Direct Fused Deposition Modeling 4D Printing and Programming of Thermoresponsive Shape Memory Polymers with Autonomous 2D-to-3D Shape Transformations

polymers into various hinges of a structure


Introduction
In recent years, additive manufacturing technology, commonly known as 3D printing, [1][2][3][4][5][6] has made significant strides, enabling the digital design and manufacturing of intricate structures with high spatial resolution and microstructural heterogeneity.Alongside these developments, a new frontier has emerged in 3D printing known as 4D printing, which incorporates the element of time or shape change over time.The term "4D printing" was first coined by Tibbits [7] in 2013, defining it as a combination of 3D printing and the ability of printed objects to change shape over time.Today, researchers describe 4D printing as the fusion of 3D printing technology with smart materials that can undergo controlled changes in response to external stimuli. [8,9][12] Among the numerous types of smart materials, thermoresponsive shape memory polymers have garnered significant attention within the realm of 4D printing.Ge et al. [13,14] utilized an Object Connex multimaterial printer to fabricate printed active composites by embedding active glassy polymer fibers into an elastomer matrix.By subjecting the printed objects to a thermomechanical programming process, which involves inducing prestrain, the structures can exhibit shape conformation when heated to high temperatures.Additionally, the proposed printed active composites can be applied to origami structures, serving as hinges. [14]This method has been further explored by Wu et al., [15] from the same research group, who developed multiple shapechange structures comprising two different glassy polymers as fibers within an elastomer matrix.Consequently, the structure undergoes two-state shape changes as the temperature reaches the respective glass transition temperature of each glassy polymer.Mao et al. [16] incorporated different shape memory polymers into various hinges of a structure, achieving sequential shape transformations triggered by different temperatures.Bodaghi et al. [17] created actuator units, stents, and metamaterial DOI: 10.1002/adem.202300334Herein, direct 4D printing of thermoresponsive shape memory polymers (SMPs) by the fused deposition modeling (FDM) method that enables programing of 2D objects during printing for autonomous 2D-to-3D shape transformations via simply heating is focused on.The programming process during printing is investigated through designs and experiments.The capability of programming SMPs during printing is illustrated by prestrain and bending capabilities, which are highly related to printing settings, such as nozzle temperature, print speed, layer height, infill patterns, and ratio of active parts in a bilayer structure.A nearly linear relationship for prestrain and bending parameters is experimentally revealed for different printing factors.Quantitative results are presented to be used as a guidance for designing complex 3D structures via 4D printing of 2D structures.Helix structure, twisting structure, DNA-like structures, and functional gripper are designed to demonstrate the potential of direct FDM 4D printing for creating complex 3D structures from simple 2D structures with advantages over traditional manufacturing methods.It is shown that, by removing the need for a layer-by-layer stacking process to achieve a complex 3D shape, FDM can promote sustainability via 4D printing of autonomous 2D-to-3D shape transformer structures with lower materials, time, energy, and longer service life.lattices that, following post thermomechanical programming, could expand and subsequently return to their original shapes upon heating.However, it is worth noting that the aforementioned 4D printing methods utilizing thermoresponsive shape memory polymers with multimaterial printers tend to be costly due to the high expenses associated with the printers and materials involved.
Thermoresponsive shape memory polymers have emerged as the most commonly used materials in 4D printing.However, many existing 4D printing methods utilizing thermoresponsive shape memory polymers require complex thermomechanical programming processes.These processes often necessitate special fixtures and are challenging to implement for postprinting programming in various engineering applications.[20] Direct 4D printing refers to the ability of a 3D-printed object to undergo shape changes directly through heat actuation, without the need for complex thermomechanical programming processes.In our method, we incorporate the programming process into the 3D printing process itself.Additionally, compared to other 3D printing techniques, FDM 3D printing offers a low-cost solution in terms of both materials and machinery.In recent years, researchers have provided insights into programming shape memory polymers during the printing process.For instance, Zhang et al. [21,22] discovered that FDM 3D-printed polymer materials exhibit heat-shrinkage properties that can be utilized in shape-changeable lattice structures.Manen et al. [23,24] investigated shrinkage differences in different printing patterns and used heat transfer delays resulting from varying porosity and thickness to demonstrate sequential shape changes.Yao's group [25][26][27][28] presented numerous functional designs, such as chairs, helmets, flowers, and intricate artworks, that can undergo shape changes from simple line-type and mesh structures.Moreover, 4D printing has demonstrated its potential in various practical applications.One prominent area is the construction of self-assembling structures, including boxes, flowers, and clamps. [13,29,30]Additionally, 4D printing technology has been employed in the development of smart structures for energy absorption. [31,32]Furthermore, 4D printing holds promise in the medical field, where it can be utilized for applications such as drug delivery devices [33] and adaptive stents. [34]These applications benefit from the ability to transport devices in a compact, initial shape and trigger shape changes remotely.FDM printers are the most affordable additive manufacturing options available.During the FDM 3D printing process, the printer nozzle is heated to melt the filament, which is then extruded onto the building platform or adjacent layers after reaching the material's glass transition temperature.As the extrusion head moves horizontally, the extruded material is often partially stretched.Therefore, we can leverage FDM to induce prestrain in the direct 4D printing process.Additionally, we can easily control the printing path and parameters during our designs.While some studies have explored direct FDM 4D printing, such as those by Zou et al. [35] and Wang et al., [36] their designs consisted of only two layers with thicknesses below 1 mm, limiting their practical applications.Consequently, there is a need for comprehensive studies providing detailed information on achieving an autonomous 4D printing method with an FDM 3D printer.Moreover, there is a lack of examples and guidance for fabricating complex structures, such as helixes and twisting structures, using the direct 4D printing method.Therefore, it is both timely and significant to develop a systematic understanding of building a direct FDM 4D printing method capable of autonomous 2D-to-3D shape transformations.Such knowledge would make a valuable contribution to sustainability.
This article introduces a novel direct FDM 4D printing method, providing a comprehensive analysis of how to achieve in-plane and out-of-plane shape changes.The programming capabilities of direct FDM 4D printing, including prestrain and bending capabilities, are thoroughly investigated through a combination of design and experimentation.The influence of key printing parameters, such as temperature, speed, and layer height, as well as the printing pattern and ratio of active to passive layers, is quantitatively examined to offer valuable insights for data-driven design in direct FDM 4D printing.The research findings demonstrate that the FDM 4D printing approach is capable of fabricating complex 3D structures, exemplified by the successful creation of helix structures, twisting structures, DNA-like structures, and functional grippers through autonomous shape transformations.The presented results and the methodology for FDM 4D printing are expected to play a crucial role in achieving 2D-to-3D shape transformations and facilitating the fabrication of intricate structures with notable benefits such as reduced material usage, shorter production time, lower energy consumption, enhanced structural integrity for extended service life, and overall promotion of sustainability.

