Vat Photopolymerization 3D-Printing of Dynamic Thiol-Acrylate Photopolymers Using Bio-Derived Building Blocks

As an energy-efficient additive manufacturing process, vat photopolymerization 3D-printing has become a convenient technology to fabricate functional devices with high resolution and freedom in design. However, due to their permanently crosslinked network structure, photopolymers are not easily reprocessed or repaired. To improve the environmental footprint of 3D-printed objects, herein, we combine the dynamic nature of hydroxyl ester links, undergoing a catalyzed transesterification at elevated temperature, with an acrylate monomer derived from renewable resources. As a sustainable building block, we synthesized an acrylated linseed oil and mixed it with selected thiol crosslinkers. By careful selection of the transesterification catalyst, we obtained dynamic thiol-acrylate resins with a high cure rate and decent storage stability, which enabled the digital light processing (DLP) 3D-printing of objects with a structure size of 550 µm. Owing to their dynamic covalent bonds, the thiol-acrylate networks were able to relax 63% of their initial stress within 22 min at 180 °C and showed enhanced toughness after thermal annealing. We exploited the thermo-activated reflow of the dynamic networks to heal and re-shape the 3D-printed objects. The dynamic thiol-acrylate photopolymers also demonstrated promising healing, shape memory, and re-shaping properties, thus offering great potential for various industrial fields such as soft robotics and electronics.


Introduction
Additive manufacturing (AM) or 3D-printing is gaining increasing attention as a versatile processing technique for designing polymeric prototypes [1][2][3][4][5]. To date, 3D-printed polymers have been successfully integrated in electronics [6,7], automotive [8], robotics [7,9], tissue engineering, and biomedical industries [10][11][12][13][14][15][16][17]. With the rising development of new raw materials and technological innovations for 3D-printing, the global market potential of AM technologies has grown exponentially over the past years and is projected to reach approximately $20 billion by 2029 [18,19]. Using vat photopolymerization 3D-printing for the fabrication of 3D-objects offers numerous advantages, including ease of raw materials' handling, ambient printing conditions, and the ability to form complex structures at a reasonable build time with high resolution. However, a major drawback of photopolymers is their lack in recyclability and repairability, as liquid resin transforms to a solid covalently-crosslinked polymer network during the printing process.
The introduction of dynamic covalent bonds is one way to render 3D-printed photopolymers repairable and reusable. In particular, dynamic polymer networks following an associative bond exchange mechanism exhibit excellent thermal and mechanical properties, and behave like classic thermosets below their topology freezing temperature (T v ). Increasing the temperature above the T v promotes a rearrangement of the network's topology, which induces a macroscopic viscoelastic reflow [20,21]. This material flow follows

Synthesis of Acrylated Linseed Oil
Acrylated linseed oil (AELO) was synthesized by adding epoxidized linseed oil (ELO) (100.05 g), acrylic acid (23.076 g, 0.32 mol), and butylated hydroxytoluene (0.33 g) to a three-necked round-bottomed flask. After stirring the mixture at 60 • C for 20 min, triphenylphosphine (8.12 g, 0.037 mol) was added. The mixture was stirred for another hour at 60 • C. Further acrylic acid (22.055 g, 0.306 mol) was then added, and the mixture was kept at 60 • C overnight under stirring. Subsequently, the mixture was allowed to cool down to room temperature and was dissolved in 100 mL of ethyl acetate. Excess acrylic acid was removed by washing the organic phase three times with saturated sodium bicarbonate solution followed by a washing step with brine. The organic layer was dried over sodium sulfate and the solvent was removed under reduced pressure. The product was obtained as yellow liquid with an acrylate functionality of 3.

