Digital light processing based multimaterial 3D printing: challenges, solutions and perspectives

Multimaterial (MM) 3D printing shows great potential for application in metamaterials, flexible electronics, biomedical devices and robots, since it can seamlessly integrate distinctive materials into one printed structure. Among numerous MM 3D printing technologies, digital light processing (DLP) MM 3D printing is compatible with a wide range of materials from hydrogels to ceramics, and can print MM 3D structures with high resolution, high complexity and fast speed. This paper introduces the fundamental mechanisms of DLP 3D printing, and reviews the recent advances of DLP MM 3D printing technologies with emphasis on material switching methods and material contamination issues. It also summarizes a number of typical examples of DLP MM 3D printing systems developed in the past decade, and introduces their system structures, working principles, material switching methods, residual resin removal methods, printing steps, as well as the representative structures and applications. Finally, we provide perspectives on the directions of the further development of DLP MM 3D printing technology.


Introduction
Additive manufacturing (also known as 3D printing) is an emerging advanced manufacturing technology that generates parts in a layer-by-layer fashion and can fabricate highly complex 3D objects with nearly zero constraints.In contrast to Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence.Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
the subtractive nature of conventional machining, the additive nature of 3D printing minimizes the material waste, and makes it applicable to a wide variety of materials such as polymers, ceramics, metals, composites, and others.Due to the advantages such as customization freedom, faster production, less waste, cost-effectiveness, as well as tangible design and product testing, 3D printing has found applications in diverse fields including electronics, automotive, medical, and aerospace.A report from Grant View Research reveals that the worldwide 3D printing market reached a valuation of $16.75 billion in 2022, and is anticipated to experience a compound annual growth rate of 23.3% from 2023 to 2030 [1], which supports the notion that the continuous research and development in 3D printing is crucial to global economic competitiveness.
FDM is an extrusion-based 3D printing technology.During the printing process, the heated printing nozzle melts or softens thermoplastic filament, depositing it layer by layer onto a build tray to create 3D structures.As shown in figure 1(a), MM 3D printing capability can be readily realized on an FDM 3D printer by simply adding extra printing nozzles.By printing primary building material along sacrificial material, FDM MM 3D printer can print structures with overhanging or spanning features, which cannot be produced by a single material FDM printer.As presented in figure 1(b), an FDM MM 3D printer fabricated a Hilbert cube where the white sacrificial material can be manually removed [35].MM 3D printing also enables the fabrication of structures consisting of stiffness and soft components.Figure 1(c) shows a bioinspired spring origami where the stiff polymer facets were printed onto the rubberlike substrate [36].However, the weak interfacial bonding between stiff polymer and soft substrate may lead to delamination during folding process.To overcome this issue, Ye et al proposed a wrapping method for FDM MM 3D printing where the stiff part was wrapped by soft skin to form robust interfacial bonding between two parts.Figure 1(d) presents a thick-panel Miura origami structure which can be folded more than 100 cycles without failure [37].
DIW, as another extrusion-based 3D printing technology, deposits viscoelastic inks with a shear thinning effect and viscosity ranging from ∼10 3 to ∼1 Pa•s as the shear rate increases from 1 to 100 s −1 .As illustrated in figure 1(e), MM 3D printing of DIW can also be achieved by adding extra independent printing nozzles.Figure 1(f) presents a custom-designed, large-area 3D bioprinter with four independently controlled printheads which was used to fabricate engineered tissue constructs replete with vasculature, multiple types of cells, and cells extracellular matrix [38].Although MM 3D printing can be realized by adding extra nozzles, the printing speed would be inevitably compromised with the number of nozzles.To address this issue, as shown in figure 1(g), Liu et al reported an MM extrusion bioprinting platform that can continuously extrude multiple encoded bioinks with fast material switching between different reservoirs, facilitating rapid manufacturing of complex constructs and even tissues [39].To further improve the frequency and resolution of the switching materials, as shown in figure 1(h), Skylar-Scott et al reported an MM multinozzle 3D printing technique to create voxelated soft matter that incorporates ultrarapid MM switching to enhance printing speed and structural complexity [40].
Different from extrusion-based 3D printing technologies, as shown in figure 1(i), inkjet 3D printing technology constructs 3D objects by employing piezoelectric-based inkjet printing head to deposit low-viscosity ink droplets at the nano/microscale onto a build tray.During the printing process, the inkjet printing head selectively deposits ink droplets onto the build tray to form 2D patterns, which are immediately solidified upon exposure to the UV light.Inkjet is an ideal technology for MM 3D printing as the printing head can simultaneously deposit ink droplets formed with different materials [47].The commercialized inkjet MM 3D printer can fabricate an MM 3D structure with a maximum size of 1000 mm × 800 mm × 500 mm, and a droplet deposition resolution of 600 dots per inch [48].Thus, inkjet MM 3D printing enables the control of droplet deposition at voxel scales, and can be used to realize the physical visualization of data sets commonly associated with scientific imaging, such as white matter tractography data of the human brain visualizing bundles of axons, which connect different regions of the brain (figure 1(j)) [41].By printing structures with active materials, which exhibit large deformation in response to environmental stimulus, inkjet MM 3D printing can be used to realize fourthdimensional (4D) printing [49][50][51][52][53] to fabricate 3D structures that can change their shapes over the fourth dimension-'time' (figure 1(k)) [42].Although the commercialized inkjet MM 3D printer is a powerful tool to fabricate MM 3D structures, it has several shortcomings, including expensive price of each printer and printing materials, limited materials available for the printer as well as the proprietary and inextensible hardware and software architectures for current MM 3D printers.To address these issues, Sitthi-Amorn et al developed a highresolution, low-cost, and extensible inkjet MM 3D printing platform which was built exclusively from off-the-shelf components and can achieve a resolution of at least 40 µm with up to 10 different materials that can interact optically and mechanically (figure 1(l)) [43].
Vat photopolymerization (VP) is a 3D printing technology that provides a benefit of relatively lower cost and higher resolution.During VP printing process, the photopolymer resin contained in resin vat is cured layer-by-layer through laser scanning or ultraviolet (UV) light patterning that triggers localized photopolymerization, converting liquid resin to 3D objects.Depending on the curing light source, VP can further be classified into stereolithography (SLA) [54][55][56], DLP [57][58][59][60][61], two-photon polymerization [62][63][64][65][66][67], and volumetric 3D printing [68][69][70][71].Compared with the other three techniques, DLP performs localized photopolymerization through projecting 2D UV patterns on the surface of liquid polymer resin, and combines the feature of high resolution with fast speed.As illustrated in figure 1(m), DLP MM 3D printing can be realized by adding multiple resin vats, and it has been successfully used to fabricate MM structures with exceptional properties  and functions.Figure 1(n) presents an MM 3D-printed negative thermal expansion metamaterial structure consisting of two different materials with distinct thermomechanical properties [44].Figure 1(o) shows a stiff lattice reinforced hydrogel composite where the local stiffness of the hydrogel composite can be tuned by adjusting the diameter of lattice structure [45].In order to enlarge the printing area and enrich the range of printable materials for DLP MM 3D printing, Cheng et al report a centrifugal MM 3D printing that utilizes centrifugal force to remove the residual resin stuck to the printed structure during material switching so the printer can fabricate heterogeneous 3D structures in large volume structure (figure 1(p)) made of materials ranging from hydrogels to functional polymers, and even ceramics [46].
Figure 2 compares the MM 3D printing capabilities of the above-mentioned four 3D printing technologies in terms of printing area, speed, resolution, printable materials, complexity of printed structures, and economy.Detailed information can be found in table S1 in supplementary information.Among them, DLP MM 3D printing can print MM 3D structures with high resolution, high complexity and fast speed.It is compatible with a wide range of materials, from hydrogels to ceramics.Its printing area has been significantly improved by the recently developed centrifugal MM 3D printing [46].More importantly, a DLP MM 3D printer mainly consists of a UV light engine and a few translational stages, which are off-theshelf components.Therefore, it is affordable for most research labs.
This review paper aims to summarize the recent advances in DLP MM 3D printing technologies.Section 2 introduces the fundamental mechanisms for DLP 3D printing including the printing principle, resin composition, photopolymerization reaction, light propagation, and UV pattern projection.In section 3, we introduce the material switching methods, soft composites [45], composite metamaterials [46], robots [76], flexible electronics [81], and biomedical devices [84].From [45].Reprinted with permission from AAAS.Reproduced from [46].CC BY 4.0.From [76].Reprinted with permission from AAAS.Reprinted with permission from [81].Copyright (2023) American Chemical Society.Reproduced with permission from [84].material bonding mechanism, layer-slicing manners and printing sequences, the material contamination issues during material switching, as well as solutions to address the material contamination issues.In section 4, we introduce the typical DLP MM 3D printing systems using different material switching and residual resin removal methods.Finally, section 5 concludes the article and provides perspectives on the directions of the further development of DLP MM 3D printing technology.

