@phdthesis{Wang2018,
  author    = {Wang, Li},
  title     = {Reprogrammable, magnetically controlled polymer actuators},
  school      = {Universit{\"a}t Potsdam},
  pages     = {xxviii, 107},
  year      = {2018},
  abstract  = {Polymeric materials, which can perform reversible shape changes after programming, in response to a thermal or electrical stimulation, can serve as (soft) actuating components in devices like artificial muscles, photonics, robotics or sensors. Such polymeric actuators can be realized with hydrogels, liquid crystalline elastomers, electro-active polymers or shape-memory polymers by controlling with stumuli such as heat, light, electrostatic or magnetic field. If the application conditions do not allow the direct heating or electric stimulation of these smart devices, noncontact triggering will be required. Remotely controlled actuation have been reported for liquid crystalline elastomer composites or shape-memory polymer network composites, when a persistent external stress is applied during inductive heating in an alternating magnetic field. However such composites cannot meet the demands of applications requiring remotely controlled free-standing motions of the actuating components. The current thesis investigates, whether a reprogrammable remotely controlled soft actuator can be realized by magneto-sensitive multiphase shape-memory copolymer network composites containing magnetite nanoparticles as magneto-sensitive multivalent netpoints. A central hypothesis was that a magnetically controlled two-way (reversible bidirectional) shape-memory effect in such nanocomposites can be achieved without application of external stress (freestanding), when the required orientation of the crystallizable actuation domains (ADs) can be ensured by an internal skeleton like structure formed by a second crystallizable phase determing the samples´s geometry, while magneto-sensitive iron oxide nanoparticles covalently integrated in the ADs allow remote temperature control. The polymer matrix of these composites should exhibit a phase-segregated morphology mainly composed of cyrstallizable ADs, whereby a second set of higher melting crystallites can take a skeleton like, geometry determining function (geometry determining domains, GDs) after programming of the composite and in this way the orientation of the ADs is established and maintained during actuation. The working principle for the reversible bidirectional movements in the multiphase shape-memory polymer network composite is related to a melting-induced contraction (MIC) during inductive heating and the crystallization induced elongation (CIE) of the oriented ADs during cooling. Finally, the amount of multivalent magnetosensitive netpoints in such a material should be as low as possible to ensure an adequate overall elasticity of the nanocomposite and at the same time a complete melting of both ADs and GDs via inductive heating, which is mandatory for enabling reprogrammability. At first, surface decorated iron oxide nanoparticles were synthesized and investigated. The coprecipitation method was applied to synthesize magnetic nanoparticles (mNPs) based on magnetite with size of 12±3 nm and in a next step a ring-opening polymerization (ROP) was utilized for covalent surface modification of such mNPs with oligo(ϵ-caprolactone) (OCL) or oligo(ω-pentadecalactone) (OPDL) via the "grafting from" approach. A successful coating of mNPs with OCL and OPDL was confirmed by differential scanning calorimetry (DSC) experiments showing melting peaks at 52±1 °C for mNP-OCL and 89±1 °C for mNP-OPDL. It was further explored whether two-layered surface decorated mNPs, can be prepared via a second surface-initiated ROP of mNP-OCL or mNP-OPDL with ω-pentadecalactone or ϵ-caprolactone. The observation of two distinct melting transitions in DSC experiments as well as the increase in molecular weight of the detached coatings determined by GPC and 1H-NMR indicated a successful synthesis of the twolayered nanoparticles mNP-OCL-OPDL and mNP-OPDL-OCL. In contrast TEM micrographs revealed a reduction of the thickness of the polymeric coating on the nanoparticles after the second ROP, indicating that the applied synthesis and purification required further optimization. For evaluating the impact of the dispersion of mNPs within a polymer matrix on the resulting inductive heating capability of composites, plain mNPs as well as OCL coated magnetite nanoparticles (mNP-OCLs) were physically incorporated into crosslinked poly(ε-caprolactone) (PCL) networks. Inductive heating experiments were performed with both networks cPCL/mNP and cPCL/mNP-OCL in an alternating magnetic field (AMF) with a magnetic field strength of H = 30 kA·m-1. Here a bulk temperature of Tbulk = 74±2 °C was achieved for cPCL/mNP-OCL, which was almost 20 °C higher than the melting transition of the PCL-based polymer matrix. In contrast, the composite with plain mNPs could only reach a Tbulk of 48±2 °C, which is not sufficient for