3D Printed Shape Memory Objects Based on Olefin Ionomer of Zinc

especially for extrusion based printing such as Fused Deposition Modeling (FDM). Here we demonstrate shape memory behavior of 3D printed objects with ...
0 downloads 0 Views 2MB Size
Subscriber access provided by GRIFFITH UNIVERSITY

Article

3D Printed Shape Memory Objects Based on Olefin Ionomer of Zinc-Neutralized Poly(ethylene-co-methacrylic acid) Zhiyang Zhao, Fang Peng, Kevin A. Cavicchi, Mukerrem Miko Cakmak, Robert A. Weiss, and Bryan D. Vogt ACS Appl. Mater. Interfaces, Just Accepted Manuscript • DOI: 10.1021/acsami.7b07816 • Publication Date (Web): 25 Jul 2017 Downloaded from http://pubs.acs.org on July 26, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Applied Materials & Interfaces is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

3D Printed Shape Memory Objects Based on Olefin Ionomer of Zinc-Neutralized Poly(ethylene-comethacrylic acid) Zhiyang Zhao,1 Fang Peng,1 Kevin A. Cavicchi,1 Mukerrem Cakmak,2 R.A. Weiss,1 Bryan D. Vogt1,* 1

Department of Polymer Engineering, University of Akron, Akron, OH 44325

2

Departments of Materials and Mechanical Engineering, Purdue University, West Lafayette, IN

47907 KEYWORDS: Surlyn, Fused Deposition Modeling, Shape Memory Polymer, 4D printing, Additive Manufacturing

ABSTRACT: 3D printing enables the net shape manufacturing of objects with minimal material waste and low tooling costs, but the functionality is generally limited by available materials, especially for extrusion based printing such as Fused Deposition Modeling (FDM). Here we demonstrate shape memory behavior of 3D printed objects with FDM using a commercially available olefin ionomer, Surlyn 9520, which is zinc neutralized poly(ethylene-co-methacrylic acid). The initial fixity for 3D printed and compression molded samples was similar, but the initial recovery was much lower for the 3D printed sample (R=58%) than the compression

ACS Paragon Plus Environment

1

ACS Applied Materials & Interfaces

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 35

molded sample (R=83%). The poor recovery on the first cycle is attributed to polyethylene crystals formed during programming that act to resist the permanent network recovery. This effect is magnified in the 3D printed part due to the higher strain (lower modulus in the 3D printed part) at fixed programing stress. The fixity and recovery on subsequent shape memory cycles is greater for the 3D printed than compression molded part. Moreover, the programmed strain can be systematically modulated by inclusion of porosity in the printed part without adversely impacting the fixity or recovery. These characteristics enable the direct formation of complex shapes of thermoplastic shape memory polymers that can be recovered in 3 dimensions with the appropriate trigger, such as heat, through the use of FDM as a 3D printing technology.

INTRODUCTION The direct fabrication of parts of arbitrary shape and size is enabled by 3D printing, which holds tremendous promise for energy, materials and cost savings through the additive manufacture of customizable objects.1-2 An extension to this net shape production approach is using functional printed material that allows the part to respond to external stimuli to rearrange into a new shape.3 This responsive nature provides an additional dimension to the shape of the part, i.e., time, where the shape evolves. This has been termed 4D printing.4 One route to enable this time response is to print glassy shape memory polymer (SMP) fibers within an elastomer to provide shape memory behavior in the 3D printed part.3 In this case, the initial shape is a flat sheet of the elastomer, but pre-programming the SMP fibers printed in the elastomer enables transformation to complex shapes. This process can be extended as a type of externally programmed origami as the fibers can induce bending of the elastomer to generate the transformed shape.5 Temperature dependent programming to different shapes is possible using a family of SMP fibers with