Fundamental mechanisms for DLP 3D printing
According to the direction where the UV pattern passes through, the single material DLP 3D printing systems can be classified as the 'top-down' configuration (figure 4(a)) and the 'bottom-up' one (figure 4(b)).On both systems, the printed part grows on the printing stage, which moves vertically to leave space for printing the coming layer.As shown in figure 4(c), the resin vat contains the liquid polymer resin, which mainly includes monomers, crosslinkers, photoinitiators, and photoabsorbers (figure 4(d)).Upon UV irradiation, photoinitiators decompose into free radicals that initiate photopolymerization.Photoabsorbers absorb excess light and confine photopolymerization to the required space to improve printing resolution in both horizontal and vertical directions.Monomers and crosslinkers are the building blocks (figure 4(e)) that form the photopolymerized networks (figure 4(f)).
Propagation : Oxygen inhibition : UV light irradiation splits one photoinitiator molecule into two free radicals R * ( * represents that the species have an active site), as shown in equation (1).Photopolymerization is initiated when a free radical R * attacks a monomer or crosslinker molecule M, and the free radical becomes RM * (equation ( 2)).This process propagates as the free radical attacks the neighboring unreacted monomer or crosslinker molecules to form a polymer chain with an active site RM * m (equation ( 3)).Photopolymerization is terminated when two radicals react with each other.The termination reaction occurs in two mechanisms: combination where two radical chains form one dead polymer chain RM m+n R; disproportionation where two dead polymer chains are produced RM m + RM n (equation ( 4)).Photopolymerization can also be inhabited when free radical reacts with oxygen (equation ( 5)).
As illustrated in figure 5(a), during DLP 3D printing, the degree of photopolymerization gradually attenuates along the path where the UV light travels through.The photopolymerization process can be modeled using a set of first-order reaction equations.As shown in figure 5(b), the light intensity at the surface of polymer resin is I 0 , and it is attenuated when the light is absorbed by photoinitiators, photoabsorbers, and other light reactive species in the resin.The attenuation of light intensity I (z, t) follows the Beer-Lambert law as: where α and C I (z, t) are the molar absorptivity and concentration of photoinitiator [88,90].
After absorbing the light energy, the photoinitiators decompose into active species.The evolution of the photoinitiator concentration can be described by a first order chemical reaction differential equation: where β is the decomposition rate of photoinitiator [88,90].
Likewise, the concentrations of radicals C R (z, t) and oxygens C O (z, t) can be calculated as: where m is the number of radicals generated in photodecomposition, k Term is the termination rate, and k O is the reaction rate between oxygen and radicals [88,90].Moreover, the unreacted species (monomers and crosslinkers) in the solution are gradually consumed during photopolymerization.In this process, the C=C double bonds on the functional groups are reduced, and the concentration of the unconverted C=C double bonds (unreacted monomers and crosslinkers) can be calculated as: where C M (z, t) is the concentration of unreacted species, k p is the propagation rate [88,90].Finally, the degree of conversion can be calculated as: During printing process, DLP projector generates the UV light patterns (figure 6(a)), which are modulated by the digital micromirror device (DMD) chip.A DMD is a small silicon chip consisting of millions of mirrors which are only a few microns wide.Each mirror has a reflective aluminum surface, and can be titled to the 'on' or 'off' positions by electrostatical switching.A UV pattern or image is created by the UV light hitting the mirrors in the 'on' position, while the UV light hitting the mirrors in the 'off' position creates black pixels.Moreover, the grayscale UV patterns can be generated by switching the mirror from the black to white pixels multiple times with different frequencies.As shown in figure 6(b), the intensity of a UV pattern distributed on a horizontal plane is the summation of the light intensity from each pixel, which follows the Gaussian distribution (figure 6(c)).Thus, the intensity from one pixel and a UV pattern can be calculated as: where I pixel (g) is the light intensity of a single pixel with grayscale value g, σ is the radius of Gaussian beam, x and y correspond to an arbitrary point in the plane z = 0, x i and y j represent the coordinates for the center position of each Gaussian beam [89,90].Figures 6(d)-(f) give one dimensional illustrative example that shows how individual Gaussian beams add up. Figure 6(d) presents a single Gaussian beam illustrated in black, where the purple dashed lines mark the nominal boundaries of a single 10 µm pixel.Figure 6(e) shows a ten-pixel light pattern.The blue line represents its light intensity distribution, which is the summation of light intensity of ten individual Gaussian beams.The purple dashed lines mark the nominal boundaries of ten 10 µm pixels.Figure 6(f) shows a ten-pixel light pattern with different grayscales.The left five beams have 100% of maximum light intensity for each pixel (full bright), while the right five beams have 50% of maximum light intensity for each pixel (50% grayscale).The overall light intensity of the pattern is still the summation of ten individual Gaussian beams with different grayscales.