Methods of Direct FDM 4D Printing
Figure 1a demonstrates the thermomechanical programming procedure.In this process, the printed object in State-a undergoes heating, programming, and cooling to achieve a temporary new shape in State-d.The subsequent shape transformation in 4D printing occurs from the object in State-d back to the original shape in State-a through reheating and cooling cycles, following the shape memory cycle of thermoresponsive shape memory polymers.However, programming the printed structures after the printing stage proves to be challenging in many engineering applications, often requiring specialized fixtures.To address this challenge, we utilize FDM, the most widely adopted consumer 3D printing technology, to program thermoresponsive shape memory polymers during the material deposition process.Figure 1b illustrates the similarity between the FDM printing process and the thermomechanical programming procedure.Initially, the nozzle is heated to melt the filament, transitioning from State-a to State-b.Subsequently, the material is heated above its glass transition temperature and extruded onto the build platform or existing layers.As the extrusion head moves horizontally, the extruded material experiences partial stretching, corresponding to the transition from State-b to State-c.The thin printed layer of the polymer bonds with the platform or underlying layers, undergoes cooling, and solidifies, transitioning from State-c to State-d.As a result, pre-strain is induced into the printed objects.Finally, the printed objects can directly undergo shape changes after printing when exposed to a heated environment, returning to State-a.Considering that thermomechanical programming relies on temperature and stretching, [37] this research work investigates the influence of various printing factors on the degree of pre-strain.These factors include printing temperature, printing speed, and printing layer height.Additionally, the design of the printing pattern of structures is found to impact both prestrain and bending capabilities.
Therefore, the direct FDM 4D printing with thermoresponsive shape memory polymer filaments is proposed to enable printed objects perform in-plane and out-of-plane shape change directly after printing.The in-plane capability is realized by inducing even prestrain into objects as shown in Figure 2a,b.The printed object shrinks in the longitudinal direction and expands in the transverse direction by simply heating and cooling via free-strain recovery.The out-of-plane capability could be induced by a bending moment, which is realized by creating a bilayer structure with mismatched prestrain as shown in Figure 2c,d.Through the heating and cooling process, the active red part would like to shrink and is constrained by passive gray part, creating an out-of-plane shape change.
As a result of the prestrain stored in printed beam structures, shrinkage occurs in the longitudinal direction, while expansion takes place in the other two directions.
As shown in Figure 3a, the prestrain is defined by where L p and L f denote length after printing and heating respectively.Expansion in height and width is defined as where h p =w p and h f =w f indicate height/width after printing and heating respectively.The out-of-plane bending is the basis of shape transformation of direct FDM 4D printing.In this case, we use a bending angle per unit length of beam structures to quantify the bending capability as where α is bending angle of beam structures and l is required length that is also the original length after printing of beam structures.
The length of the samples can be directly measured using a Vernier scale with a resolution of 0.01 mm and an accuracy of 0.02 mm.Additionally, the angle of the samples can be measured using photo processing software (Tracker).To determine angle α, we identify the tangent lines AC and BC in Figure 3b to obtain angle β.