Resin Preparation
AELO (70 wt%), the transesterification catalyst DMEP (8 wt%), the photoinitiator BAPO (2 wt%), and 0.05 wt% Sudan II as photo-absorber were mixed and stirred in an ultrasonic mixer for 30 min at 40 • C until BAPO and Sudan II were dissolved. The photo-absorber (Sudan II) was added to prevent over-polymerization and to ensure high resolution during DLP 3D-printing. The respective thiol crosslinker (20 wt%) was added after cooling the formulation to room temperature. Light-protected glass vials were utilized for the preparation of the resins. The compositions are summarized in Table 1 and the chemical structures of the components in the resin formulations are provided in Figure 1.

Digital Light Processing (DLP) 3D-Printing
DLP 3D-printing was performed on an Anycubic Photon Zero printer (Shenzhen, China) with a LED 405 nm light source. The overall light exposure was controlled using bottom and normal exposure settings with a corresponding duration of 90 s and 70 s, based on the optimal monomer conversions estimated from the FTIR spectroscopy. The 3D-printed structures were built up layer-by-layer with a layer thickness of 50 µm. Building and retracting speed of the platform were set to 2 mm s −1 . An optical microscope (Olympus BX 51, Tokyo, Japan) was utilized to record images of the DLP 3D-printed test structures with a Color View IIIu digital camera (Soft Imaging System, Münster, Ger-  2.3. Digital Light Processing (DLP) 3D-Printing DLP 3D-printing was performed on an Anycubic Photon Zero printer (Shenzhen, China) with a LED 405 nm light source. The overall light exposure was controlled using bottom and normal exposure settings with a corresponding duration of 90 s and 70 s, based on the optimal monomer conversions estimated from the FTIR spectroscopy. The 3D-printed structures were built up layer-by-layer with a layer thickness of 50 µm. Building and retracting speed of the platform were set to 2 mm s −1 . An optical microscope (Olympus BX 51, Tokyo, Japan) was utilized to record images of the DLP 3D-printed test structures with a Color View IIIu digital camera (Soft Imaging System, Münster, Germany).

Evaluation of Reaction Kinetics and Material Characterization
Fourier transform infrared radiation (FTIR) spectroscopy was performed on a Vertex 70 spectrometer (Bruker, Billerica, MA, USA) to analyze the reaction progress over the illumination time. 16 scans were measured for each exposure time interval in transmission mode between 4000 to 700 cm −1 , with a resolution of 4 cm −1 . For sample preparation, the liquid resin (1.2 µL) was cast between two CaF 2 slides. Photocuring reactions were studied by illuminating the resin with a LED light lamp (zgood ® wireless) with a power density of 3.3 mW cm −2 (λ = 420-450 nm). Reaction conversions were estimated from the normalized characteristic peak areas of acrylate (1635 cm −1 ) and thiol (2570 cm −1 ) groups with the corresponding exposure dose using the software OPUS. Normalization of the acrylate and thiol peaks was carried out using C=O peak (1750 cm −1 ), which remained unaltered during the curing process (Supporting Information S2).
FT-IR spectra of cured and thermally-treated samples (180 • C for 4 h under air) were recorded with a FT-IR spectrometer (Vertex 70, Bruker, Billerica, MA, USA) integrated with a reflection diamond attenuated total reflection (ATR) accessory (Platinum ATR). A total of 16 scans were measured for each sample from 4000 to 400 cm −1 .
Stress relaxation experiments were carried out at temperatures between 140 and 180 • C on DLP 3D-printed samples (d = 10 mm, w = 1 mm) with a rheometer (Anton Paar Physica MCR-501, Graz, Austria) in parallel plate configuration. Prior to the relaxation measurements, samples were equilibrated at 20 N normal force and the corresponding temperature for 20 min. Subsequently, 3% strain was applied and the decreasing stress was recorded over a period of time. Tensile properties of the materials were characterized on DLP 3D-printed specimens (l = 30 mm, w = 10 mm, and h = 1.5 mm) with Zwick Roell Z1.0 static material testing equipment (Ulm, Germany) at a cross-head speed of 250 mm min −1 .