Critical challenges in DLP MM 3D printing
For DLP 3D printing, the printing process occurs in the resin vat, which is a liquid environment.To achieve the MM 3D printing capability, it is necessary to switch the liquid environments.Over the past decade, there have been two major methods of MM switching: vat switching and resin switching.For the vat switching method, the 3D printing system needs to be equipped with multiple resin vats, and the resin vats are alternatively switched for printing different materials.As illustrated in figures 7(a) and (b), the vat switching method enables the MM 3D printing capability for the DLP 3D printing systems following top-down and bottom-up projection manners respectively.When a layer of blue material is being printed, the printed part is placed in the blue resin.If the next layer is printed with yellow material, the yellow resin vat will be switched to the position below the printing stage so that the printed part can be placed in yellow resin for the next layer printing.For the resin switching method, the 3D printing system is equipped with only one resin vat, but the resin inside the vat can be switched through a fluidic system.The resin switching method enabled MM 3D printing systems, as illustrated in figures 7(c) and (d), for the top-down and bottom-up projection manners respectively.As shown in the illustrations, the resin vat is fixed and connected with multiple resin reservoirs.During the resin switching process, the valve connected to the outlet is open first so that the current resin in the vat can be quickly removed.After that, the valve to the outlet is closed, and the valve connected to the reservoir containing the resin to print the next layer is opened so that the vat can be filled with the corresponding resin.The resin switching process iterates to complete the MM 3D printing job.
Different from single material printing, the layer-slicing manners and printing sequences greatly depend on the spatial arrangements of the different constituent materials.As shown in figure 8(a), to print a pure-white Eiffel Tower (height: 50 mm), we only need to slice its CAD model into 500 layers of patterns (layer thickness: 100 µm), and use the white-color resin to print the whole tower.As it is a single material printing process, it only takes about 125 min to print this pure-white Eiffel Tower. Figure 8(b) shows an Eiffel Tower where the bottom half is white and the top half is black.Since the direction along which the materials alter is parallel to the printing direction, we just need to assign the printer to print the first 210 layers in a white color and the other 290 layers in a black color.To print such a two-color Eiffel Tower only takes 126 min as it only requires to switch material from white to black once.Different from figure 8(b), to print a two-color Eiffel Tower in figure 8(c) is a complex and time-consuming process.Because the direction along which the materials change is perpendicular to the printing direction, the printer needs to respectively print the black and white parts at every layer.It requires about 1000 min to print such a two-color Eiffel Tower in figure 8(c) as the printer needs to switch materials twice for printing each layer.The material arrangement of the Eiffel Tower in figure 8(d) seems more complicated than that of the one in figure 8(c), but the number of switching times and the total printing time required for both of them are the same.
Figures 9(a)-(d) illustrate detailed processes to print a layer consisting of two materials.As shown in figure 9(a), the printing platform is first placed in the blue polymer resin (chemical details can be seen in figure 9(e)).The UV pattern projection results in the solidification of the blue polymer, and the chemical structure of the photopolymerized blue polymer is described in figure 9(f).It should be noted that there is a small amount of unreacted monomer or crosslinker molecules remaining in the polymerized structure (figure 9(b)).In figure 9(c), the printing platform is moved to the yellow polymer resin to print the yellow part.The projection of the UV pattern leads to the free radical photopolymerization that cures the yellow part, where the chemical structure of the photopolymerized yellow polymer can be seen in figure 9(g).More importantly, at the interface, the radicals within yellow polymer can also attack those unreacted monomers, crosslinking with the blue polymer network and resulting in robust chemical bonding between the yellow and blue parts (figure 9(h)).
So far, the MM DLP 3D printing has demonstrated the excellent capability of bonding various materials with distinctive properties or functionalities.As presented in figures 9(i)-(k), Ge et al reported an MM DLP 3D printing approach for creating complex heterogeneous 3D structures, which are composed of highly stretchable acrylamide poly(ethylene glycol) diacrylate (PEGDA) (AP) hydrogels covalently bonded with a wide variety of photocurable polymers such as rigid polymers, elastomers, shape memory polymers (SMPs), acrylonitrile butadiene styrene (ABS)-like polymers, and many other (meth)acrylate-based resins [45,82].To present the robust interfacial bonding between the AP hydrogel and other polymer, as shown in figure 9(i), the authors stretched the heterogeneous dog-bone specimens where the upper half part is AP hydrogel and the lower half part is another polymer [82].Since the photocurable polymers exhibit greater stiffness than AP hydrogel, deformations predominantly occur in the hydrogel part.The energy required to break the hydrogelpolymer interfacial bonding exceeds that needed to break the AP hydrogel.Consequently, each heterogeneous specimen fractures on the hydrogel part rather than at the interface.The strong interfacial bonding between AP hydrogel and other polymers allows rapid fabrication of hydrogel-polymer hybrids for various applications.Figure 9(j) presents a rigid polymer reinforced hydrogel composite fabricated by MM DLP 3D printing technology, where the horseshoe-shaped rigid polymer can effectively adapt to the great stretchability of the hydrogel [45].Figure 9(k) shows an MM DLP 3D printed cardiovascular SMP stent with drug releasing function by integrating hydrogel into the SMP rods [45].Figure 9(l) demonstrates a soft pneumatic actuator (SPA), where its body is printed with elastomer and the hydrogel circuit is printed on its back to sense the bending deformation [45].Moreover, the recent research shows that the MM DLP 3D printing can even bond ceramic with elastomer (figure 9(m)) [46].The uniaxial tensile tests on the ceramic-elastomer hybrid specimen indicate that the interfacial strength surpasses that of the elastomer (figure 9(n)) [46].The 90 • peeling test results confirm that interfacial fracture is cohesive and the measured interfacial toughness reaches ∼1200 J•mm −2 (figure 9(o)) [46].The robust interfacial bonding between ceramic and elastomer allows the fabrication of complex ceramic structures with overhanging parts.To demonstrate this distinctive capability, the authors printed a ceramic spider with its body supported by a polymer part (figure 9(p)) [46].The debinding and sintering processes remove the polymer part and organics in the ceramic green body, leaving a ceramic spider with an overhanging body.
As DLP 3D printing occurs in a liquid resin environment, the material contamination generated from the MM switching process is inevitable for both vat switching and resin switching methods.Figure 10(a) illustrates a top-down MM DLP 3D printing system with the vat switching method.After a layer of blue material is printed, the printing stage is lifted from the blue resin vat, and the residual blue resin is stuck to the printed part as well as the printing stage.In order to print the next layer with yellow material, the printing stage needs to be submerged into the yellow resin vat.It is unavoidable that the yellow resin would be contaminated by residual blue resin.The contamination turns to be worse as multiple rounds of MM switching take place.As shown in figure 10(b), compared with the topdown DLP MM 3D printing system, the contamination is less severe but also inevitable in the bottom-up system as only a few top layers of the printed structure need to be submerged into the resin vat.For the resin switching method, as shown in figure 10(c), in the top-down DLP MM 3D printing system, the residual resin would be stuck to the printed structure, the printing stage, and the inner wall of the resin vat.As the printed part does not have the direct contact with the resins contained in the resin reservoirs, the degree of contamination does not accumulate as more rounds of MM switching take place.However, the resin switching method leads to a huge amount of unnecessary resin waste.In order to print a thin layer of new material, the current resin that fills the vat has to be completely switched.The level of resin waste can be greatly reduced by applying the resin switching method to the bottom-up system (figure 10(d)).Nevertheless, the contamination issue still exists if the residual resin is not removed.Comparing the four resin contamination scenarios introduced in figure 10, it is apparent that the bottomup DLP 3D printing system is more suitable to be extended to an MM 3D printer as less contamination would be introduced during the MM switching process.Compared with the resin switching enabled MM 3D printing (figures 10(c) and (d)), the vat switching enabled MM 3D printing is more materialefficient as it is not required to completely switch the resin in the vat.
As discussed in the above paragraph, the material contamination is the most critical challenge for achieving the MM 3D printing capability for DLP.How to efficiently and quickly remove the residual resin resulted from the MM switching step is the key to addressing the material contamination issue.In the past decade, for different DLP MM 3D printing systems, many efforts have been made to explore efficient approaches to remove the residual resin.For the vat switching enabled top-down DLP MM 3D printing system in figure 10(a), the major approach to remove the residual resin is to immerse the printing stage along with the printed structure into the vat of cleaning solution.For the resin switching enabled DLP MM 3D printing systems in figures 10(c) and (d), the residual resin issue can be resolved by flushing new resin to completely replace the current resin in the vat [91,92].Compared with the other three MM 3D printing systems, the vat switching enabled bottom-up DLP MM 3D printing system in figure 10(b) confronts less material contamination.As summarized in figure 11, the residual resin in such system can be removed through multiple approaches by adding various mechanisms.Figure 11(a) illustrates an approach that uses the cleaning solution to wash away the residual resin stuck to the printed structure, and an air dryer to evaporate the cleaning solution remaining on the printed part [93,94].However, this approach requires two additional steps to remove the residual resin (solution washing and air drying), which would greatly extend the time for MM 3D printing.Moreover, the cleaning solution may diffuse into the partially cured structure, which would weaken its mechanical property.Instead of using cleaning solution, the approach of using wiper to remove the residual resin was also proposed (figure 11(b)) [44].Compared with the approach of using cleaning solution, this approach only needs one step to remove the residual resin.Nevertheless, this wiping approach may not be applicable to print MM structure made of soft materials as the wiping process would either severely deform or even damage it.Moreover, after multiple rounds of MM printing, the wiper might be badly contaminated so that cannot be further used to remove the residual resin.To address the above issue, as shown in figure 11(c), researchers also proposed the approach of using air jetting to remove the residual resin [95].This approach enables MM 3D printing of two material structures with complex geometries and clear material transition interface.However, the highpressure air jetting is not suitable to remove the residual resin stuck to a soft material structure.In addition, the approach could only be used to achieve small-area MM 3D printing as removing the residual resin on the larger area part requires much higher pressure, which may distort or damage the printed part badly.In the above three approaches, the residual resin removal process necessitates direct contact with solid wiper or airflow onto the printed part, which constrains DLP MM 3D printing to small build sizes, limited available materials, and slow printing speed.To address these issues, researchers proposed the approach of using centrifugal force to remove the residual resin [46].As shown in figure 11(d), after lifting the printed structure from the yellow resin vat, the printing stage spins to generate centrifugal force that quickly removes the residual resin.Such an approach allows DLP MM 3D printing to fabricate heterogeneous 3D structures in large area (up to 180 mm × 130 mm) made of materials ranging from hydrogels to functional polymers, and even ceramics.Detailed introduction and discussion on the examples of the DLP MM 3D printers that use various approaches to remove residual resin will be discussed in the following section.

