Owing to the brittle nature, the mechanical properties of ceramics are highly sensitive to processing defects. As the currently available AM techniques combine both 3D printing and conventional manufacturing, such as debinding and sintering, which can result in defects by both manufacturing techniques. In this review, a variety of defects were found in the final products as: porosity, agglomerates, cracks, surface defects, large particles, delamination, deformation, etc. While some of these defect types are also common in conventional techniques, there can be differences in shapes, sizes, and locations due to different formation causes, resulting in different influences on the mechanical properties of the final products.

Porosity, as a common type of defect for both AM and SM, can be formed during the entire AM process, including slurry preparation, printing, debinding, and sintering, existing as voids or bubbles in the final products. Highly viscous ceramic slurry can trap air bubbles during preparation or printing, which are not effectively removed by subsequent thermal treatments, leading to micro-sized residual pores of diverse shapes in the final parts (20, 22, 26, 76, 77). Additionally, various situations can contribute to the formation of porosity or cracks: weak bonding between successive layers can lead to insufficient layer fusion (23, 25); solvent evaporation from the ceramic paste as it dries on the exposed surface before the next layer is added may cause uneven shrinkage during sintering (76); sedimentation of ceramic particles can lead to voids between layers after sintering (78, 79). For post-treatments, the removal of binders during the debinding process can also leading to the presence of residual pores in ceramic parts (20, 76). Unsuitable sintering conditions (52) or insufficient ceramic particles dispersion (76) may not be fully densified, resulting in the formation of pores. Porosity have been addressing as one main type of critical defects in AM ceramics. Compared with milling, it was found that DLP zirconia have a larger scale of critical defect size, which can explain its relatively low Weibull modulus (20). Despite zirconia, Abreu et al. (49) also found individual surface pores as the fracture origin of lithium disilicate by robotic material extrusion, while Marsico et al. (48) observed porous region as one of the main fracture initiations in lithium disilicate by DLP. The existence of porosity could also affect hardness, resulting in lower value in comparison with milled ceramic with similar compositions (27, 28).

Agglomerate is another common defect for AM and SM. Agglomerates are clumps or clusters of particles that stick together in a material, which can be attributed to incomplete dispersion during the slurry-forming process (44). The high viscosity of ceramic slurry can hinder the uniform dispersion of these particles (21). Furthermore, the presence of residual powder particles or incomplete removal of powder can lead to the creation of agglomerates in the printed ceramic structure (20). Additives or binders in the ceramic formulation may also influence particle interactions and lead to agglomeration (80). During sintering, agglomerations can persist within the ceramic parts (81). The presence of agglomerates can act as stress concentrators, leading to the initiation of crack propagation and fractures, consequently reducing the overall strength of the material (82). In addition, they correspond to large under-densified zones, which can have negative effects on its translucency and mechanical properties (25).

Cracks, including microcracks along the interface of layers and macrocracks, pose a significant risk to the material strength in AM ceramic. These cracks mainly initiate and propagate at the layer interfaces and grain boundaries, where the bonding is weaker compared to other regions (82). High solids volume fraction is essential to achieve sufficient strength. In the study by Jang et al. (17) and a decreasing zirconia volume fraction resulted in an increasing number of cracks. This can be attributed to a low solid volume fraction, which results in the presence of undensified regions, thereby leading to porous regions along the layers. While maintaining a high solid volume fraction is important, the ceramic slurry needs to retain sufficient flowability and dispersibility to achieve a high quality of the final product (17). Transportation, handling, and movement of the green part throughout the printing and post-processing stages may lead to deformation. Consequently, the green part can generate cracks along the layers' interface due to the influence of uneven internal stress (82). Additionally, uneven shrinkage during the debinding and sintering processes can induce stress at the interfaces between different layers, resulting in the formation of cracks and deformation of the green part (83). Moreover, a reduced solid content in the ceramic slurry results in a higher proportion of polymer content in the green part, which leads to increased shrinkage during curing, consequently leading to internal stress (76). In addition, particle segregation and sedimentation of coarse particles after sintering can lead to larger particles at layer boundaries, which can lead to inhomogeneous grain distribution after sintering and different mechanical properties in different printing layer orientations. This can lead to the formation of cracks and compromise the mechanical properties (79). Delamination was observed as a result from weak bonding strength between successive layers (25). This can cause weak interfaces, making the green part more vulnerable to cracks. In addition, thicker layers can lead to layer union issues because larger layer line defects or even delamination of the layers can occur (44, 84).

Another concern for AM ceramics is surface roughness, or surface defects, as high roughness has been proven to impact mechanical properties, fit, aesthetics, bacterial adhesion and wear on opposing teeth or restorations (26, 67, 85). The main contributor to roughness is the “Step effect” or visible layer lines caused by layer-by-layer construction (21, 44), while the residual ceramic slurry after printing serves as the secondary factor (37). A study on SLA zirconia (37) showed that the surface parallel to the printing platform exhibited an average surface roughness of 0.71 μm (Ra), while the surface perpendicular to the printing platform was significantly rougher (2.91 μm). Both orientations considerably exceeded the required value (0.2 μm) for dental restorations (86), indicating the significance of subsequent surface finishing, such as polishing or glazing. Thorough surface polishing has been proven to reduce the effect of roughness, therefore significantly promoting the flexural strength and fatigue strength of AM zirconia (25, 37, 55). Conventional glazing technique or glass infiltration can also generate a shiny and smooth surface for monolithic ceramics, while their influence on mechanical properties of AM ceramics yet lacks reports.