The glass transition temperatures (T g ) of the DLP 3D-printed samples were measured by differential scanning calorimetry (Perking Elmer DSC 400, Waltham, MA, USA). A heating rate of 20 K min −1 was applied from −50 • C to 200 • C under nitrogen atmosphere. The T g of the samples was determined by taking the midpoint of the heat capacity from the second heating run.
The viscosity of the resins was determined by using a compact modular rheometer (MCR 102 Anton Paar, Graz, Austria) with a CP60-0.5/TI cone (49.97 mm diameter and 1.982 • opening angle). Each measurement was performed with 1.2 mL resin at 30 • C and a shear rate ranging from 0.1 to 300 s −1 .
The gel content of the photopolymers was measured by immersing the 3D-printed samples in dichloromethane (DCM) at 20 • C for 48 h. After removing and drying the samples with paper tissues, the specimens were placed in an oven at 40 • C until constant weight. The gel content was determined by taking the mass ratios m dry /m initial , where m dry represents the weight of the final dried samples and m initial represents the weight of the initial samples. Five samples of each resin were measured and the arithmetic average was determined.
Thermal healing experiments were carried out on DLP 3D-printed rectangular bars (l = 30 mm, w = 10 mm, h = 1.5 mm) with a central cavity (d = 5 mm). Circular discs (5 × 1.5 mm) and control samples without cavity (same dimensions) were also printed with the resin formulation AELO-EGMP. For evaluating the healing efficiency, the circular discs were placed inside the cavities of the rectangular bars and the samples were thermally treated at 180 • C for 4 h in the oven. Tensile tests were performed on rectangular bars with a cavity, healed samples, and control samples with Zwick Roell Z1.0 static material testing equipment (Ulm, Germany) at a cross-head speed of 250 mm min −1 .
Shape memory experiments were performed with DLP 3D-printed grippers (d = 50 mm, w = 1 mm) after a thermal treatment at 180 • C for 4 h. The first shape of a 3D-printed gripper was fixed by heating the sample to 180 • C for 2 h. For programming the second shape, the gripper was cooled down to 30 • C, re-shaped, and further cooled down to −20 • C in a deep freezer. By heating the gripper to 30 • C, the first shape was regained within 30 s. The permanent shape of the specimen was restored by heating the gripper to 180 • C for 2 h.

Synthesis and Network Evolution of Dynamic Thiol-Acrylate Photopolymers
The acrylated linseed oil (AELO) was obtained via the nucleophilic ring opening reaction of an epoxidized linseed oil with acrylic acid ( Figure S1 in supporting information), as reported in the literature [53]. The one-pot synthesis afforded the acrylated product at high yields, which was confirmed by 1 H NMR spectra in deuterated chloroform (Figures S2 and S3 in supporting information). The signal at 2.3 ppm is assigned to the three CH 2 groups (H f ) attached at the α-position of the ester carbonyl. This peak was also employed as a reference to quantify the introduced acrylates (5.85, 6.15, 6.4 ppm). By setting the integral value of H f to 6, the integrals of H a-c could be directly used to determine the average acrylate functionality per triglyceride molecule. Furthermore, the decrease in epoxide functionality (3.0 ppm) could be observed from 5 to 0.4 per triglyceride molecule, showing a successful reaction. According to the NMR data, an AELO with an acrylate functionality of 3.2 was synthesized. Since the residual epoxide functionality was 0.45, it was calculated that 90% of the epoxide groups participated in the reaction. However, since more epoxides reacted than acrylates formed, it is assumed that a part of the epoxy groups did hydrolyze or participated in other side reactions. An acrylate functionality of 3.2 was sufficient for all further reactions; therefore, no effort was made to achieve higher epoxide conversions.
