

a A density method-based design framework (Image reproduced with permission from ref. 6, Copyright Springer Nature) and b a level set method-based design framework (Image reproduced with permission from ref. 82, Copyright Springer Nature) that simultaneously optimizes macroscopic composite topology and microscopic fiber orientation. c Nonlinear topology optimization design of a soft swimming robot (Image reproduced with permission from ref. 8, Copyright John Wiley and Sons). From left to right: robot components, optimization objective versus iteration number, DIW printing process, and the swimming robot in operation. d Optimization design of magnetization distribution in magnetic composites for a frog-inspired swimming robot. Top images depict the design domain and candidate magnetizations, while the bottom images illustrate the target and optimized encoded shapes under magnetic actuation (Image reproduced with permission from ref. 9, Copyright Elsevier). e Topology optimization design for temperature-regulating composite cold plates. Different design outcomes result from varying inlet and outlet positions (Image reproduced with permission from ref. 10, Copyright Elsevier).

The aforementioned studies predominantly employed the density method to represent structural shapes. While effective, this approach requires filtering and projection techniques to control feature size and accurately define solid-void boundaries. Conversely, the Level Set Method (LSM) addresses these challenges by providing a precise description of structural shapes throughout the optimization process, which eliminates the need for material interpolation schemes. As one of the pioneering studies, Mokhtarzadeh et al.82 developed a design optimization methodology for 3D-printed continuous fiber composites using LSM (Fig. 7b). In this framework, fiber orientation is explicitly described as a continuous field parameterized by higher-order B-splines with the B-spline coefficients serving as optimization variables. This strategy is shown to notably promote smooth designs, prevent spurious features, and eliminate the need for additional filtering techniques.

Optimization design for innovative functionalities

Beyond conventional structural stiffness analysis, topology optimization has recently been extended to design functional composites with nonlinear deformation and multi-physics mechanisms. In these studies, objective functions are carefully selected based on specific performance requirements. Additional design variables are introduced to capture the constitutive behaviors of active phases or their interactions with external stimuli. In certain cases, multi-objective functions are employed to balance competing design criteria.

As one of the first studies, Maute et al.83 utilized LSM-based topology optimization to design active SMP composites fabricated using material jetting. The design framework included a simplified model to predict shape-changing behaviors, with an objective function aimed at minimizing the mismatch between achieved and target shapes. The study successfully determined the optimal distribution of soft and hard SMPs to enable a composite plate to achieve target bending or twisting deformations. Bhattacharyya et al.84 integrated a thermoviscoelastic constitutive model of SMPs with topology optimization to design tailored shape-changing structures, including a morphing beam capable of achieving non-axial shape recovery through simple axial loading programming. Recently, Boddeti et al.8 utilized density-based topology optimization to optimize DIW-printed soft composites for swimming robotic fish (Fig. 7c). By minimizing the difference between the simulated composite kinematics and prescribed target kinematics, the design framework optimized the composite microstructure, including the active fiber orientation and volume fraction. The resulting robotic fish demonstrated significant performance improvements, including a 50% increase in swimming speed, a 28% increase in turning rate, and a 55% reduction in turning radius compared to unoptimized prototypes.

Beyond shape-changing composites, topology optimization has recently been extended to other domains, including the design of active magnetic composites with optimal topology, magnetization distribution, and applied magnetic fields9 (Fig. 7d), thermal design of composite cold plates10 (Fig. 7e), and the development of porous multi-material composites with tunable thermal expansion11.

Outlook

Enhancing mechanical properties and scalability of composites 3D printing

Despite the exceptional design flexibility offered by AM, it still lags behind conventional composite fabrication methods in terms of structural integrity, scalability, and production efficiency. Currently, composite printing techniques remain predominantly limited to laboratory-scale research or prototyping applications. These limitations are primarily attributed to inherent constraints of the layer-by-layer deposition process, including weak filament-to-filament and layer-to-layer bonding, interfacial imperfections, and void formations within printed structures, all of which negatively impact mechanical performance. Future research should aim to address these critical challenges by developing high-performance resins that facilitate rapid, uniform curing even in the presence of reinforcing additives, such as frontal polymerization resins or dual-curable (thermal and UV) resins. Additionally, integrating advanced in-situ consolidation methods, including microwave or laser-assisted heating (as highlighted in Section “Continuous fiber reinforced composites”), could significantly enhance interlayer bonding and reduce void formation.

