Monolithic shape-programmable dielectric liquid crystal elastomer actuators | Science Advances

Soft robotics technology can realize many new technologies for the physical interaction between humans and robots, but the necessary high-performance soft actuators still do not exist. The best soft actuator must be fast, powerful, and have programmable shape changes. In addition, they should have high energy efficiency in unfettered applications and be easy to manufacture. Here, we have combined the ideal characteristics of two completely different active material systems: the fast and efficient drive of dielectric elastomers and the directional shape programmability of liquid crystal elastomers. Through the top-down photo-orientation method, we program the molecular orientation and local huge elastic anisotropy into the liquid crystal elastomer. The linearly driven liquid crystal elastomer monolith achieves a strain rate of more than 120% per second with an energy conversion efficiency of 20% per second, while the moving load exceeds 700 times the weight of the elastomer. Electric actuators provide unprecedented opportunities to achieve miniaturization, shape programmability, efficiency and greater freedom, suitable for soft robots and subsequent fields.

In traditional robotics technology, the basic rigid actuation mechanism of electric motors or hydraulic and pneumatic actuators hinders the miniaturization of robots, and more importantly, hinders the use of robots in human collaboration environments. The compatible actuator is to make the robot and the man-machine interface (

). The ideal compliant actuator will have high efficiency, strength-to-weight ratio, work capacity and shape programmability to perform complex functions. Like artificial muscles, soft actuators with these characteristics will greatly promote technology in aerospace, robotics, medical equipment, energy harvesting equipment and wearable devices (

-

). Among the many soft actuators that have been explored, the dielectric elastomer (DE) seems promising in some respects, even surpassing skeletal muscle (

). In addition, the liquid crystal elastomer (LCE) exhibits reversible large mechanical deformation (

,

). The latest advances in optical alignment and microfabrication technologies have enabled the pre-programming of liquid crystal alignment in microscopic areas to achieve complex shape deformations (

). However, both actuators have their shortcomings: DE membrane needs to be macroscopically pre-strained (

) Or requires a multi-step manufacturing method, which makes it difficult to program small actuators with local shape changes (

). At the same time, due to the small strain generated, the direct conversion of electrical energy to mechanical work using LCE is still limited (

). However, we proved the ability to pattern the LCE molecules in a locally varying arrangement, thereby adjusting the spatial variation of mechanical compliance, so that a more efficient DE driver can be realized with a pre-programmed drive degree and direction.

Generally, the DE actuator forms a variable resistor-capacitor through electrostatic attraction between two compliant electrodes coated on opposite sides of the isotropic DE (

) (

). The high voltage applied to the compliant electrode induces electrostatic pressure, the so-called Maxwell stress, which deforms the DE. Compared with LCE, electric actuation mechanism can produce higher working efficiency (ratio of mechanical work to input electric energy) and higher actuation speed (

). In addition to acting as a soft linear actuator, DE actuators can also be applied to fixtures, haptic devices or optical devices, but they require complex shape changes (

). Despite some impressive demonstrations, DE actuators have not yet been widely used in soft robots, partly because of the need for pre-straining or the challenges of manufacturing equipment with complex deformation profiles (

). Overcoming these challenges and expanding the application range of DE actuators requires material innovation for the next generation of high-performance DE with shape programmability (

).

(

) Schematic diagram of traditional isotropic DE actuator in closed and open state. (

) Schematic diagram of a single-axis aligned dielectric LCE actuator (DLCEA) in closed and open states. Liquid crystal molecular orientation; director

Use double-headed arrows to indicate the harder direction of the LCE. When driven by voltage,

, The material will become thinner and stretch perpendicular to the alignment direction, greater than the direction parallel to the director. (

) DLCEA mechanical stress and normalized capacitance (

) The strain response in the linear range of DLCEA is characterized by a strain rate of 0.1% per second.