Investigation on Fundamental Shape Changes
Given the process of inducing prestrain during printing, it is reasonable to assume that several factors can influence the level of prestrain achieved in the printed objects.As the prestrain induction process involves thermomechanical programming, it follows  that temperature plays a significant role in determining the extent of prestrain in the printed objects.Furthermore, the movement of the nozzle during printing leads to stretching of the shape memory polymer, making printing speed a crucial factor to consider.In considering the interplay between the stretch of the shape memory polymer and the 3D printing process, it can be hypothesized that the printing layer height for each layer also impacts the degree of polymer stretching.Additionally, different printing patterns may yield distinct directions of stretch for the shape memory polymer, potentially resulting in varying levels of prestrain in the printed objects.
In this section, the fabrication process utilizes a 3DGence FDM 3D printer.Mechanical objects are printed using polyurethane-based SMP filaments (provided by SMP Technologies Inc.) with a diameter of 1.75 mm.These filaments possess a glass transition temperature of 60 °C.The build platform temperature and chamber temperature are both maintained at 24 °C.The printer is equipped with a 0.2 mm diameter extrusion nozzle.To evaluate prestrain levels, beam structures are designed and printed accordingly, as demonstrated in Figure 4.The printed samples of the beam structures have dimension 40 Â 4 Â 4 mm:.They are printed using a straight-line pattern in the longitudinal direction, as illustrated in Figure 4b.For each parameter setting, three samples are printed and subsequently tested in hot water at a prescribed temperature of 85 °C for a duration of 10 min.This duration ensures sufficient time for free-strain recovery and triggers the shape transformation.The specific printing parameter settings used in each experiment are listed in Table 1.
Figure 5a clearly illustrates a noticeable trend indicating that the prestrain induced during the printing process decreases as the printing temperature increases.The graph shows that higher printing temperatures result in lower prestrain values, with a significant difference observed between 210 and 260 °C.This finding highlights the substantial impact of temperature on prestrain induction.Thermally induced shape memory effect (SME) is closely tied to temperature.The shape recovery process involves releasing prestrain by reheating the printed object to a high temperature during free-strain recovery. [38]Higher printing temperatures provide more opportunities for the partial curing of prestrain during the printing process, leading to a decrease in prestrain.Thus, an inverse relationship between prestrain and printing nozzle temperature is observed.Furthermore, the expansion observed in both height and width, as depicted in Figure 5b, corresponds well with the changes in prestrain across different printing temperatures.It is worth noting that filament melting is inadequate at temperatures below 210 °C.Consequently, the printing temperature must not be set lower than 210 °C to ensure the desired quality of printed samples. [39,40]igure 6a depicts an increasing trend in prestrain as the printing speed ranges from 20 to 40 mm s À1 .This trend can be attributed to the fact that the thermal mechanical programming of shape memory polymers (SMP) relies on stretch deformation at high temperatures.Therefore, the printing speed has an impact on the stretch deformation of the extruded-melt SMP.In a given printing length for a prescribed design, a higher printing speed corresponds to a shorter printing time within that length.As a result, less material is extruded and stretched to cover the prescribed length, leading to a higher stretch rate of the material and, consequently, a higher prestrain.However, when the printing speed surpasses 40 mm s À1 , the prestrain tends to stabilize and even exhibit a slight decrease from 40 to 50 mm s À1 .This can be attributed to the diminishing quality of layer-to-layer connections as the printing speed exceeds 50 mm s À1 .The deteriorating connection quality can result in a lower prestrain, which offsets the positive effect of increasing speed.Consequently, achieving functionally good printing quality becomes challenging when the printing speed exceeds 70 mm s À1 .The expansion observed in height and width in Figure 6b aligns well with the changes in prestrain across different printing speeds.These findings further support the relationship between printing speed and prestrain induction.
The choice of printing layer height is constrained by the precision capabilities of the printer itself.The diameter of the extrusion nozzle and the accuracy of the step motor impose limitations on the viable printing layer height.In Figure 7a, it can be observed that the prestrain decreases as the printing layer height increases.This decrease in prestrain is influenced by two key factors associated with smaller printing layer heights.First, when the printing layer height is smaller, a lesser amount of material is extruded for each layer within a prescribed length.Consequently, a higher stretch rate is induced in the material, resulting in a    higher prestrain compared to larger printing layer heights for the same printing length.Second, smaller printing layer heights tend to yield better connection quality between the printed layers.Conversely, attempting to print with layer heights larger than 0.25 mm on the present printer leads to subpar function and surface quality of the printed samples due to poor connection quality.Additionally, layer heights smaller than 0.05 mm cannot be achieved due to the precision limitations of the printer employed in this study.Furthermore, Figure 7b demonstrates a consistent correlation between expansion in the height and width directions and the increase in layer height.These findings further support the relationship between printing layer height and the resulting prestrain and expansion characteristics.
Furthermore, the movement of the printer's nozzle during the printing process creates stretching of the shape memory polymers.It is worth considering that different printing patterns can generate varying directions of stretch in the shape memory polymer, leading to different levels of prestrain in the printed objects.To investigate this, we conducted tests using eight different printing patterns available in open-source slicing software, as illustrated in Figure 8a.The beam structures used for testing had dimensions of 40 Â 4 Â 4 mm: The printing parameters were set as follows: nozzle temperature of 220 °C, printing speed of 40 mm s À1 , and a printing layer height of 0.1 mm.For each printing pattern, three samples were printed and subjected to testing in hot water at a prescribed temperature of 85 °C for 10 min to allow for free-strain recovery and trigger shape transformations.The strain was measured in the longitudinal direction.The results, shown in Figure 8b, clearly indicate that certain printing patterns, such as P1 and P6 (single-direction-line patterns), were able to induce high levels of prestrain.On the other hand, P2, which was printed with a straight-line pattern in the transverse direction, exhibited the lowest prestrain.These findings demonstrate the influence of different printing patterns on the resulting prestrain, thereby validating our previous assumption.It is well known that prestrain is induced through the stretch deformation of the shape memory polymer.In the case of P1 and P6, the single-direction-line patterns allow for optimal stretching in the length direction, resulting in higher prestrain.Conversely, P2, with its transverse straight-line pattern, exhibits the least amount of stretch.The other printing patterns displayed varying degrees of stretch in the length direction, which were attenuated by the presence of crossline printing patterns.In summary, the choice of printing pattern can have a significant impact  on the prestrain achieved in the printed objects, as it directly influences the direction and extent of stretch deformation in the shape memory polymer.