ACS Paragon Plus Environment

2

Page 3 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

different triggering temperatures.6 Since the initial shape is generally a flat sheet, the distribution of fibers within the sheet ultimately determines the shape to which it transforms. However, the relationship between the fiber distribution and the desired shape is not trivial. An easy solution around this challenge is to directly form the desired shape in a SMP via 3D printing, transform the printed object into an arbitrary temporary shape and then recover the printed permanent shape with an appropriate external stimulus or trigger. The difficulty of this facile approach is the minimum requirement for an SMP to have two distinct networks, a permanent network that defines the permanent shape and drives recovery, and a temporary network that holds the temporary shape in place.7 For common 3D printing technologies, the formation of two networks can be an obstacle. Stereolithography (SLA) can be used to generate a covalently crosslinked network in the part,8 but inclusion of a temporary network is less straightforward. A recent report demonstrated how to 3D print a SMP for flexible electronics by using a custom semicrystalline macromonomer and SLA.9 Additional photopolymers have also been developed to add and/or improve the shape memory behavior of SLA 3D printed parts.10-11 One limitation with SLA is the generally poor mechanical properties of the final part, so there is interest to extend these concepts away from photopolymers to thermoplastics. An alternative to SLA is to 3D print a shape memory thermoplastic using extrusion methods, such as fused deposition modeling (FDM).12 These methods require flow of the polymer for the part to be printed, which does not allow the use of covalently crosslinked materials like formed in SLA. Physically crosslinked polymers, such as filled thermoplastic elastomers that exhibit shape memory behavior13 resolve the flow constraint. However, FDM requires the sufficiently stiff polymer filaments such that the unmelted filament can push the melt out of the extruder, which limits the use of most thermoplastic elastomers (TPEs)14 in FDM printing. For most FDM

ACS Paragon Plus Environment

3

ACS Applied Materials & Interfaces

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 35

printers, the filament is pulled into the heated extrusion nozzle by rollers on the filament drive wheel; common TPEs tend to bend, which prevents application of sufficient force push the melted material out of the extrusion nozzle. In this work, we describe the use of FDM to 3D print filaments of an ionomeric thermoplastic SMP, zinc neutralized poly(ethylene-co-methacrylic acid), PEMA.15 The shape memory behavior of PEMA is enabled by its two networks from nanodomains of supramolecular crosslinks formed by ionic bonds and ethylene crystals. Due to the ionic aggregates, the polyethylene crystals melt over a wide temperature range (40 °C-110 °C) and form the basis of the temporary network, while the unmelted crystals and the ionic clusters provide a permanent network – in this case, the term permanent implies that the relaxation times of the reversible supramolecular bonds are long compared to shape-changing application. The broad melting point for the ethylene crystals provides the opportunity for multi-shape memory behavior,15 but here we focus on a single shape using a transition temperature at 70 °C. This temperature is selected to preserve some crystals in the FDM printed part to avoid irreversible shrinkage during programing from the residual stress during the printed processing. Shape memory behavior of FDM printed PEMA was directly compared with compression molded (conventional) samples. The results indicated that the both printed samples and compression molded sample exhibited lower recovery on the initial shape memory cycle, while the fixity was highest for the initial cycle. Both the fixity and recovery were nearly invariant for subsequent cycles. Moreover, the printing of the PEMA enabled the fabrication of porous SMP structures, whose shape memory performance, in terms of fixity and recovery, did not depend on the porosity (from 8 - 45 %). This work demonstrates a simple route to generate complex shape memory parts using 3D