AELO-thiol resins were prepared by employing two different types of thiols with varying functionality: the trifunctional thiol TMP3MP and the difunctional thiol EGMP. BAPO was used as a radical photoinitiator, whose extended absorption profile allowed for the curing of the AELO-resins under visible light exposure ( Figure 2). The curing kinetics of the reaction were monitored by following the depletion of the characteristic stretching bands of S-H bonds at 2570 cm −1 and the C=C bond wagging band at 1635 cm −1 over the exposure dose (IR spectra prior to and after illumination are shown in Supporting Information Figures S4 and S5). In both resins, the final acrylate conversion was significantly higher (87% and 97%) compared to the thiol conversion, which amounted to 46% and 48% for AELO-TMP3MP and AELO-EGMP, respectively. This can be explained by the participation of the acrylate moieties in both reaction pathways-the chain growth homo-polymerization (acrylate-acrylate) as well as the step-growth thiol-ene addition reaction (thiol-acrylate)-leading to a mixed-mode reaction mechanism [26,27,54]. While chain growth polymerization reactions show fast conversions and lead to high molecular weights, the step growth polymerization favors the formation of lower molecular weight species, resulting in delayed gelation, more homogenous networks, and lower shrinkage stress [26,27,50,54,55]. It is reported in the literature that the rate of homo-polymerization is 1.5 times higher than the rate of hydrogen abstraction from thiol moieties [50,56], which ultimately leads to the higher conversion of acrylates in thiol-acrylate-based networks. action (thiol-acrylate)-leading to a mixed-mode reaction mechanism [26,27,54]. Wh chain growth polymerization reactions show fast conversions and lead to high molecu weights, the step growth polymerization favors the formation of lower molecular wei species, resulting in delayed gelation, more homogenous networks, and lower shrink stress [26,27,50,54,55]. It is reported in the literature that the rate of homo-polymerizat is 1.5 times higher than the rate of hydrogen abstraction from thiol moieties [50,56], wh ultimately leads to the higher conversion of acrylates in thiol-acrylate-based networks  The final conversion of acrylate moieties was higher in the case of AELO-EGMP (97%), in contrast to AELO-TMP3MP (87%), while the thiol conversion also decreased with increasing thiol functionality (Figure 3a,b), which might be due to the diffusion limitation of the reactive monomers. The increase in thiol functionality leads to a decrease in gel point conversion and the polymerization rate slows down due to diffusional constraints. Moreover, almost similar reaction kinetics were recorded during the first 70 s of illumination, and acrylate conversion was above 70% in both cases. It should also be noted that the trifunctional thiol acts as a crosslinker, whereas the difunctional thiol acts as a chain extender and increases the molecular weight between adjacent crosslinks [27,55,57]. This was confirmed by additional sol-gel analysis. The gel content of the cured samples was determined by immersing the 3D-printed samples for 48 h in dichloromethane, and the residual weight of the dried samples was recorded. The networks formed with difunctional thiol (EGMP) exhibited a gel content of 71.5% ± 1.9%, while the networks formed with trifunctional thiol (TMP3MP) had a gel content of 77.7% ± 2.3%. The higher gel content of the AELO-TMP3MP system can be explained by an increased crosslink density due to the higher functionality of the thiol used. Generally, the low gel content can be related to an incomplete monomer conversion and, especially, the thiol crosslinkers, which were extracted from the networks during swelling.