LCE is a polymer with rubber elasticity, and its molecular anisotropy gives it anisotropic overall properties. Most previous work on LCE actuation has focused on thermal or light drive mechanisms. Heat or light will temporarily disrupt the order of anisotropic molecules, called the guiding field (

), resulting in internal stress and anisotropic overall deformation (

). The local LCE director field can be pre-programmed to create complex shape changes at startup. However, light actuation is inefficient, while thermal actuation is slow and inefficient. Therefore, they are not suitable for applications that require high energy efficiency and fast startup, such as robotics. The direct electric drive of LCE is a highly sought after technology (

). Some previous studies have demonstrated that by coupling the electric field to the molecular dielectric anisotropy or sometimes to the intrinsic polarization of the LCE or LCE composite, the electrical drive of the LCE can be achieved. Therefore, the electric field drives the molecules to reorient to produce overall strain. However, these methods require high temperatures or the use of carbon nanotubes to enhance electrical response. Otherwise, only a small driving strain will be generated at room temperature (

). In this work, we directly use the large mechanical anisotropy of LCE without relying on molecular rotation in the electric process. We further use the latest developments in the patterning of LCE films to customize local anisotropic elasticity and Poisson's ratio to achieve an efficient and shape-programmable DE, which we call a dielectric LCE actuator (DLCEA;

). By arranging the LCE molecules in the local domain, we achieved electric drive and shape deformation at room temperature, and showed large, rapid and powerful strain.

LCE film is manufactured by a two-step process recently developed by some authors (

). In short, before preparing the LCE film, oligomers were synthesized through the thiol acrylate click reaction. Ordinary diacrylate reactive liquid crystal monomers are chain extended by Michael addition of dithiol linking molecules. The exact component ratio, monomer selection and dithiol linking group can all be adjusted to adjust the specific mechanical properties of the final LCE film (

). We have produced a large area of ​​ordered uniaxial LCE (Figures S1 and S2), and have huge elastic anisotropy (

). In all experiments, we only activated DLCEA at room temperature in the linear state of the strain (Figure S3). We can also program the LCE director field locally through optical alignment to create a command surface for spatial programming and locate the LCE director locally (

). Finally, we used compatible grease electrodes on both sides of the LCE membrane to create the DLCEA device (Figure S4). Further details can be found in materials and methods.

In order to characterize the basic characteristics of DLCEA, we first produced a single-domain, uniaxially arranged LCE film. The electrode coated on the uniaxial DLCEA can simultaneously measure the capacitance and the stress and strain applied to the LCE film (

). When strain (

) Parallel application with the director,

, Compared with when strain is applied perpendicular to the director, the rigidity of LCE is one order of magnitude higher.

⊥

, Which shows a high degree of elastic anisotropy. Similarly, the difference in the slope of the normalized capacitance between DLCEA devices with different director directions indicates the anisotropy of Poisson's ratio. DLCEA capacitance is directly proportional to the area of ​​the film covered by the electrode and inversely proportional to the film thickness. Therefore, Poisson's ratio anisotropy causes the thickness and area of ​​the film to change at different rates according to the strain direction relative to the LCE director field (Figure S5A). Using a simplified finite element model, we found that for a large elastic anisotropy, the linear expansion strain produced by a given Maxwell stress is almost twice the linear expansion strain observed in an isotropic material, while the elastic The expansion modulus is equal to the soft direction (Figure S5B). Other works with similar chemical properties of LCE have also observed one-dimensional (1D) translational crystallinity, which can explain the particularly large elastic anisotropy observed in this work (

Then, we feature isometric drawings (constant strain; uniaxial DLCEA).

) And isotonic (constant force;

) Configuration. In the isometric test, we applied an initial strain to the DLCEA device and performed a relaxation period before applying high pressure (Figure S6A). Through the reduction of active stress, we observe two relationships between strain, applied voltage and active stress, which are consistent with Maxwell’s stress model.