Other microscopic and macroscopic defects, such as large particles, deformation, contamination, and machining damage, were also observed previously (20, 29, 53, 82). Previous studies also reported small amounts of heavy metals detected in zirconia, possibly originating from contamination during the purification and production processes (29, 87). Furthermore, printed zirconia is very sensitive to manual defects introduced before sintering and after printing because of the extreme fragility of the products (37).

Printing layer orientation has been recognized as a significant factor influencing the mechanical properties of AM materials. Studies have shown that AM ceramics can achieve the highest flexural strength, Young's modulus, and fracture toughness when the printing layers are perpendicular to load in the bending tests (25, 26, 29, 44). One explanation is that this orientation avoided the step effect on the tensile surface, thereby reducing the effect of surface roughness on flexural strength (44). Though the step effect can be removed by delicate surface polishing, sometimes microcracks along the layers still remain. When printing layers are perpendicular to the load, these interlayer cracks are less likely to appear in the zone where the stress is concentrated. The effect of orientations could also be related to the difference in microstructures of different orientations. During the printing process, large particles are likely to settle before the layer is cured, resulting in coarse grain at the bottom (79). Additionally, Lu et al. (37) observed the presence of elliptical-shaped pores that acted as fracture initiators, in the long axis parallel to the layers. These pores may be distributed within the sintered ceramic, but specimens with a printing orientation parallel to the load may be more prone to the influence of these pores due to their sharper shape in the direction. However, it is yet unclear how the different mechanical properties of different printing layer orientations would influence the mechanical performance of clinical restorations, and if this risk can be eliminated by optimization of restoration design and printing orientation/angle.

Ceramic restorations fabricated by AM can exhibit a satisfying visual quality, including structural integrity, smooth surface, and fine details, indicating the promising clinical application of the technique, as shown in Figure 4. However, literature regarding the clinical performances of AM restorations is still very limited. The fracture resistance of AM 3Y-TZP crowns was reported as weaker than SM crowns both before and after 1.2 million loading cycles (56). While another study found that AM 3Y-TZP can achieve a comparable fatigue strength to SM 3Y-TZP by optimizing printing layer orientation and surface polishing (55). Regarding low-temperature degradation on AM 3Y-TZP, an increase in the monoclinic phase can be observed, without a significant decreasing in flexural strength (24). However, hardness and fracture toughness (31) seem to be affected by aging. For wear, both DLP-manufactured 3Y-TZP specimens with horizontal and vertical printing orientations have similar friction efficiency and negligible volumetric substance loss after occlusal wear by zirconia antagonists (57). In addition, no difference was found between the bond strengths of printed and milled 3Y-TZP substrate and porcelain veneer, by either Schwickerath adhesion test (54) or shear bond strength test (27). However, the adhesion between AM ceramic and abutment tooth is another essential topic, but relevant reports are still lacking. Beyond geometry, clinical performance can be compromised by defects that directly impact the ceramic's structural integrity and longevity. Layer cracks and delamination are common in additively manufactured ceramics due to insufficient bonding between layers and accumulated internal stresses. These cracks can act as initiation points for failure, particularly in the cyclic loading environment of the oral cavity where chewing forces can exacerbate structural weaknesses. Furthermore, density variations and porosity arise from incomplete filling between layers or air entrapment during printing, leading to stress concentration points that reduce the overall mechanical robustness. Such inconsistencies are especially concerning for dental restorations, as they may lead to premature wear or fracture under masticatory loads (82).

In this review, AM ceramics have reached a level close to milled ceramics in terms of mechanical properties, but the reliability of the final product is still a general problem associated with processing defects. Therefore, technological development and further research are still necessary for defect identification and control to improve the reliability of the materials and fabricated restorations. The second common issue of AM ceramics is the influence of printing layer orientations. Though some studies have addressed the strongest or weakest orientations for standard specimens, the situation of clinical prostheses is much more complex. Yet it is unclear how the real restorations would be impacted and if/to which extent such effects can be avoided by the optimization of design. Regarding different ceramics, while 3Y-TZP has received the most in-depth investigations, its long-term clinical performance as a framework material requires further research, such as the duration of the restorations and the interaction with the bonding to natural teeth or porcelain veneer. For monolithic zirconia and glass ceramics, scientific reports are still very limited in comparison to 3Y-TZP. More technological explorations that combine both mechanical and optical properties are expected for these materials. Additionally, the effects of humidity and wear in the oral medium are of concern when evaluating the properties. Literature investigating surface treatments, coatings, or post-processing methodologies to improve surface quality and toughness are still lacking from the literature and should be encouraged. The future trends for the development of dental ceramic 3D printing include the development of higher-performance and more reliable ceramics, translucent ceramics for aesthetic applications, biomimetic gradient ceramics through multi-material printing, and advanced technologies to increase production efficiency.