DLP MM 3D printing systems
In this section, we review a number of typical examples of the DLP MM 3D printing systems developed in the past decade.These systems use the material switching methods that are summarized in figure 7.In each example, we introduce the structural details of each system, the key steps during MM 3D printing, as well as the representative structures and even applications from each system.

MM stereolithography via vat switching
Before the emergence of DLP MM 3D printing technology, researchers have made a lot of efforts to realize multimaterial stereolithography (MMSL) [96][97][98][99][100].In 2011, as shown in figure 12(a), Choi et al designed and built a top-down MMSL system based on a commercial 3D Systems SL printer [99,100].The laser scanning system is retained from the 3D Systems printer, including the optical system, laser, the rim assembly with beam profilers, and associated scanning controllers.The MMSL system contains two parts: manufacturing center and control center.The manufacturing center consists of a new machine frame and all fabrication hardware, including four vats mounted on a rotary stage.Three of the four vats are filled with different resins, and the other one is filled with a cleaning solution.To avoid the possible interferences from previously printed layers, the MMSL system does not include a sweeping blade recoating system, but adopts a deepdip coating strategy with 3 mm pre-dip using diluted resins.The control center consists of the automatic resin height control system, the 3D Systems controller, and a new LabVIEW program that manages the entire MMSL process.Figure 12(b) depicts the general MMSL 3D printing process [96].The process starts with immersing the printing platform into the first resin vat, followed by a printing process to fabricate part with the first resin.Then, by vertical movement of platform and rotation of vat, the platform is immersed into the cleaning vat for removal of residual resin.Once the cleaning and drying processes are finished, the vat filled with the second resin is rotated under the platform to continue the subsequent printing process.Figure 12(c) demonstrates the building sequence of an MM rook fabricated by the MMSL system [99].The rook model is divided into three sub-models with three different materials.Each sub-model is individually sliced and then printed either on the printing platform or directly on the previously printed part.The material switching occurs three times during the whole MMSL process.Although the MMSL system can fabricate various 3D objects with different material combinations (figure 12(d)) [99], there are still some issues that need to be addressed, such as laser blocking and material contamination.The laser beam may be physically blocked by the previously printed part.During material switching, the resin remaining on the entire platform and printed part is difficult to be removed completely, especially for those structures with microchannels.These problems limit the design flexibility of MM structures.