Large-format additive manufacturing (LFAM) is an emerging approach for printing polymers and composites at the meter scale, offering substantial potential for structural applications across industries such as aerospace, marine, and construction. However, several primary challenges remain, including limited deposition speed, poor interlayer bonding, significant thermal shrinkage, and deformation during and after printing. Addressing these challenges requires coordinated innovations in printing conditions, post-processing strategies, material formulations, and process modeling. For example, pellet-based extrusion systems20,21 have gained widespread use in LFAM due to their high deposition rates and lower material costs compared to filament-based systems. Further improvements in productivity can be achieved through robotic arm integration85, which enables multi-axis printing and increased flexibility, as well as through post-processing techniques such as polymer welding or mechanical joining to assemble multi-segmented parts86. On the materials side, LFAM necessitates the development of printable materials with reduced shrinkage and warping, improved interlayer adhesion, and enhanced mechanical performance87. These properties are especially critical as the scale of the printed part increases and thermal gradients become more pronounced. In terms of modeling and simulation, gravitational effects, thermal shrinkage, and internal stress development must be accounted for to ensure structural integrity and dimensional accuracy. For readers interested in further exploring this topic, Goh et al.88 provide a comprehensive overview of fabrication processes, materials, and design considerations in LFAM; Pignatelli et al.89 offer a market-oriented review focused on polymer pellet-based 3D printing; and Urhal et al.85 emphasize the role of robot-assisted additive manufacturing in enabling flexible, large-scale printing.

Integrating adaptable multifunctionality into printed composites

Contemporary composites are increasingly expected to provide multifunctionality in addition to mechanical support. However, current CFRP printing methods primarily achieve specific functionalities through pre-selected combinations of matrix materials and additives. Future studies are suggested to enable the printed composites to dynamically integrate multifunctional components during operation or adapt functionality to environmental or application-specific demands. One promising strategy involves incorporating microvascular networks90, which are traditionally used as extrinsic self-healing systems, to deliver various functional fluids dynamically within printed composites. For instance, embedding electrochemical or magnetorheological fluids, or phase-change materials within microchannels, can enable composites to reversibly alter their properties under external stimuli such as electric or magnetic fields, or temperature fluctuations.

Another promising direction is the integration of reprogrammable functional matrix materials. For example, LCEs, known for their shape-changing and actuation behaviors in 4D-printed composites, can be further enhanced by incorporating dynamic covalent bonds91. These dynamic bonds facilitate random bond-exchange reactions, disrupting mesogen alignment and the monodomain state, allowing the material to revert to a polydomain state. Consequently, this process enables the material’s actuation pathways and functional responses to be reprogrammed repeatedly during operation.

Advancing theoretical modeling of composite printing processes

While significant experimental progress has been made in composite AM, theoretical and computational modeling remains comparatively underdeveloped. Future research should develop advanced modeling frameworks capable of capturing the complex interplay among diffusion, reaction kinetics, and mechanical behavior during composite printing. These advanced computational models should accurately describe the kinetics of material deposition, particle or fiber alignment under external stimuli, resin curing processes in the presence of additives, and the resulting mechanical and functional properties of the printed composites.

These comprehensive computational models would substantially deepen the understanding of the relationships among materials, processing conditions, and composite properties. Additionally, integrating these models with in-situ sensing and monitoring technologies to establish digital twins would facilitate real-time predictive control and virtual prototyping. This approach would significantly enhance manufacturing productivity, improve quality control, and better align AM practices with industry standards and requirements.

Incorporating manufacturing processes into design frameworks

Current design frameworks for composite materials often overlook manufacturing constraints, defects, and the complex constitutive behaviors of composites. While ignoring these challenges enables an efficient design process, it potentially results in printed composites with properties that deviate from theoretical predictions or, in some cases, designs that are not manufacturable.

To address these limitations, there is a growing need to integrate constitutive modeling into existing design frameworks. This integration would enhance the ability of existing design technologies to address multi-scale and multi-physics challenges. Furthermore, by incorporating manufacturing constraints, as well as material and process parameters, as active design variables, these comprehensive frameworks could support a product-process codesign methodology, which allows for simultaneous optimization of product topology, fiber placement, and manufacturing conditions, thereby maximizing the mechanical performance and functionality of 3D-printed CFRPs.

It is important to note that developing such design frameworks would involve numerous parameters and processing variables, making calibration and iterative numerical simulations time-consuming. To address these challenges, developing advanced computational techniques, such as integrating machine learning techniques92 with constitutive modeling, could enable rapid predictions and streamline the design of material structures. This approach has the potential to dramatically accelerate the design process, improve the accuracy of predictions, and facilitate the creation of innovative composite materials and structures.

Conclusions

In summary, design optimization and 3D printing serve as two foundational pillars advancing the development of modern composite materials. Recent advancements in 3D printing technologies have significantly enhanced composite quality and expanded the range of printable materials. Design optimization plays a pivotal role in guiding 3D printing to produce composites with exceptional mechanical properties and diverse functionality. Conversely, 3D printing presents enriched challenges that drive the continuous development of advanced design methodologies. Future research should focus on overcoming current limitations, such as improving the mechanical performance and scalability of 3D-printed composites, incorporating adaptable multifunctionality, and advancing theoretical modeling frameworks. By addressing these needs, the field of composite 3D printing can evolve toward realizing its full potential in both industrial and research applications, which paves the way for groundbreaking materials and scalable manufacturing processes.