As mentioned earlier, the creation of a bilayer structure can facilitate out-of-plane bending by introducing a mismatched prestrain.We propose investigating three types of bilayer structures, as depicted in Figure 9a,c,e.The experimental settings are chosen to demonstrate significant bending capabilities.The first type (a) features a bilayer structure with two different printing patterns.The upper layers are printed with a straight-line pattern in the longitudinal direction, while the lower layers employ a straight-line pattern in the transverse direction.This arrangement results in crossed stretching directions, which are expected to produce the most pronounced bending capability.The second type (c) consists of two printing layer heights.The upper layers are printed with a layer height of 0.05 mm, while the lower layers have a layer height of 0.2 mm.The design with a 0.05 mm layer height induces the highest prestrain, while the 0.2 mm layer height promotes better connection quality and lower prestrain compared to a 0.25 mm layer height.This combination is anticipated to exhibit favorable bending capability.Finally, the third type (e) comprises two different printing speeds.The upper part is printed at a speed of 40 mm s À1 , while the bottom part is printed at a speed of 20 mm s À1 .The design with varying printing speeds of 40 and 20 mm s À1 demonstrates a noticeable difference in prestrain along with good printing quality, making it an ideal demonstration for bilayer structures.By investigating these three types of bilayer structures, we aim to explore their bending capabilities and evaluate the effects of different printing patterns, layer heights, and printing speeds on the resulting prestrain and connection quality.
The beam structure samples are fabricated using an FDM printer with dimensions.The printing parameters used for each type of bilayer structure are outlined in Table 2. To assess the bending capabilities, three samples are printed for each type and subjected to a hot water bath at a specified temperature of 85 °C for a duration of 10 min.This duration ensures sufficient time for free-strain recovery, enabling the triggering of shape transformation.Finally, photos of the tested samples are captured for image processing, facilitating the calculation of the bending capability, as illustrated in Figure 9b,d,f.
The distinct variations in bending capability are evident from Figure 9b,d,f,g.The bilayer structures employing two different printing patterns exhibit significantly higher bending angles, indicating a greater bending capability.Referring to the results from previous experiments, we observed differences in prestrain values of 1.35 for P1 and P2, 0.60 for 0.05 and 0.2 mm layer height, and 0.46 for 40 and 20 mm s À1 printing speed.These discrepancies primarily account for the differences in bending capability.Therefore, using different combinations of printing  patterns, layer heights, and printing speeds, it is possible to establish a database of structures with varying bending capabilities to create bilayer structures.
In bilayer structures, we create a mismatch of prestrain by combinations of settings for producing different prestrains for two parts.We can treat one layer with higher prestrain as active part and the other with lower prestrain as passive part.Therefore, we can change the portion of two parts in bilayer structures.We take the Type 1(two patterns) bilayer structure as an example.The dimension of sample is 40 Â 4 Â 4 mm: We fix the printing layer height as 0.1 mm, printing speed as 40 mm s À1 , printing nozzle temperature as 220 °C.In such a case, the beam structure is printed with 40 layers.The upper 20 layers are printed with a straight-line pattern in the longitudinal direction, and lower 20 layers are printed with a straight-line pattern in the transverse direction.Besides, we design extra four beam structures with proportion of the upper active part over the lower passive part, which are 36:4, 32:8, 28:12, and 24:16.Then, three samples printed with each ratio are tested in the hot water with prescribed temperature 85 °C for 10 min, which promise adequate time for free-strain recovery, to trigger the shape transformation.Finally, photos of tested samples are taken for image process to calculate the bending capability as shown in Figure 10a.
To explore the impact of varying proportions of active and passive parts in bilayer structures, we will focus on the Type 1 bilayer structure (two patterns) as an example.The sample dimensions are 40 Â 4 Â 4 mm: The printing parameters are as follows: printing layer height of 0.1 mm, printing speed of 40 mm s À1 , and printing nozzle temperature of 220 °C.In this case, the beam structure consists of 40 layers, with the upper 20 layers printed in a straight-line pattern in the longitudinal direction and the lower 20 layers printed in a straight-line pattern in the transverse direction.Additionally, we design four additional beam structures with different ratios of the upper active part to the lower passive part, namely 36:4, 32:8, 28:12, and 24:16.For each ratio, three samples are printed and tested in hot water at a prescribed temperature of 85 °C for 10 min, allowing sufficient time for free-strain recovery and triggering the shape transformation.Finally, photos of the tested samples are captured for image processing to calculate the bending capability, as illustrated in Figure 10a.
Figure 10b demonstrates a nearly linear relationship between the bending capability and the ratio of upper layers to lower layers in the bilayer structures.The bending capability increases as the ratio approaches 1:1 (20:20), indicating that a higher proportion of the active part leads to a larger radius and smaller bending capability.This can be attributed to the fact that a larger amount of active parts tends to contract in order to recover prestrain, while fewer passive parts result in fewer constraints, which are favorable for achieving larger bending angles.Notably, the bilayer structures constructed with an equal proportion of active and passive parts (1:1 or 20:20) exhibit the smallest radius and the largest bending angle, indicating the optimal bending capability in this configuration.