∝

/

,where is it

Is the applied voltage,

Is the thickness of the LCE film. First, driving with a larger initial strain will produce a higher active nominal stress reduction; the isometric pre-strain will cause the LCE to become thinner, so at a given voltage, a higher Maxwell stress (

). For LCE, it seems that the speed at which the additional strain causes the material to thin is sufficient to offset the increased restoring force, so that a given drive voltage will result in greater activity stress. We also observed that for each fixed strain, the effective nominal stress reduction during isometric measurement increases squarely with the increase in voltage (Figure S6B). At the highest voltage tested, we measured the peak effective nominal stress reduction by more than 50 kPa. However, for devices with directors,

Since the modulus is much higher, the active stress is relatively small. When maintaining isometric strain, DLCEA behaves like a spring with variable stiffness. In this case,

, The initial strain of 5% and the driving voltage of 2 kV, the LCE expansion caused by Maxwell stress almost compensates the stress caused by the entire equidistant strain. We also performed an equipotential test in which DLCEA was strained under a constant voltage. These tests show the expected drive stroke of the DLCEA load when voltage is applied (Figure S7).

) Isometric (constant strain) test. Active nominal stress reduction measured under various initial equidistant strains (

) For equipment assembled with LCE Director

with

‖

And photos of the assembled DLCEA equipment,

. (

Isotonic (constant force) test. The contraction discharge strain trajectory of a high-speed camera with an actuation voltage of 3 kV under various loads. Illustration: Corresponding discharge measurement value. (

) The basic actuator characteristics are based on the contraction trajectory and the discharge current (including strain (

), peak strain rate (

), peak specific power (

), specific energy (

) And efficiency. Image source: Zoey S. Davidson.

Next, we use the same DLCEA and suspend different weights from the free end of DLCEA to generate a constant load force and initial nominal strain for isotonic testing.

(

And figure. S8). DLCEA in

Even at the highest voltage tested, this configuration did not show any appreciable active strain at initial load because their elastic modulus was significantly higher (Movie S1). However, DLCEA and

Under the heaviest load tested, the applied voltage of 3 kV was 0.27 N, which was approximately 790 times the weight of the bare LCE film (35 mg). This configuration showed up to 5% rapid active strain. We performed the isotonic contraction test by suddenly releasing the weighted DLCEA device, and used high-speed video (Figure S8B and Movie S2) to capture the subsequent motion. As the load and initial strain increase, the DLCEA capacitance also increases, which can be seen from the offset discharge curve in the insert diagram.

. In all cases, the discharge time within about 60 ms (approximately 1 ms) is much faster than the discharge time required for DLCEA contraction, which indicates that the system is currently limited by the viscoelasticity of LCE. We also observed a significant viscosity loss in contraction, which is evident from the continuous creep contraction after the initial elastic response (Figure S8B). From the contraction trajectory of DLCEA, we can calculate the basic performance indicators of pure elastic response and expansion creep contraction (

And figure. S8C).

In order to achieve complex shape driving, LCE usually functions by programming the spatially varying in-plane shrinkage strain when heated above the phase change. However, the DE drive mechanism is not based on a thermally induced phase change, but generates an in-plane expansion strain. Therefore, the boundary conditions play an important role in determining the realized DLCEA shape change. In order to better understand the effect of boundary conditions on DLCEA, we have performed a basic characterization of the buckling effect caused by the expansion of the elastic body between the fixed boundary (

, And the movie S3). The buckling amplitude increases with the increase of the voltage, and at 2.5 kV, an out-of-plane peak-to-peak stroke of 1800% of the LCE film thickness greater than about 80 μm is generated, corresponding to a linear strain of about 5% (

And figure. S9A). Actuation speed is another important characteristic of potential DLCEA applications. We applied a sinusoidally varying 1-kV potential to measure the change in drive amplitude with the applied frequency (

And figure. S9B). The excitation amplitude decays exponentially with frequency, but 50μm can still be sensed at 30 Hz and 1 kV.

)with(

) The state of a single-axis DLCEA device with fixed boundary conditions. Expansion in the soft direction will produce out-of-plane bending, thereby displacing the thin lines tightened on the surface. (

) The relationship between the experimental measurement of buckling and the applied voltage. (

The frequency response of 1kV buckling uniaxial DLCEA is 0.1 Hz and the excitation amplitude is about 130μm.