ACS Paragon Plus Environment

4

Page 5 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

printing and illustrates the potential for lightweight shape memory objects through controlled porosity from 3D printing. EXPERIMENTAL SECTION Materials. The PEMA ionomer, Surlyn 9520, obtained from DuPont de Nemours Co had a nominal methacrylic acid concentration of 10 wt% that was partially neutralized to the zinc methacrylate. The neutralization level was estimated from FTIR (see Figure S1) based on the intensity of the stretching vibration vs(C=O) at 1699 cm-1, which is associated with the acid form, and the symmetric vibration of carboxylate group vs(COO-) at 1585 cm-1, which is associated with the zinc salt.16 For Surlyn 9520, the absorbance of the vs(C=O) is 22 (integrated area underneath the peak), while the absorbance of vs(COO-) is 32, which indicates approximately 60 % neutralization. Prior to use the PEMA pellets were dried at 30°C in a vacuum oven for 24 h. For a control, the dried PEMA was compression molded at 140 °C using a Carver Hydraulic Press (model #3912) to produce a film that was 0.5 mm thick. The film was cut into rectangular bars (20 mm × 5 mm) using a razor blade. The shape memory properties of these compression-molded samples were compared to those of 3D printed samples of a similar shape. Filament preparation. One requirement for 3D printing using Fused Deposition Modeling (FDM) is the availability of high quality filaments with constant diameter of the material of interest. In this case, the filaments of PEMA were produced by extrusion through a capillary rheometer (Rosand RH7 advanced capillary rheometer) at 190 °C with a die diameter of 3 mm and capillary length-to-diameter (L/D) ratio=10. The extrusion speed was 18 mm/min and filaments of the PEMA were collected on an 86 mm diameter tube that was rotating at a constant angular velocity (2 rpm). The filaments were 2.60 mm in diameter with a standard deviation in

ACS Paragon Plus Environment

5

ACS Applied Materials & Interfaces

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 35

diameter of 0.04mm as determined by using multiple point measurements with calipers. This low variance is critical to printing high quality parts with FDM. 3D printing of PEMA using FDM. The PEMA filaments were printed using a customized Cartesio W09 3D printer. The original-platform bed was replaced by an aluminum plate (30 mm thick) that was isolated from the stage using ceramic spacers. The aluminum was heated from below using silicone flexible heaters (Watlow) and the temperature of the print-bed was monitored using a type K thermocouple (Omega Engineering). To improve the adhesion and clean release of the printed part from the bed, the aluminum print-bed was covered with Kapton tape. For all samples printed in this work, a constant bed temperature of 40 °C was used. A hotend assembly with an aluminum heat sink (E3D-V6) was installed to minimize the preheating of the filament. The extrusion temperature of the hot-end was set to 260 °C for all printing conditions examined here. The parts were printed at 10 mm/s with a layer height of 0.3 mm. The input diameter for the printing was 2.6 mm. The G-code of printing commands was generated using Slic3r. In the printing of the part, the porosity was modulated using the infill densities, which introduces a designated gap to adjacent extruded strand on the same layer. In all cases, a dense perimeter around the edges of the part was printed for each layer to generate a vertical shell. In order to enable direct comparisons of the shape memory performance between 3D printed parts and traditional compression molding, simple rectangular samples (50 mm × 5 mm × 1 mm) to match standard DMA specimens were 3D printed with FDM. The parts were printed at two different infill orientations relative to the long axis of the sample: 45°/-45° and 0°/90°. For assessing the influence of porosity on the shape memory performance, the DMA samples were printed with 4 infill densities (100%, 80%, 60%, and 40%). To demonstrate shape memory

ACS Paragon Plus Environment

6

Page 7 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

behavior in complex shapes from 3D printing, a spring with a constant helix pitch of 20 mm was printed at 100 % infill. The spring was clockwise helical with an outer diameter of 25 mm and a wire diameter of 5 mm. The support material was generated during printing under a threshold of 90° using the same PEMA, but with a much lower infill and 0.2 mm contact Z distance between the object and the support material. These parameters were selected to provide good support during the print and easy removal of the support by hand peeling. Structural characterization. A Bruker Skyscan 1172 with 50kV X-ray source (microCT) was used with a 180° scan of the samples without a filter to obtain tomography images. These images were reconstructed to a set of cross-section images of the sample with NRecon software where CTvox and CTanalysis were also used to obtain section images and full 3D-models of the part. An ATOS Core 200 3D scanner (GOM) was used to obtain the 3D shape of the spring optically. The spring was fixed on the black plastic flat with air-dry clay (up&upTM). The sample with four reference points that were evenly spaced on the spring surface was scanned from both top and bottom. The 3D image of the object was reconstructed from these two scans using ATOM Hotfix 6 software. Thermal and mechanical characterization. Differential scanning calorimetry (DSC, PerkinElmer DSC8000) was used to determine transition temperatures of the PEMA on heating from 20°C to 150°C at 10°C/min. The PEMA (2-5 mg) was sealed inside of crimped aluminum pans and blanketed with a dry nitrogen atmosphere during the DSC measurement. The temperature and enthalpy were calibrated with an indium standard. The melting temperatures (Tm) for the polyethylene in the PEMA was defined as the maximum of the melting endotherm, A TA Instruments Q800 dynamic mechanical analyzer (DMA) with a tensile film fixture was used to determine the viscoelastic properties of the filaments, 3D printed samples, and the