Complex Structures Design
The natural world provides us with examples of macroscopic helical structures, one of which can be observed in the seedpods of peas.When these seedpods dry out, they undergo a humiditydriven deformation, resulting in the formation of helical shapes.This deformation is facilitated by the arrangement of organic fibers within the seed pods, as depicted in Figure 11a.
Inspired by this natural phenomenon, we can incorporate the concept of helical structures into our bilayer structure design.Utilizing different printing patterns, specifically two distinct straight-line patterns, we can create a bilayer structure capable of triggering out-of-plane bending.Interestingly, the arrangement of these straight-line patterns closely resembles that of the organic fibers found in the seedpods.Taking this inspiration into account, we propose a design approach that enables the formation of helical structures by strategically arranging the printing patterns within the bilayer structures.
In bilayer structures, we choose straight-line patterns which are likely made by fibers.The bottom layers are fixed in straight-line pattern with 90°angle as passive segments as demonstrated in Figure 11b.The upper layers are printed with varied angles from 15°to 75°as active segments.The printing settings are set as 220 °C printing temperature, 40 mm s À1 speed, and 0.1 mm printing layer height.The bilayer beam structures are printed with dimension of 80 Â 4 Â 2 mm ðL Â W Â HÞ. Figure 11c illustrates the deformed shapes after being triggered in the hot water with prescribed temperature 85 °C.A series of helical structures are formed with different pitches and diameters.The helical structure with 15°has the largest diameter and shortest pitch.The pitch of helical structure increases, and the diameter decreases when angles of fibers are elevated.That is because structures with small angles owe a higher bending moment produced by a relatively higher bending force and smaller bending stiffness.Therefore, when the angles of upper layers reach 60°, the bending moment is too small to form helical structures with a very high bending stiffness.
In our bilayer structures, we deliberately choose straight-line patterns to mimic the arrangement of fibers found in natural helical structures.The bottom layers are printed with a fixed straight-line pattern, forming passive segments with an angle of 90°(illustrated in Figure 11b).On the other hand, the upper layers consist of active segments that are printed with varying angles ranging from 15°to 75°.The printing parameters for these structures include a printing temperature of 220 °C, a printing speed of 40 mm s À1 , and a printing layer height of 0.1 mm.The dimensions of the printed bilayer beam structures are 80 Â 4 Â 2 mm ðL Â W Â HÞ.Upon immersing these printed structures in hot water at a prescribed temperature of 85 °C, they undergo shape transformations.As shown in Figure 11c, a series of helical structures are formed, each with a different pitch and diameter.Notably, the helical structure with 15°exhibits the largest diameter and the shortest pitch.As the angles of the upper layers increase, the pitch of the helical structures also increases, while the diameter decreases.This phenomenon can be explained by the bending moment generated by the structures.Structures with smaller angles experience a higher bending moment due to a relatively higher bending force and smaller bending stiffness.Consequently, when the angles of the upper layers reach 60°, the bending moment becomes too small to form helical structures with a very high bending stiffness.
By applying a similar design strategy as that of helical structures, we introduce bilayer structures with straightline printing patterns of opposite angles, as depicted in Figure 12a.The dimensions of the printed beam structures are 80 Â 8 Â 2 mm ðL Â W Â HÞ and the printing parameters remain the same as those used for the helical structures.The outcomes of these varied-angle structures are shown in Figure 12b after undergoing heat treatment in hot water.The observed results reveal that smaller angles have a higher probability of inducing shrinkage, as long fibers are printed, resulting in fewer twisting segments and a larger twisting diameter.As the angles increase, constraints for shrinkage begin to emerge due to the shorter printed fibers and higher bending stiffness of the two layers.Consequently, more twisting segments are formed, resulting in a smaller twisting diameter.However, when the angle reaches 75°the printed fibers become too short to generate sufficient bending moment for both the upper and lower layers to form twisting structures.This limitation leads to a higher bending stiffness, ultimately preventing the formation of twisting structures.Overall, the opposite-angle bilayer structures exhibit varied twisting capabilities, with the twisting diameter and the number of twisting segments dependent on the angle of the printing patterns.
DNA is a well-known double-helix structure as shown in Figure 13a.Due to the complexity of the double-helix structure, it is difficult to print such structures with traditional printing methods.The hanging parts require support structures during printing, and layer-by-layer printing manner would weaken the mechanical strength in certain directions.As we have made a helix structure in the previous section, we wish to fabricate a double-helix structure by direct FDM 4D printing and an autonomous 2D-to-3D shape transformation mechanism.We propose a design as shown in Figure 13b.The double-helix parts are printed in the same bilayer structure with specific printing patterns.The base pairs in the middle of DNA-like structure are printed with a passive pattern which is not sensitive to the heat activation.The printing settings are set as 220 °C printing temperature, 40 mm s À1 speed, and 0.1 mm printing layer height.Finally, the printed flat structures in Figure 13c are immersed in the hot water (85 °C) and start to change into double-helix structure, as demonstrated in Figure 13d.During this process, step effect would be eliminated by shape change, which could enhance the strength of structures.It is true that printing condition of programming capabilities corresponds with the minimization of step effect after 3D printing.The better connection quality of layers tends to smaller step effect during 3D printing, which also contributes to stronger 4D-printed structures.This autonomous 2D-to-3D shape transformation deleted the need for printing extra support materials.This shape transformation offers the following advantages.1) Printing with less materials: less materials are consumed in this case as the printer does not need to print support materials.2) Shorter printing time: the printing time is shorter as the printer does not need to print extra support materials in 3D space.3) Less electric power consumption: the two points mentioned above reduce the printer power consumption.4) Better surface quality/mechanical strength/fatigue life: since FDM is a layer-by-layer stacking procedure, there is always a step effect between layers along the z axis when printing complex 3D shapes with or without supports. [35]The step effect increases the surface roughness automatically and lowers the mechanical strength significantly and consequently their fatigue life. [36]In the current work, the printer fabricates a 2D planar structure with no supports and a better surface quality and its transformed 3D structure has a better quality compared to its 3D counterpart printed with support materials in 3D space.Surface accuracy has a direct effect on the mechanical strength and fatigue life.5) Sustainability: the four points mentioned above promote sustainability via 4D printing of 3D structures with lower materials, time, energy, and longer service life.
DNA is a well-known double-helix structure, as depicted in Figure 13a.However, due to its complexity, traditional printing methods struggle to accurately reproduce such structures.Support structures are typically required for the hanging parts during printing, and the layer-by-layer printing approach often weakens the mechanical strength in specific directions.Building upon our previous success in creating helix structures, we now aim to fabricate double-helix structures using direct FDM 4D printing and an autonomous 2D-to-3D shape transformation mechanism.Our proposed design, as shown in Figure 13b, involves printing the double-helix parts within the same bilayer structure, utilizing specific printing patterns.The base pairs in the middle of the DNA-like structure are printed using a passive pattern that is not sensitive to heat activation.The printing parameters are set as follows: a printing temperature of 220 °C, a printing speed of 40 mm s À1 , and a printing layer height of 0.1 mm.Subsequently, the initially printed flat structures, illustrated in Figure 13c, are immersed in hot water at 85 °C, initiating the transformation process into double-helix structures, as demonstrated in Figure 13d.During this process, the step effect typically associated with 3D printing is eliminated through the shape change, resulting in enhanced structural strength.Optimizing the printing conditions for improved programming capabilities minimizes the step effect, thereby enhancing the connection quality between layers and contributing to stronger 4D-printed structures.This autonomous 2D-to-3D shape transformation eliminates the need for printing additional support materials, offering several advantages. 1) Reduced material consumption: The absence of support material printing reduces the overall material consumption, making the process more efficient and cost-effective.2) Shorter printing time: Without the need for printing extra support materials in 3D space, the printing time is significantly reduced, allowing for faster fabrication of the desired structures.3) Lower electric power consumption: The reduction in printing time and material consumption also leads to lower power consumption, making the process more energy efficient.4) Improved surface quality, mechanical strength, and fatigue life: In traditional layer-by-layer 3D printing, a step effect between layers along the z-axis often results in surface roughness, decreased mechanical strength, and reduced fatigue life.In contrast, the proposed 2D planar printing approach yields structures with superior surface quality, leading to improved mechanical strength and fatigue life compared to 3D counterparts printed with support materials.5) Enhanced sustainability: The aforementioned advantages, including reduced material usage, shorter printing time, lower energy consumption, and improved longevity, contribute to a more sustainable approach to 4D printing of 3D structures.By leveraging the autonomous 2D-to-3D shape transformation mechanism, we can achieve the fabrication of complex double-helix structures with improved efficiency, quality, and sustainability.
Given the capability of 4D printing structures to undergo shape change through heat stimulation, it is natural to explore its potential applications.In this context, we propose the design of a functional gripper that leverages this shape-changing property.The gripper design, depicted in Figure 14a, consists of a three-fingered structure printed in a 2D plane using specific printing patterns that enable a catching motion.The printing parameters are configured as follows: a printing temperature of 220 °C, a printing speed of 40 mm s À1 , and a printing layer height of 0.1 mm.Upon applying heat, the initially flat 2D structure initiates a shape transformation, allowing it to effectively catch and hold a 3D toy, as demonstrated in Figure 14b,c.This gripper design showcases its potential in applications where remote triggering, easy transportation, and grasping capabilities are required.The functional gripper represents just one example of how 4D printing structures can be utilized for practical purposes.By harnessing the shapechanging behavior enabled by heat stimulation, we can envision a wide range of applications in fields such as robotics, automation, and material engineering, where adaptable and programmable structures are highly desirable.The ability to remotely trigger shape changes and facilitate the transportation and grasping of objects opens up new possibilities for enhanced manipulation and handling capabilities in various industries.