DLP MM 3D printing via resin switching
Other than vat switching method, the resin switching method can also be used to realize MM 3D printing capability for topdown DLP system.In 2019, Han et al developed a DLP MM 3D printing system using a solvent-free cleaning process to realize rapid material switching [91].As shown in figure 13(a), the system adopts a dynamic fluidic control approach to switch materials in a fluidic cell that integrates material containers and a pump.Through a computer-controlled operation sequence of valves and pumps, the fluidic cell is filled with the material intended for printing.When switching material, a new material in the corresponding container is quickly pumped into the fluidic cell to replace the previous material.The building platform and printed part are vigorously flushed by the pressure-driven flow of new material.This material switching method can be applied to various liquid materials including resins, hydrogel solutions, and particle-containing suspensions.Figure 13(b) presents MM 3D-printed heterostructures at millimeter scale.Confocal fluorescent microscope images clearly display that these structures are composed of two different materials.In addition, the resin switching method can also be applied to the bottom-up DLP system to achieve MM 3D printing capability.In 2022, Wang et al reported a bottom-up composable-gradient DLP (bio)printing system integrated with a microfluidic mixer for the generation of desired composite (bio)inks in real time, as shown in figure 13(c) [92].The gradients of printed structures can be precisely controlled by adjusting the flow ratio of different (bio)inks.The microfluidic chaotic mixer and the shallow ink vat minimize the material waste when switching (bio)inks.Figure 13(d) shows the printed gradient objects including a vascular network with 4 colors, a 3D maple leaf with vertical gradients, and a cube consisted of 6 × 6 × 6 units with gradient colors.The two examples in figure 13 confirm that the resin switching method can be implemented to both top-down and bottom-up DLP systems to print MM 3D structures.However, the resin switching method leads to a substantial resin waste.In order to print a thin layer of new material, the current resin that fills the vat has to be completely (figure 13(a)) or partially switched (figure 13(c)).Since the resin waste is proportional to the size of the resin vat, the printed MM 3D structures demonstrated in figure 13 are limited to small size at millimeter scale.In addition, the use of dynamic fluidic channels to switch resins requires the viscosity of the material resin to be sufficiently low.Thus, the DLP MM 3D system with the resin switching method may not be able to print MM ceramic structures as the ceramic resins are more viscous than ordinary polymer resins [101].

DLP MM 3D printing via rotating wheel-based vat switching
Compared with the MM 3D printing systems introduced in figures 12 and 13, the bottom-up DLP 3D printing systems implemented with vat switching method can print larger area MM 3D structure and have less constraints on the viscosity of material resin.Therefore, this type of DLP MM 3D printing has been intensively explored in terms of vat delivery approaches, means of residual resin removal, as well as application explorations.As shown in figure 14(a), in 2013, Zhou et al developed a bottom-up DLP MM 3D printing system which adopts a two-stage cleaning strategy to avoid contamination during the vat switching [93].A soft brush is used to remove most of the liquid resin remaining on the printed part, and then the bottom of the printed part is immersed into a cleaning solvent assisted with an ultrasound cleaning to remove the residual resin thoroughly.After final cleaning, a fan is used to blow dry air to dry out the residual alcohol on the part.Besides the cleaning system, a new two-section system is developed to address the large separation force between the cured layer and the polydimethylsiloxane (PDMS) membrane in printing process.As illustrated in figure 14(b), a transparent PDMS membrane is attached to half of the bottom surface of the resin vat, dividing the vat into two sections.The patterned UV light is only exposed on the section with PDMS.By combining the horizontal movement of resin vat and the vertical movement of printing platform, the separation force is significantly reduced.Figure 14(c) presents MM 3D-printed objects including brush, tai chi pattern, and digital material.Although the printing resolution and efficiency need to be improved, this work validates the feasibility of vatswitching-based DLP MM 3D printing.To further reduce the complexity of the printing system and avoid the direct contact between printed parts and resins stored in the containers, Wang et al reported a droplet-delivery-based DLP MM 3D printing system which was used to fabricate MM mechanical metamaterials [44].As shown in figure 14(d), a rotating wheel is designed for delivering polymer resin droplets which are used to form the corresponding layers.In addition, the rotating wheel is also equipped with cleaning vats and wipers to remove the residual resin remaining on the printed structure.Different from the system in figure 14(a), the droplet-deliverybased system avoids the direct contact between printed parts and resins stored in the containers, which averts the contamination of the resins in the containers.Figure 14(e) presents the mechanical metamaterial printed with photocurable PEGDA solutions and copper nanoparticles reinforced PEGDA (Cu-PEGDA).Due to the different thermal expansion coefficients and stiffness of PEGDA and Cu-PEGDA, these MM lattices  [93].(b) Two-section system for reducing the separation force during printing process [93].(c) MM 3D-printed brush, tai chi pattern and digital material.Reproduced with permission from [93].Copyright © 2013, Emerald Group Publishing Limited.(d) Residual resin removed by washing and wiping in vat switching for bottom-up DLP MM 3D printing [44].(e) MM 3D-printed mechanical metamaterials with tunable negative thermal expansion.Reprinted figure with permission from [44], Copyright (2016) by the American Physical Society.exhibit negative thermal expansion through special designs of microstructures.