Network Properties of Dynamic Thiol-Acrylate Photopolymers
The residual thiols act as chain lubricants and in combination with flexible thioether bonds, and the mobility of the networks is significantly enhanced, which leads to glass transition temperatures (T g ) of −18 • C and −10 • C for AELO-EGMP and AELO-TMP3MP, respectively. The related DSC curves are shown in Figures S7 and S8 in supporting information. For uniaxial tensile tests, samples were prepared via DLP 3D-printing. Based on the low T g , the networks were soft at room temperature and were characterized by a low tensile strength, which did not exceed 1.18 MPa and 0.58 MPa for TMP3MP-and EGMP-based formulations, respectively. The ultimate elongation of AELO-EGMP (19 %) was lower compared to AELO-TMP3MP (38 %) (Figure 3c). By thermal annealing of the 3D-printed samples at 180 • C for 4 h, the Young's modulus and the mechanical properties were improved (Figure 3d). After the thermal treatment, both networks underwent an increase in the tensile strength, and exhibited a maximum strain of 31 % (AELO-EGMP) and 37 % (AELO-TMP3MP), and a breaking stress of 1.0 MPa (AELO-EGMP) and 1.3 MPa (AELO-TMP3MP), respectively. This improvement in toughness and modulus can be attributed to the thermally-induced rearrangements and the reduction in shrinkage stress. Previous work demonstrated that the thermal treatment also enables the formation of hydrogen bonds, which further increases the tensile properties [27,58,59]. The analysis of ATR FTIR spectra of the thermally-treated samples also clearly indicates the depletion in the -OH groups' region (3300-3600 cm −1 ), confirming the formation of hydrogen bonds ( Figure S6). gel point conversion and the polymerization rate slows down due to diffusional con-straints. Moreover, almost similar reaction kinetics were recorded during the first 70 s of illumination, and acrylate conversion was above 70% in both cases. It should also be noted that the trifunctional thiol acts as a crosslinker, whereas the difunctional thiol acts as a chain extender and increases the molecular weight between adjacent crosslinks [27,55,57]. This was confirmed by additional sol-gel analysis. The gel content of the cured samples was determined by immersing the 3D-printed samples for 48 h in dichloromethane, and the residual weight of the dried samples was recorded. The networks formed with difunctional thiol (EGMP) exhibited a gel content of 71.5% ± 1.9%, while the networks formed with trifunctional thiol (TMP3MP) had a gel content of 77.7% ± 2.3%. The higher gel content of the AELO-TMP3MP system can be explained by an increased crosslink density due to the higher functionality of the thiol used. Generally, the low gel content can be related to an incomplete monomer conversion and, especially, the thiol crosslinkers, which were extracted from the networks during swelling.  The kinetics of the thermo-activated bond exchange reactions was characterized by stress relaxation experiments between 140 • C and 180 • C. Samples with a diameter of 10 mm were DLP 3D-printed and the impact of the thiol crosslinker on the exchange kinetics was determined. Based on the Maxwell model, the characteristic time for stress relaxation τ* can be determined from the measured relaxation plots when the normalized relaxation modulus (G t /G o ) reaches 37 % (1/e) of its initial value [60]. Generally, the relaxation time (τ*) in vitrimers follows an Arrhenius type temperature dependency τ* = τ o exp (E a /RT), where E a represents the activation energy, T the temperature, and R the general gas constant [60][61][62]. networks, albeit at different rates. Furthermore, the thiol-acrylate networks were able to relax stress in the absence of a catalyst; however, the rate was much lower. This relaxation is contributed by the thermal release of volumetric shrinkage stresses generated during the network formation [26,63]. Without the catalyst, AELO-EGMP reached 37% stress within a time period of 384 min at 180 • C (Figure 4a), whilst the catalyzed AELO-EGMP network exhibited a τ* of 22 min.

Thermally Triggered Healing and Shape Memory Properties of DLP 3D-Printed Objects
Due to the lower viscosity of AELO-EGMP resin (2.1 Pa·s obtained at a shear rate between 0.1-300 s −1 ) compared to AELO-TMP3MP (11.9 Pa·s obtained at the same shear rate range), it was further applied to DLP 3D-print objects to study the material resolution, thermal healing, and shape-memory performance.
A comb-like structure was DLP 3D-printed with holes of different sizes to monitor the printing capabilities (Figure 5a). We observed that it was possible to print holes with a diameter up to 250 µm. At lower feature sizes, the material started to lose resolution and no open holes could be obtained (Figure 5b). The relaxation rate of catalyzed AELO-EGMP (22 min) is much faster in comparison to catalyzed AELO-TMP3MP (77 min), which can be attributed to the higher crosslinking degree of AELO-TMP3MP, as the number of -OH and ester groups was comparable in both networks. Moreover, the stress-relaxation of both networks shows a clear temperature dependency and the bond exchange rate increases with rising temperature (Figure 4b,c). In Figure 4d, τ* is plotted versus 1/T in a semilogarithmic scale and the linear trend obtained for both networks confirms the Arrhenius type temperature dependency and the vitrimeric behavior of the developed networks [64]. The E a of the networks was also determined, by taking the slope (m = E a /R) of the straight line fitted to the data, and it amounted to 77.58 and 79.88 kJ·mol −1 ·K −1 for AELO-EGMP and AELP-TMP3MP, respectively.