Conclusion
In this article, we have proposed a direct FDM 4D printing method and conducted a series of experiments to gain a comprehensive understanding of its mechanism and provide insights on how to build a database for driving more designs in 4D printing.Through our experimental investigations, we have observed nearly linear relationships between prestrain and printing parameters such as temperature, speed, and layer height.Notably, the steepest curve was observed for prestrain over printing temperature, indicating that temperature has the greatest impact on induced prestrain.Based on our experiments, we have determined the optimal printing temperature, speed, and layer height to be 210 °C, 40 mm s À1 , and 0.05 mm, respectively.However, it is important for designers to consider the trade-offs between time cost and printing quality when choosing the optimal combination.Furthermore, we have explored the influence of printing patterns on the induced prestrain and found that different patterns lead to variations in the prestrain induced during printing.Building upon the proposed bilayer structures capable of out-of-plane bending, we have demonstrated that beam structures with different printing patterns exhibit superior bending capabilities compared to other bilayer structure types.Moreover, by varying the proportion of upper and lower layers in the bilayer structures, we have discovered a good nearly linear relationship between the bending capability and the ratio of upper layers to lower layers.This finding has significant implications for constructing a comprehensive database for direct FDM 4D printing.
Our research has demonstrated the potential of the proposed direct FDM 4D printing method in creating complex structures that are challenging to produce using traditional 3D printing and manufacturing techniques.By eliminating the need for a layer-by-layer stacking process to achieve complex 3D shapes, FDM-based 4D printing promotes sustainability by reducing material consumption, printing time, and energy consumption, while also extending the service life of the printed structures.With this method, we can easily transform 2D planar structures into intricate 3D structures.By leveraging different printing patterns and adjusting printing settings to achieve varying prestrain, we can create a diverse range of helical and twisting structures.Additionally, we have successfully fabricated a DNA-like structure that would be difficult to realize using traditional manufacturing methods.Finally, we have showcased the functional capabilities of the direct FDM 4D printing method by designing a gripper that can catch objects through shape-changing motion triggered by hot water immersion.
In conclusion, the proposed direct FDM 4D printing method offers a new avenue for structural design and has the potential to be applied in various applications for a more sustainable future.Its advantages include the ability to create complex structures, simplified fabrication processes, improved surface quality and mechanical strength, and enhanced manipulation capabilities.