ACS Paragon Plus Environment

7

ACS Applied Materials & Interfaces

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 35

compression molded samples at 1 Hz with a 0.01 N preload force and strain amplitude of 1%. The DMA and the PEMA sample were first equilibrated at 0 °C and then the mechanical response in DMA was measured as the temperature was increased to 100 °C at 3°C/min. Shape memory performance. Shape memory cycles (SMCs) were measured with a dynamic mechanical analyzer, DMA (TA Instruments Q800), using a tensile film fixture in force control mode. Sandpaper (3M 400 Grit) was used in contact with the sample to prevent the specimen from slipping in the grips. A SMC consisted of five steps: (1) heating to 70 °C at 10 °C/min and holding isothermally for 3 min, (2) isothermal stretching at 70 °C with a fixed stress to strain the material, (3) cooling to 20 °C at 10 °C/min at constant stress and holding at 20 °C for 3 min to fix the temporary shape, (4) removal of the applied stress and holding at 20 °C for another 3 min, and (5) re-heating to 70 °C to recover the permanent shape. A preload force of 0.01N was used to prevent sagging of the specimen and ensure that all samples were loaded under identical conditions on the tensile fixture. The PEMA had a broad melting region from 40 °C to 110 °C due to a distribution of various size crystallites.15 The fixing temperature was set at 70 °C in this work due to challenges associated with residual stress in the sample (see Figure S1). This temperature has been demonstrated previously to provide good shape memory behavior of PEMA.15 The shape memory performance of the PEMA parts was quantified in terms of the shape fixity (F) and the shape recovery (R). The shape fixity represents the effectiveness of the temporary shape maintaining its dimensions and is defined by eqn (1):    

=

  

× 100%

(1)

where εu is the strain of the temporary shape after removing the stress (end of step 4 of the SMC), εm is the strain after stretching the sample before the stress was removed (end of step 3 of

ACS Paragon Plus Environment

8

Page 9 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

the SMC), εp is the residual strain of the sample at the completion of the SMC (when N=1, the initial strain of the sample at the beginning of the experiment εp(N-1) is equal to zero), and N is the cycle number (for experiments with only a single SMC, N=1). The original, permanent shape of the sample was recovered by reheating the unconstrained sample to 70 °C. The recovery efficiency (R) represents how close the dimensions of the recovered sample are to the dimensions at the start of the SMC as defined by eqn (2):  = 

  

  

× 100%

(2)

For an ideal shape memory material, F = R = 100% where all of the applied strain energy is stored by temporary shape and recovered during the recovery step so that the part returns to its original dimensions. Finite Element Analysis. The geometry was built based on printed parts with 60 % infill density. The printed filament was modeled as a rectangular cross-section with rounded edges with 0.5mm width and 0.3mm height, with a corner radius of 0.075 mm. The gap between two consecutive print lines was 0.3 mm. The stacked layers were offset by 0.05 mm into the adjacent layer to simulate the interpenetration of the layers. The geometry gives a porosity of 57.1%, close to the infill density value. Two patterns were set at 45° and 90° to simulate the printed direction. Then, both the 90° and 45° were meshed into tetrahedral shaped elements with default method for static structural analysis, with the element size set at default. A 100N force was applied to the object for the deformation. The maximum principal stress was calculated using the ANSYS 16.2 Workbench software.