Thermally Triggered Healing and Shape Memory Properties of DLP 3D-Printed Objects
Due to the lower viscosity of AELO-EGMP resin (2.1 Pa·s obtained at a shear rate between 0.1-300 s −1 ) compared to AELO-TMP3MP (11.9 Pa·s obtained at the same shear rate range), it was further applied to DLP 3D-print objects to study the material resolution, thermal healing, and shape-memory performance.
A comb-like structure was DLP 3D-printed with holes of different sizes to monitor the printing capabilities (Figure 5a). We observed that it was possible to print holes with a diameter up to 250 µm. At lower feature sizes, the material started to lose resolution and no open holes could be obtained (Figure 5b). Due to the dynamic nature of the developed thiol-acrylate networks, the AEL EGMP photopolymer exhibits triple shape memory and the ability to undergo controll and macroscopic deformation when the sample is heated and programmed beyond two transition temperatures, Tv and Tg. As presented in Figure 6a,b, the permanent sha of a DLP 3D-printed and thermally-annealed flower was altered by heating it above t highest transition temperature (Tv), along with applying an external deformation force temperature of 180 °C was applied since rheological experiments revealed a fast str relaxation at this temperature (Figure 4b). The first temporary shape was fixed by cooli the sample to 30 °C, which is well above the Tg of the material and enabled the seco programming. By applying a deformation force and cooling the sample below Tg (−20 ° the second temporary shape was fixed. Subsequently, the specimens were able to seque tially recover the different shapes under thermal stimulus without applying any force recovery time of 30 s was recorded for the first temporary shape, which also highlig the potential application of partly bio-based thiol-acrylate resins in biomedical and s robotic systems. In a second step, the thermally-induced healing and repair abilities of the AEL EGMP photopolymer have been studied. Rectangular bar-shaped samples with (defe and without (defect-free control sample) a circular cavity in the center were DLP 3 Due to the dynamic nature of the developed thiol-acrylate networks, the AELO-EGMP photopolymer exhibits triple shape memory and the ability to undergo controlled and macroscopic deformation when the sample is heated and programmed beyond its two transition temperatures, T v and T g . As presented in Figure 6a,b, the permanent shape of a DLP 3D-printed and thermally-annealed flower was altered by heating it above the highest transition temperature (T v ), along with applying an external deformation force. A temperature of 180 • C was applied since rheological experiments revealed a fast stress relaxation at this temperature ( Figure 4b). The first temporary shape was fixed by cooling the sample to 30 • C, which is well above the T g of the material and enabled the second programming. By applying a deformation force and cooling the sample below T g (−20 • C), the second temporary shape was fixed. Subsequently, the specimens were able to sequentially recover the different shapes under thermal stimulus without applying any force. A recovery time of 30 s was recorded for the first temporary shape, which also highlights the potential application of partly bio-based thiol-acrylate resins in biomedical and soft robotic systems.  Figure 6. (a) Photographs representing the triple shape memory of DLP 3D-printed AELO-EGMP structures. The first shape was programmed above the Tv of AELO-EGMP while the second shape of material was programmed using the Tg of the photopolymer. Heating of the programmed object sequentially recovered the permanent shape of the 3D-printed flower; (b) DLP 3D-printed structure representing the potential application as gripper by employing shape memory properties; (c) DLP 3D-printed control sample, damaged sample with a circular cavity, the disc, and the healed sample for analysis of thermal welding and healing properties; (d) Stress-strain curves of control, damaged, and healed samples. Thermal healing was performed at 180 °C for 4 h.