Future Work
We highlight areas for potential improvement and further exploration here.In order to achieve a more comprehensive study of FDM 4D printing, it would be valuable to explore different cooling conditions and speeds, as they could significantly impact the shape fixity and shape recovery of shape memory polymers.Modifying commercial 3D printers to incorporate different cooling capabilities would be necessary to facilitate such investigations.
Regarding the coupling of printing parameters in creating prestrain, it is indeed challenging to fully decouple them using a control variable method.While reducing the printing layer height may increase the reheating zone, it also leads to higher prestrain.Thinner layers promote increased stretching and result in better connections, thus contributing to higher prestrain compared to the influence of the reheated zone.Consequently, the prestrain is influenced by the stretching degree, connection quality, and reheated amounts, which are affected collectively by printing temperature, printing speed, and printing layer height.This interdependency of parameters should be considered in future studies.
The advantages of 4D printed helical and twisting structures, such as not requiring support materials and exhibiting better structural integrity and isotropy, have been highlighted.In contrast, 3D printed helical and twisting structures, which are printed layer by layer, tend to be anisotropic and prone to breaking, often necessitating the use of support materials.The enhanced performance of 4D printing in these aspects has been well-established.However, it would be meaningful to conduct a numerical study to quantify the extent of the mechanical strength improvement achieved by 4D printing compared to traditional 3D printing.This would provide valuable insights for future developments.