RESULTS AND DISCUSSION

ACS Paragon Plus Environment

9

ACS Applied Materials & Interfaces

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 35

The polyethylene crystals of the PEMA provide the temporary network, so the transition temperature for shape memory is determined by the melting temperature of the crystallites. Figure 1(a) shows the DSC thermograms for the first heating cycle of the PEMA after different processing steps, i.e., the original ionomer pellets, the extruded filament and the 3D printed sample. Irrespective of processing history the PEMA exhibits two broad melting endotherms due to the primary (higher temperature) and secondary (lower temperature) crystallization of the ethylene sequences (see Scheme 1). In the scheme, straight segments indicate PE crystals stems, while red circles represent ionic aggregates. The thinner (smaller) secondary PE crystals are located between bigger and thicker primary PE crystals.17 There is no chemical difference between the primary and secondary PE crystals, but the secondary PE crystals melt at lower temperatures due to their smaller size as a result of the Gibbs-Thompson effect. In the DSC, the peak melting temperature of the primary PE crystallites (T1) for all samples is nearly the same, which is about 96°C. However, the peak temperature of the low endotherm (secondary PE crystallites) is quite different between different samples. This is because the secondary PE crystals can slowly develop during room temperature annealing18. For the resin, it has been sit at room temperature for a while (several month), which let it develop larger secondary crystallites compared with the rapidly cooled samples (filament and 3D printed samples). Therefore, the peak melting temperature of the secondary PE crystallites (T2) for PEMA resin was at 69.4 °C, which is high than the T2 of the filament and 3D printed samples (ca. 49 °C)

ACS Paragon Plus Environment

10

Page 11 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

Figure 1. (a) Impact of processing on the DSC thermograms of PEMA. The melting temperature of the secondary crystals is suppressed in the filament (blue) and 3D printed (red) relative to asreceived resin (black). (b) Temperature dependence of mechanical properties (E’, solid and Tan δ, dashed line) from DMA for the filament (red) and 3D printed (blue) PEMA.

Scheme 1. Schematic of the molecular structure of the PEMA ionomer. The thinner secondary crystals are used here as the second network for fixing the temporary shape.

As shown in Figure 1(b), the thermomechanical behavior measured from DMA shows two decreases in the storage modulus (E’) between 0 – 150 ºC for the filament and 3D printed sample

ACS Paragon Plus Environment

11

ACS Applied Materials & Interfaces

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 35

that correspond to the two melting points detected by DSC. The two decreases in E’ are accompanied by maxima in tan δ. As a result, three rubber-like plateaus exist in E’ as the PEMA was heated. The first plateau, below about 50 °C represents the plastic-like behavior of the unmelted semi-crystalline PEMA. In this case, three types of physical crosslinks are present due to the ionic domains and the primary and secondary polyethylene crystallites. This results in a relatively high modulus of ~500 MPa, which indicates a high crosslink density. The one order of magnitude decrease in E’ between 50 and 65°C corresponds to melting of the secondary crystallites, which lowers the effective crosslink density. From 65 °C - 90 °C, the second plateau is not flat due to continuous melting of the crystals with increasing temperature. From 90 °C – 110 °C, E’ decreases to ~ 1 MPa as the primary crystals melt. Above ca. 110 ºC, where all the polyethylene crystals have melted, the PEMA is a physically crosslinked rubber with a supramolecular network formed by the ionic interactions. The ionic clusters that define the permanent network persist even at high temperatures as determined by the rheological properties (Figure S2). However, the strength of the ionic interaction is softened at high temperatures, which leads to a lossy network (in terms of energy dissipation from shear rheology) at high temperature (> 150 °C) that allows for flow. Melting most of the primary crystals at 100 °C leads to significant shrinkage of the 3D printed part. Figure S3 shows the result of heating a printed single layer of PEMA in boiling water for 2 min where the initial rectangular shape is transformed to a shrunken, curled shape. This deformation is attributed to the residual stress generated during the FDM printing from the rapid cooling during solidification. By maintaining most of the primary crystals in the PEMA, this deformation on heating can be avoided, so a transition temperature of 70°C was selected. In this case, the crystallites that melt below 70 °C (mainly the secondary crystallites) are used as the