Conclusions
In the present work, acrylated linseed oil was synthetized and used as a building block for the preparation of dynamic thiol-acrylate photopolymers, which relied on thermo-activated transesterification. Structure-property relationships with two different thiols, varying in their functionality (bi-and tri-functional), were comprehensively described. The bifunctional thiol yielded a lower crosslinked photopolymer with a Tg of −18 °C compared to the trifunctional counterpart (Tg = −10 °C). The lower crosslinking degree enhanced the mobility of the AELO-EGMP networks, which accelerated the bond exchange reactions at an elevated temperature (>140 °C). With both thiols, thiol-acrylate networks were rapidly formed upon visible light exposure, which is crucial for DLP 3Dprinting. Taking advantage of the lower viscosity, EGMP-AELO was applied in subsequent printing experiments and it was found that objects with a feature size up to 550 µm could be conveniently fabricated. The EGMP-AELO network comprised ample free ester and hydroxyl groups, which, in the presence of an organic phosphate ester as a catalyst, underwent bond exchange reactions at elevated temperatures (>140 °C). The related macroscopic reflow was used for the healing and re-shaping of the printed structures. Tensile Figure 6. (a) Photographs representing the triple shape memory of DLP 3D-printed AELO-EGMP structures. The first shape was programmed above the T v of AELO-EGMP while the second shape of material was programmed using the T g of the photopolymer. Heating of the programmed object sequentially recovered the permanent shape of the 3D-printed flower; (b) DLP 3D-printed structure representing the potential application as gripper by employing shape memory properties; (c) DLP 3D-printed control sample, damaged sample with a circular cavity, the disc, and the healed sample for analysis of thermal welding and healing properties; (d) Stress-strain curves of control, damaged, and healed samples. Thermal healing was performed at 180 • C for 4 h.
In a second step, the thermally-induced healing and repair abilities of the AELO-EGMP photopolymer have been studied. Rectangular bar-shaped samples with (defect) and without (defect-free control sample) a circular cavity in the center were DLP 3D-printed. A disc fitting in the cavity was further printed and inserted in the defect sample. For the thermal healing/welding, the sample was placed at 180 • C for 4 h in the oven without applying any pressure (Figure 6c). Subsequent tensile testing of the healed samples revealed that the stress-strain curves are nearly identical to the defect-free control sample, giving rise to the high healing efficiency of the AELO-EGMP photopolymer (Figure 6d).

Conclusions
In the present work, acrylated linseed oil was synthetized and used as a building block for the preparation of dynamic thiol-acrylate photopolymers, which relied on thermoactivated transesterification. Structure-property relationships with two different thiols, varying in their functionality (bi-and tri-functional), were comprehensively described. The bifunctional thiol yielded a lower crosslinked photopolymer with a T g of −18 • C compared to the trifunctional counterpart (T g = −10 • C). The lower crosslinking degree enhanced the mobility of the AELO-EGMP networks, which accelerated the bond exchange reactions at an elevated temperature (>140 • C). With both thiols, thiol-acrylate networks were rapidly formed upon visible light exposure, which is crucial for DLP 3D-printing. Taking advantage of the lower viscosity, EGMP-AELO was applied in subsequent printing experiments and it was found that objects with a feature size up to 550 µm could be conveniently fabricated. The EGMP-AELO network comprised ample free ester and hydroxyl groups, which, in the presence of an organic phosphate ester as a catalyst, underwent bond exchange reactions at elevated temperatures (>140 • C). The related macroscopic reflow was used for the healing and re-shaping of the printed structures. Tensile properties of defect free samples could be nearly fully recovered after a thermal treatment at 180 • C for 4 h. Furthermore, the 3D-printed structures were able to undergo triple-shape memory and two different shapes were programmed by exploiting the topological freezing temperature and the glass transition temperature of the dynamic thiol-acrylate photopolymers.