a complete melting of all PCL crystallites as required for actuation. The inductive heating capability of a multiphase copolymer nanocomposite network (designed as soft actuators) containing surface decorated mNPs as covalent netpoints was investigated. Such composite was synthesized from star-shaped OCL and OPDL precursors, as well as mNP-OCLs via reaction with HDI. The weight ratio of OPDL and OCL in the starting reaction mixture was 15/85 (wt\%/wt\%) and the amount of iron oxide in the nanocomposite was 4 wt\%. DSC experiments revealed two well separated melting and crystallization peaks confirming the required phase-segregated morphology in the nanocomposite NC-mNP-OCL. TEM images could illustrate a phase-segregated morphology of the polymer matrix on the microlevel with droplet shaped regions attributed to the OPDL domains dispersed in an OCL matrix. The TEM images could further demonstrate that the nanoparticulate netpoints in NC-mNP-OCL were almost homogeneously dispersed within the OCL domains. The tests of the inductive heating capability of the nanocomposites at a magnetic field strength of Hhigh = 11.2 kA·m-1 revealed a achievable plateau surface temperature of Tsurf = 57±1 °C for NC-mNP-OCL recorded by an infrared video camera. An effective heat generation constant (̅P) can be derived from a multi-scale model for the heat generation, which is proportional to the rate of heat generation per unit volume of the sample. NC-mNP-OCL with homogeneously dispersed mNP-OCLs exhibited a ̅P value of 1.04±0.01 K·s- 1 at Hhigh, while at Hreset = 30.0 kA·m-1 a Tsurf of 88±1 °C (where all OPDL related crystallite are molten) and a ̅P value of 1.93±0.02 K·s-1 was obtained indicating a high magnetic heating capability of the composite. The free-standing magnetically-controlled reversible shape-memory effect (mrSME) was explored with originally straight nanocomposite samples programmed by bending to an angle of 180°. By switching the magnetic field on and off the composite sample was allowed to repetitively heat to 60 °C and cool to the ambient temperature. A pronounced mrSME, characterized by changes in bending angle of Δϐrev = 20±3° could be obtained for a composite sample programmed by bending when a magnetic field strength of Hhigh = 11.2 kA·m-1 was applied in a multi-cyclic magnetic bending experiment with 600 heating-cooling cycles it could be shown that the actuation performance did not change with increasing number of test cycles, demonstrating the accuracy and reproducibility of this soft actuator. The degree of actuation as well as the kinetics of the shape changes during heating could be tuned by variation of the magnetic filed strength between Hlow and Hhigh or the magnetic field exposure time. When Hreset = 30.0 kA·m-1 was applied the programmed geometry was erased and the composite sample returned to it´s originally straight shape. The reprogrammability of the nanocomposite actuators was demonstrated by one and the same test specimen first exhibiting reversible angle changes when programmed by bending, secondly reprogrammed to a concertina, which expands upon inductive heating and contracts during cooling and finally reprogrammed to a clip like shape, which closes during cooling and opens when Hhigh was applied. In a next step the applicability of the presented remote controllable shape-memory polymer actuators was demonstrated by repetitive opening and closing of a multiring device prepared from NC-mNP-OCL, which repetitively opens and closes when a alternating magnetic field (Hhigh = 11.2 kA·m-1) was switched on and off. For investigation of the micro- and nanostructural changes related to the actuation of the developed nanocomposite, AFM and WAXS experiments were conducted with programmed nanocomposite samples under cyclic heating and cooling between 25 °C and 60 °C. In AFM experiments the change in the distance (D) between representative droplet-like structures related to the OPDL geometry determining domains was used to calculate the reversible change in D. Here Drev = 3.5±1\% was found for NC-mNP-OCL which was in good agreement with the results of the magneto-mechanical actuation experiments. Finally, the analysis of azimuthal (radial) WAXS scattering profiles could support the oriented crystallization of the OCL actuation domains at 25 °C. In conclusion, the results of this work successfully demonstrated that shape-memory polymer nanocomposites, containing mNPs as magneto-sensitive multifunctional netpoints in a covalently crosslinked multiphase polymer matrix, exhibit magnetically (remotely) controlled actuations upon repetitive exposure to an alternating magnetic field. Furthermore, the (shape) memory of such a nanocomposite can be erased by exposing it to temperatures above the melting temperature of the geometry forming domains, which allows a reprogramming of the actuator. These findings would be relevant for designing novel reprogrammable remotely controllable soft polymeric actuators.},
  language  = {en}
}