ACS Paragon Plus Environment

12

Page 13 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

temporary network for generating shape memory, while the permanent network consisted of two physical networks that are formed by the primary crystallites and the ionic bonds. The initial shape memory cycle (SMC) of 3D printed PEMA sample is compared with the SMC for a compression molded PEMA sample in Figure 2. The exact same shape memory procedure using the DMA was used for both samples. The original length of the sample was ~5 mm, and the sample was heated at 10 °C/min from 20 °C to 70 °C to melt the secondary crystals. The sample was then held at 70 °C for 2 min, after which an engineering stress of 1 MPa was applied and maintained for 3 min to allow the equilibrium strain to be reached. The 1MPa load was retained and the sample was then cooled at 10°C/min to 20°C, where the applied stress was removed to fix the temporary shape. The permanent shape was recovered by heating at 10°C/min to 70°C and annealing for an additional 20 min without any load on the sample. The same SMC procedure was repeated for an additional 4 cycles for each sample to assess the cyclic response.

ACS Paragon Plus Environment

13

ACS Applied Materials & Interfaces

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 35

Figure 2. Initial shape memory cycle of the (a) compression molded and (b) 3D printed (100 % infill density) samples.

For a fixed stress of 1 MPa, the 3D printed sample exhibited a larger strain (εu=40.3%) than the compression molded sample (εu=15.3%), which was due to the higher modulus of the latter. The DSC data shows these two samples have a nearly same degree of crystallinity, which about 20% in total. The large different in modulus was mainly due to the porosity of the 3D printed sample, which leads to large differences in the size of the temporary shape of the two samples. Therefore, to obtain an equal strain for the temporary shapes of the two samples, a lower stress is required for deforming the 3D printed sample to its temporary shape. The shape fixity is surprisingly better for the 3D printed sample (F=90.2%) than the compression molded sample (F=80.9%) despite the larger strain during programming of the former. However, the shape recovery of both the 3D printed sample (R=57.7%) and compression molded sample (R=82.5%) are poor for the initial cycle. This initially poor recovery behavior is common for SMP materials. By repeating the SMC shown in Figure S4, the new permanent shape from after the 1st SMC from the 3D printed sample can be recovered to a much greater extent (98 %) with the shape fixity decreasing slightly (88%) as shown in Table 1. The compression molded sample also exhibits improved recovery, while the fixity remains lower than the 3D printed one for subsequent cycles. The relative invariance in the recovery and fixity for the PEMA after the 1st SMC indicates that there is an irreversible change in the PEMA during the first SMC. In this case, we can use the PEMA crystallization behavior to help to explain the poor recovery on the initial cycle.

ACS Paragon Plus Environment

14

Page 15 of 35

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

Table 1. Comparison of the shape memory performance of compression molded and 3D printed (100 % infill density) samples for multiple shape memory cycles

Cycle No.

Compression molded

3D printed

F%

R%

F%

R%

1

80.9

82.5

90.2

57.7

2

76.0

98.0

88.1

98.3

3

74.4

99.1

88.2

99.3

4

74.7

99.2

88.3

99.5

5

74.3

99.4

88.1

99.5

Figure 3(a) illustrates the thermograms from DSC associated with middle of the compression molded PEMA specimens at different steps of the SMC: initial (black), after programming (red), and after recovery (blue). For the initial sample, there are two broad melting points centered at 48°C (secondary crystallites) and 96°C (primary crystallites), along with a shoulder around 68°C. Any crystal structure that reversibly melts and crystallizes at or below the transition temperature (Ttrans=70 °C) will be used to generate the temporary network. There is a very wide distribution of crystal sizes as evidenced by the broad melting peaks, but the programming step could lead to the growth of crystals to increase the melting point. This is especially true for the initial sample due to its thermal history associated with rapid cooling to lead to arrested crystallization from supercooling. To understand the poor recovery for the initial cycle, the cumulative crystallinity was calculated by integrating the area in the DSC thermograms to determine the melting enthalpy at below and above 70°C (programing temperature) for compression molded sample as shown in Figure S5. This enthalpy is indicative of the relative degree of crystallinity in the sample and the fraction of reversible crystals for the SMC (