DLP MM 3D printing via translating plate-based vat switching
Other than using rotating wheel to deliver polymer resins, later studies found that using translational motion is a more efficient way to deliver polymer resins, and the translational delivery mechanism has been equipped to DLP MM 3D printing systems to fabricate various MM structures such as metamaterials [75,102], electronic devices [103], and soft robots [76].Xu et al developed a translating plate-based DLP MM 3D printing system equipped with a recoating fixture to print viscous multiple resins [102].As shown in figure 15(a), the recoating fixture employs the tape-casting technology to recoat highly viscous carbon fiber reinforced polymer (CFRP) composite efficiently.Initially, a small quantity of CFRP composite is extruded onto the oxygen-permeable window.Subsequently, a doctor blade is moved horizontally from left to right to uniformly spread the composite on the oxygen-permeable window for printing.Figure 15(b) presents a lightweight, stiff, high damping microlattice consisting of CFRP phase and soft phase, which was printed by such DLP MM 3D printing system.In addition to printing CFRP lattice structures, Hensleigh et al also reported a charge-programmed 3D printing approach to fabricate MM electronic devices through DLP MM 3D printing [103].Figure 15(c) presents the schematic of the MM 3D printing system with MM programable feedstock and resin wiper to fabricate MM 3D structures.Through controlling the surface charge between 3D-printed substrates and the materials to be deposited, the reported approach can realize volumetrically selective deposition of multiple functional materials such as nickel-phosphorus (Ni-P), copper (Cu), magnetite (Fe 3 O 4 ), zinc oxide (ZnO), and carbon nanotube.Figure 15(d) shows various 3D electronic devices fabricated by this approach, including a circuit on a 3D pyramid substrate, a 3D shape sensor with embedded electrodes, a piezoelectric lattice with selectively coated internal electrodes, and a tactile sensor with patterned electrodes.Moreover, Cui et al developed a charge-programmed MM 3D printing technology that is capable of assembling piezoactive, structural, and conductive phases into complex 3D microstructures [76].As shown in figure 15(e), negatively charged resin and highly loaded nanoparticle colloid are printed via a bottom-up DLP MM 3D printing system.After printing, the conductive metals Typical examples of DLP MM 3D printing systems using translating plate-based vat switching method.(a) DLP MM 3D printing system integrated with the tape-casting method [102].(b) 3D-printed lightweight, stiff, high damping microlattice consisting of CFRP and soft phase [102].Reprinted from [102], © 2020 Published by Elsevier B.V. (c) DLP MM 3D printing system with MM programable feedstock [103].(d) A circuit on a 3D pyramid substrate, a 3D shape sensor with embedded electrodes, a piezoelectric lattice with selectively coated internal electrodes, and a tactile sensor with patterned electrodes [103].Reproduced from [103], with permission from Springer Nature.(e) A bottom-up DLP MM 3D printing system for charge-programmed MM additive fabrication [76].(f) Schematic of the stimuli-responsive multimodal mobile microrobot [76].(g) Optical image of the fabricated mobile microrobot.From [76].Reprinted with permission from AAAS. can be selectively deposited on the charged resins to form a 3D microstructure with electrodes.Figure 15(f) shows the schematic diagram of stimuli-responsive multimodal mobile microrobot consisting of a piezoelectric metamaterial building block, embedded actuation element, self-sensing element, and contactless detection element.Figure 15(g) presents the optical image of the MM 3D-printed mobile microrobot.By allocating sections of piezoactive phases as the sensing element and utilizing its direct piezoelectric effects through 3D embedded electrodes, the robot attains proprioceptive sensibility capable of self-sensing its strain change and responding to external stimuli rapidly.
The DLP MM 3D printing systems introduced in figures 13-15 adopt the wiping approach to remove the residual resin on the printed part.However, the wiper itself would be badly contaminated after multiple rounds of MM printing, and cannot be further used to remove the residual resin.Therefore, instead of wiping approach, air flow is another more  [95].(e) Printed two-material lattice structures [95].Scale bars, 2 mm.Reproduced with permission from [95].Copyright © 2018, Mary Ann Liebert, Inc. (f) MMSL 3D printing system with a glass palette storing multiple resin droplets [94].(g) Cross shape and cube fabricated with multicolor resins [94].Scale bars, 200 µm.Reprinted with permission from [94] © The Optical Society.
efficient means to remove the residual resin and reduce the degree of material contamination of printed MM structures.As shown in figure 16(a), Mao et al proposed a bottom-up DLP MM 3D printing approach based on a new curing-on-demand printhead [74].The printhead adopts a strategy consisting of four sections: 'coating, curing, cleaning and post-curing' (C3P) (figure 16(b)).A novel coating mechanism based on surface tension of liquid resin is developed to coat the resin to the printed part at a thickness as shallow as a single layer.After the coating process, the printhead moves to the curing section, where the coated resin undergoes photocuring in accordance with the projected UV light pattern.Subsequently, the uncured liquid resin is removed through a vacuum cleaning method.The cleaning section connects a vacuum pump, which provides a negative pressure to suck the uncured resin on the printed part.To avoid material waste, the collected resin can be recycled into the material reservoir.After the vacuumcleaning process, there still exists a small amount of residual resin, which can be fully cured in the post-curing section.Photocuring residual resin will also lead to an additional ∼10% cured materials.Figure 16(c) shows the multi-color objects printed with two or three different resins.Different from the C3P MM 3D printing system, which uses the air flow generated from the vacuum pump to remove the residual resin, Kowsari et al developed an air jet-based DLP MM 3D printing system to produce high-resolution MM objects in a more efficient way (figure 16(d)) [95].The system uses a glass plate that moves translationally to facilitate the delivery of various material puddles, enabling rapid material exchange.Notably, it introduces an air jet-based cleaning step for the first time, which effectively minimizes resin waste and material contamination, and avoids the use of any cleaning solutions that may damage the printed parts.This system achieves a remarkable 58% increase in printing speed when producing complex MM microlattice structures (figure 16(e)), compared with previous studies that use cleaning solutions.Similarly, Maruyama et al reported a bottom-up MMSL 3D printing system that also uses an air jet to remove the residual resin [94].As shown in figure 16(f), the system utilizes a glass palette that can store multiple resin droplets, along with two tanks that hold the cleaning solvent for a two-step cleaning process.The second tank incorporates an air blowing unit to ensure the complete removal of the cleaning solvent.This device can fabricate MM 3D microstructures without material contamination and large amounts of resin waste.Figure 16(g) shows the printed cross shape and cube with multi-color resins.

DLP centrifugal MM 3D printing technology
The MM switching process in any above mentioned MM 3D printing system requires direct contact of solid wiper or fluidic flow onto the printed part to remove the residual resin, which in fact significantly constrains the MM 3D printing to small build size, limited available materials, slow printing speed, and low function integration.To address these limitations, Cheng et al reported an innovative DLP-based centrifugal multimaterial (CM) 3D printing technique, enabling the creation of large-volume heterogeneous 3D objects with programmable compositions, properties, and functionalities at the voxel scale [46].As shown in figure 17, by harnessing centrifugal force, the CM 3D printing system achieves non-contact high-efficiency MM switching, capable of producing heterogeneous 3D structures in a large area (up to 180 mm × 130 mm) using a wide range of materials including hydrogels, polymers, and even ceramics.Figure 17(a) illustrates the setup of bottom-up CM 3D printing system that adopts a rotating motor to spin the printing platform for the removal of residual resin stuck on the printed part when switching materials.Figure 17(b) depicts the process of printing an octet truss using two different materials (black and white resins).The procedure begins with Step I, where a slice of the black part is printed.Following this, in Step II, the printing platform is raised from the vat containing the black resin.In figure 17(c), it can be clearly observed that some black resin remains attached to the printed part.In Step III, the rotating motor spins the printing platform to remove the residual resin (figure 17(d)).Then, in Step IV, the vat containing white resin moves horizontally underneath the platform for printing of the white part.As shown in figure 17(e), the CM 3D printing system can directly fabricate a large-volume two-material octet truss structure with an overall size of 155 mm × 108 mm × 57 mm, where the black and white units are alternatively arranged in three dimensions.Figure 17(f) presents a four-material octet truss.The clear boundaries between different units shown in the zoomedin images demonstrate that the CM 3D printing system can achieve nearly zero material contamination during the MM switching process.Figure 17(g) shows the model predictions of centrifugal time for resins with different viscosities under different angular speeds, indicating that the centrifugal-based residual resin removal method is suitable for resins with a wide range of viscosity, and even high viscosity resins can be quickly removed by increasing the angular speed.Thus, the CM 3D printing system can print a variety of materials with different properties and functionalities.Figure 17(h) presents a few representative MM structures including a blood vessel system with two color hydrogels, a Kelvin foam structure with soft and hard polymers, a Miura-origami sheet with hard polymer panels connected by the SMP hinges, an octet truss consisting of an ionic conductive elastomer core within the nonconductive soft polymer lattice shell, and a Kelvin foam with two color ceramics.The CM 3D printing system demonstrates exceptional capabilities in fabricating digital materials, soft robots, and ceramic devices.Figure 17(i) presents the printed digital materials whose mechanical properties can be adjusted by manipulating the spatial arrangement of hard and soft voxels.The capability of fabricating digital materials can be applied to the 4D printing of a water-responsive hand.Figure 17(j) shows a printed SPA that features seamless integration of bending, pressure, and temperature sensors.A soft robotic gripper with multiple sensing capabilities can be obtained by assembling three SPAs.Figure 17(k) shows a printed ceramic bearing by using a polymer as support material to fill the empty space between the rollers and the inner/ outer ring during the CM 3D printing process.Sintering process can remove the polymer composition in the ceramic green body and support material to obtain a pure ceramic bearing, which could rotate freely without resistance at high temperatures.Although the CM 3D printing system can fabricate large-volume heterogeneous structures with a wide variety of materials, it compromises the printing resolution and does not fully present the high-precision advantage of DLP technology.Additionally, when printing soft or low-strength materials, the high-speed centrifugal process may damage the printed structure.Therefore, it is necessary to optimize the structural design and material distribution to ensure successful printing.