Figure 2 .
Figure 2. Methods of direct FDM 4D printing.a,b) in-plane shape change; c,d) out-of-plane shape change, red part stands for active layers and gray part stands for passive layers.

Figure 3 .
Figure 3.The schematic of calculations for a) prestrain and b) bending capability.

Figure 4 .
Figure 4.The schematic of a) design and b) printing pattern of samples; c) configuration of printed samples.

Figure 5 .
Figure 5.The schematic of experimental results; a) the prestrain induced during printing over printing temperature; b) the expansions in height and width direction over printing temperature.

Figure 6 .
Figure 6.The schematic of experimental results: a) the prestrain induced during printing over printing speed; b) the expansions in height and width direction over printing speed.

Figure 7 .
Figure 7.The schematic of experimental results: a) the prestrain induced during printing over printing layer height; b) the expansions in height and width direction over printing layer height.

Figure 8 .
Figure 8. a) Types of different printing patterns; b) experimental results of prestrain over different printing pattern types.

Figure 9 .
Figure 9. a-f ) The schematic of designs and configuration of samples with shape change of different bilayer structures; g) experimental results of bending capability over different types of bilayer structures.

Figure 10 .
Figure 10.a) The printed sample after heating and cooling process; b) experimental results of bending capability over ratio of upper layers to lower layers.

Figure 11 .
Figure 11.The schematic of a) seedpods after being dried out, b) designs for helical structures by bilayer beam structure with specified printing patterns and setting; c) the configurations of helical structures changed from designed beam structures after heat actuation with the setting of angle from 15°to 75°.

Figure 12 .
Figure 12. a) The schematic of designs for twisting structures by bi-layer beam structure with specified printing patterns and setting; b) the configurations of twisting structures changed from designed beam structures after heat actuation with the setting of angle from from 15°to 75°.

Figure 13 .
Figure 13.The schematic of designs for DNA-like structure: a) inspired DNA double-helix structure; b) designs with specified printing setting; c) printed object; d) shape transformation evolution in the hot water (Video S1, Supporting Information); e) desired DNA-like structures after shape transformation.

Figure 14 .
Figure 14.The schematic of designs functional gripper: a) designs with specified printing setting; b) configurations of printing results and desired functional gripper; c) performance of functional gripper, showing picking up a dog toy by shape transformation evolution in the hot water (Video S2, Supporting Infrmation).