Perspectives
In the above sections, we present a comprehensive review of the recent advancements of DLP MM 3D printing technologies in terms of material switching methods, residual resin removal methods, working principles, printing steps, as well as the representative structures and applications.Numerous DLP MM 3D printing systems have been developed in the past decade, and used to fabricate various MM 3D structures for application in metamaterials, flexible electronics, biomedical devices and robots.However, the adoption of DLP MM 3D printing technologies to broader applications is hindered in the following four aspects: multiscale MM 3D printing capability, suitable functional materials, powerful tools to design and optimize MM 3D structures, and commercialized products for DLP MM 3D printing.

Multiscale MM 3D printing
The previous study shows that DLP 3D printing can fabricate 3D structure with high-resolution details (up to 0.6 µm) but limit overall size (a few millimeters) due to the tradeoff between image resolution and image field area [104].To overcome this limitation and achieve multiscale 3D printing capability, researchers have made tremendous efforts and proposed various feasible approaches such as large area projection microstereolithography approach that combines DLP with a coordinated optical scanning system [105], femtosecond projection two-photon lithography technique that combines two-photon lithography with digital mirror device [65], as well as integral lithography that combines DLP with rotational microlens array [106].However, these approaches only allow us to print multiscale 3D printing with one material.The multiscale MM 3D printing capability has not yet been achieved.To achieve such capability, further explorations are required to investigate whether we can directly combine a suitable MM switching method with one of the above multiscale 3D printing systems or need to develop a new specific MM switching method which could cooperate with one of the multiscale 3D printing systems.

Functional materials for MM 3D printing
The recently developed CM 3D printing system allows us to print a wide range of materials with distinct properties and functions including hydrogels, polymers with various stiffnesses, SMPs, ionic conductive elastomers, and even ceramics [46].However, due to high viscosity or low photo-reactivity, we can rarely use DLP to print some high-performance polymers such as poly ether ether ketone (PEEK), and polyimide (PI) [107], or functional polymers such as liquid crystal elastomers (LCE) [108], and poly (3,4-ethylenedioxythiophene):poly(styrenesulfonate) (PEDOT:PSS) [109,110].It is more challenging to print them with other materials on a DLP MM 3D printing system.Besides the chemical modifications, which make these polymers more photo-reactive, the hardware modification to a DLP 3D printer may also find solutions to print these polymers.

Design for MM 3D printing
The layer-by-layer processing manner endows 3D printing with unprecedented design flexibility, which is well compatible with the concept of topology optimization (TO), whose geometrical complexity often restricts its application using conventional manufacturing methods [111].On the other hand, TO also becomes a powerful tool to guide the design for 3D printing.Thus, TO and 3D printing mutually benefit each other.Yet, the conventional TO approaches are mainly used to design lightweight structures by progressively removing structurally inefficient areas in design domain.The rise of MM 3D printing greatly enriches the manufacturability of 3D printing and extends its application from lightweight structures to functional structures such as active composites by 4D printing [72,73,112], assembly of sensors [81,113,114], and soft actuators with complex deformation [115,116].However, the current TO approaches are incompetent to guide the design of those MM functional structures which usually involve multiphysics and material nonlinearity [117].Moreover, the discrete nature of voxel based MM 3D printing methods adds further difficulties to traditional TO approaches, which are more suitable to model a structure where the materials are continuously distributed [117].Recent advancements in machine learning (ML) offer new possibilities for the design of MM 3D printing [118].Further research efforts are needed to develop various ML techniques, such as evolutionary algorithms, convolutional neural networks, and recurrent neural networks for the design of functional structures fabricated by MM 3D printing.

Commercialized products for DLP MM 3D printing
Compared with various commercialized products for single material 3D printing, the number of companies that produce MM 3D printers is limited.The most successful commercial products of MM 3D printing are Connex or J series from Stratasys (Eden Prairie, MN, USA), which are able to print full color complex 3D objects through inkjet technology.For the extrusion-based 3D printing systems, the products with MM 3D printing are provided by a number of companies such as Makerbot (Brooklyn, NY, USA), Ultimaker (Utrecht, Netherlands), Bigrep (Berlin, Germany), and many other brands for FDM; Regenovo (Hangzhou, China), CELLINK (Gothenburg, Sweden) for DIW.DLPbased 3D printing technologies have been well commercialized by various companies such as Carbon (Redwood City, CA, USA), Prodways (Montigny-le-Bretonneux, France), and BMF (Shenzhen, China), to name a few.However, compared with the rapid development in feasible MM 3D printing technologies for DLP, the progress in commercializing them stagnates.The further advancement in DLP MM 3D printing technologies necessitates the commercialization of a few representative technologies as the practical needs from daily use and real applications help mature the technical details.

Figure 1 .
Figure 1.MM 3D printing capability implemented by various 3D printing technologies.(a)-(d) FDM MM 3D printing.(a) Thermoplastic filaments extruded from different heated nozzles.(b) A Hilbert cube with numerous overhangs (front) is obtained by removing the white sacrificial support material in an FDM 3D-printed part (back).Reproduced from [35], with permission from Springer Nature.(c) Folded and unfolded states of MM 3D-printed wing.From [36].Reprinted with permission from AAAS.(d) Miura origami.Scale bar, 20 mm.Reproduced from [37].CC BY 4.0.(e)-(h) DIW MM 3D printing.(e) Inks extruded from different syringe nozzles.(f) Elastomeric lattice printed with four materials.Scale bar, 2 mm [38].John Wiley & Sons.© 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim.(g) Bioprinting of human organ-like constructs from multiple bioinks [39].John Wiley & Sons.© 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim.(h) Voxelated matter printed with four materials.Reproduced from [40], with permission from Springer Nature.(i)-(l) Inkjet MM 3D printing.(i) An inkjet process via multiple ink-jetting nozzles to deposit different photopolymerizable resins.White matter tractography data of a human brain.In this visualization, the fibers represent high-resolution bundles of axons connecting different regions of the brain (j).Scale bar, 10 mm.From [41].Reprinted with permission from AAAS.(k) Inkjet-printed flower consisting of multiple layers of petals that bloom into a configuration with different curvatures upon heating.Scale bar, 30 mm.From [42].Reprinted with permission from AAAS.(l) A landscape painting printed by inkjet MM 3D printer.Reproduced with permission from [43].Used with permission of Association for Computing Machinery, permission conveyed through Copyright Clearance Center, Inc. (m)-(p) DLP MM 3D printing.(m) Illustration of a bottom-up DLP MM 3D printing process.(n) DLP MM 3D-printed lightweight lattice with tunable negative thermal expansion.Scale bar, 2 mm.Reprinted figure with permission from [44], Copyright (2016) by the American Physical Society.(o) Rigid polymer reinforced hydrogel composite lattice structure with gradient stiffness.Scale bar, 5 mm.From [45].Reprinted with permission from AAAS.(p) A large volume two-material octet truss via centrifugal DLP MM 3D printing.Scale bar, 50 mm.Reproduced from [46] CC BY 4.0.

Figure 2 .
Figure 2. Comparison on the MM 3D printing capability of the above-mentioned four 3D printing technologies.

Figure 5 .
Figure 5. Resin photopolymerization and light intensity distribution when UV light travels in the resin.(a) The degree of photopolymerization attenuates along the UV light direction.(b) The light intensity attenuates along the UV light direction.

Figure 6 .
Figure 6.Mechanism of UV light projection for DLP 3D printing.(a) Basic optical components and light path.(b) Illustration of DMD chip.(c) Gaussian distribution of light intensity for a single pixel.(d) A single Gaussian beam illustrated in black.The purple dashed lines mark the nominal boundaries of a single 10 µm pixel.(e) Ten Gaussian beams are illustrated in black dashed lines, and their sum is depicted in blue solid line.The purple dashed lines mark the nominal boundaries of ten 10 µm pixels.(f) Ten Gaussian beams where the left five beams have 100% of maximum light intensity for each pixel (5 mW•cm −2 ), while the right five beams have 50% of maximum light intensity for each pixel (2.5 mW•cm −2 ).The individual beams are illustrated with black dashed lines, and their sum is shown in blue.The purple dashed lines mark the nominal boundaries of ten 10 µm pixels.

Figure 7 .
Figure 7. Material switching methods for DLP MM 3D printing.(a) Top-down DLP MM 3D printing via vat switching.(b) Bottom-up DLP MM 3D printing via vat switching.(c) Top-down DLP MM 3D printing via resin switching.(d) Bottom-up DLP MM 3D printing via resin switching.

Figure 8 .
Figure 8.Comparison of layer-slicing manners, printing sequences, switching times, and printed objects for Eiffel Towers with different material distributions.(a) A pure-white Eiffel Tower.(b) A two-color Eiffel Tower where the bottom half is white and the top half is black.(c) A two-color Eiffel Tower where the front and rear quarters are black and the left and right quarters are white.(d) A two-color Eiffel Tower where the black and white resins are randomly distributed.Scale bars, 10 mm.

Figure 9 .
Figure 9. Material bonding mechanism for MM DLP 3D printing.(a)-(d) Illustration of detailed process to print a layer consisting of two materials.(e) Chemicals in the photocurable resins.(f) Chemical structures of cross-linked blue resin, (g) cross-linked yellow resin, and (h) their interface, respectively.(i) Uniaxial tensile test on a hydrogel-elastomer heterogeneous dog-bone specimen.Reproduced from [82] with permission from the Royal Society of Chemistry.(j) Snapshots of 3D printed rigid polymer-reinforced hydrogel composite before uniaxial tensile test (left) and after stretched by 175% (right).(k) SMP-hydrogel stent: as printed (up), after compacted (down).Scale bars, 5 mm.(l) Soft pneumatic actuator: illustration (up), as printed (down).From [45].Reprinted with permission from AAAS.(m) A printed heterogeneous dog-bone sample consisting of elastomer and ceramic green body.Scale bar, 5 mm.(n) and (o) Uniaxial tensile (n) and 90 • peeling tests (o) on the ceramic-elastomer heterogeneous specimens.(p) A printed ceramic spider with an overhanging body.Scale bar, 5 mm.Reproduced from [46] CC BY 4.0.

Figure 10 .
Figure 10.Issues about material contamination during material switching.(a)-(d) Contaminations caused by (a) vat switching in top-down, (b) vat switching in bottom-up, (c) resin switching in top-down, and (d) resin switching in bottom-up DLP MM 3D printing systems, respectively.

Figure 13 .
Figure 13.Typical examples of DLP MM 3D printing systems using resin switching method.(a) Schematic illustration of a top-down DLP MM 3D printing system using dynamic fluidic control to switch materials [91].(b) Optical and fluorescent microscope images of MM 3D-printed microstructures [91].Reprinted from [91], © 2019 Elsevier B.V. All rights reserved.(c) Illustration of a bottom-up composable-gradient DLP (bio) printing system integrated with a microfluidic mixer [92].(d) Models and MM 3D-printed gradient objects [92] John Wiley & Sons.© 2021 Wiley-VCH GmbH.

Figure 14 .
Figure 14.Typical examples of DLP MM 3D printing systems using rotating wheel-based vat switching method.(a) Residual resin removed by washing and drying in vat switching for bottom-up DLP MM 3D printing[93].(b) Two-section system for reducing the separation force during printing process[93].(c) MM 3D-printed brush, tai chi pattern and digital material.Reproduced with permission from[93].Copyright © 2013, Emerald Group Publishing Limited.(d) Residual resin removed by washing and wiping in vat switching for bottom-up DLP MM 3D printing[44].(e) MM 3D-printed mechanical metamaterials with tunable negative thermal expansion.Reprinted figure with permission from[44], Copyright (2016) by the American Physical Society.

Figure 15 .
Figure 15.Typical examples of DLP MM 3D printing systems using translating plate-based vat switching method.(a) DLP MM 3D printing system integrated with the tape-casting method[102].(b) 3D-printed lightweight, stiff, high damping microlattice consisting of CFRP and soft phase[102].Reprinted from[102], © 2020 Published by Elsevier B.V. (c) DLP MM 3D printing system with MM programable feedstock[103].(d) A circuit on a 3D pyramid substrate, a 3D shape sensor with embedded electrodes, a piezoelectric lattice with selectively coated internal electrodes, and a tactile sensor with patterned electrodes[103].Reproduced from[103], with permission from Springer Nature.(e) A bottom-up DLP MM 3D printing system for charge-programmed MM additive fabrication[76].(f) Schematic of the stimuli-responsive multimodal mobile microrobot[76].(g) Optical image of the fabricated mobile microrobot.From[76].Reprinted with permission from AAAS.

Figure 17 .
Figure 17.CM 3D printing system and its applications [46].(a) Illustration of the CM 3D printing system.(b) CM 3D printing steps.(c) Residual resin stuck on the printed part after the platform is lifted from the resin vat.(d) Residual resin removed by centrifugal force.(e) A large-volume octet truss made of two materials.(f) A four-material octet truss.Scale bar, 5 mm.(g) Model predictions of centrifugal time for resin with different viscosities under different angular speeds.Black circles represent experimental data.(h) A variety of materials with different properties and functionalities suitable for CM 3D printing.SM: shape memory.ICE: ionic conductive elastomer.Scale bars, 10 mm.(i)-(k) CM 3D-printed structures: (i) digital materials with programmable compositions and properties, (j) soft actuator with multiple sensors, and (k) ceramic bearing using a polymer as support material.Scale bars in (i), 10 mm.Scale bar in (k), 5 mm.Reproduced from [46] CC BY 4.0.