We designed a spatially variable LCE director configuration to prove that it is possible to pre-program complex patterns in 2D and then electrically drive the film into a 3D form (

). According to the programmed director area, the LCE film will be partially positive (

) Or negative Gaussian curvature (

). These shapes are usually called conical and anti-conical. The theory describing this form of deformation in elastic media was previously described in Modes.

). We understand these shape changes by considering a simplified model of an anisotropic DE made of rigid concentric rings embedded in a soft elastomer (

). These rings prevent expansion along the ring, but allow expansion in the radial direction, resulting in frustration and out-of-plane bending. Similar arguments apply to radial rigid elements. Double-headed red arrow in

Indicates the direction of soft expansion. We create a pixelated array of topological defects by spatially programming the light polarization using the pattern of linear thin-film polarizers (

), positioning the LCE supervisor locally as

. Directors form a lattice of radial and azimuthal defect types. When electrically driven, they will bend out of the plane due to incompatible in-plane strain (

And movie S4). We measured discharge (0 V) and start (2.5 kV;

) Status. For this reason, we kept the device in its active shape for more than 3 hours under a voltage of 2.5 kV, while the current consumption was less than 1μA, thus confirming its high stability and low power consumption. From the circular trajectory around the center of the radial defect type, the height change of the local programming and the accompanying Gaussian curvature formation (

). Out-of-plane bending will produce a peak-to-peak height difference of more than 1600 μm, which is an increase of 2000% compared with the initial film thickness of about 80 μm, corresponding to a surface strain of 22%. Bottom right corner of DLCEA enabled

Indicates that the defect may bend up and down.

By patterning the pattern of the director into an azimuth-radial defect lattice, it is possible to perform programmed shape driving, such as dent pattern deformation. (

) The azimuth defect type is deformed into a cone with local positive Gaussian curvature, and (

) The radial defect type deforms into an inverse cone with locally negative (saddle) Gaussian curvature. In (A) and (B), the two-way red arrow indicates the soft direction. (

) Use a polarizing film pixelated array with a designed local direction to pattern defects. (

Observed by crossed polarizers, the finished LCE film has a pixelated uniaxial arrangement (indicated by the white dotted line), forming a defective lattice. (

) When charged to 2.5 kV, there will be a lot of visible deformation on the surface. (

) The height map measured by the grease-coated LCE profiler is almost flat, there is no charge, and the change exceeds 1.6 mm when charged to 2.5 kV. The two-dot-dash line and the dot-dash circle in (F) are (

). The change from an approximately constant height to a sinusoidally varying height represents a sign change of the local Gaussian curvature. Scale bar, 4 mm. Image source: Zoey S. Davidson.

When using DLCEA as a linear actuator, compared with similar isotropic materials, we expect the anisotropy of Poisson's ratio to be larger to produce higher efficiency and require a lower electric field for actuation (

). Compared with other LCE actuators, the execution efficiency reported here is about 20%. As far as we know, due to the low energy conversion efficiency, the actuation efficiency of LCE has not been reported. For example, according to our estimation, the actuation efficiency of thermally induced LCE is lower than 0.001%, according to (

). Note that this estimation is based only on the initial stroke; a constant current is required to keep the LCE in the contracted state. In addition, our DLCEA efficiency is comparable to the latest examples of isotropic DE actuators with highly optimized electrodes, with a reported efficiency of 1.5% (

). We expect to reduce viscosity loss and creep, which is represented here by the hysteresis loop.

And prolong the contraction

, Will further improve the fast and efficient DLCEA situation.

We believe that the high efficiency reported in our system is due to the anisotropy of elastic modulus and Poisson's ratio. Generally speaking, elastomers save volume. Therefore, extension in one direction results in contraction in the other direction. However, the shrinkage of the LCE film is perpendicular to the director. In other words, when the LCE film is tensioned perpendicular to the director, its thickness shrinks. The thickness is also perpendicular to the director, so it shrinks faster than the width parallel to the director. When parallel to the deflector strain, the thickness and width of the LCE are equal (assuming that the cross section is isotropic). It is worth mentioning that the last point is still uncertain: as mentioned earlier, in this type of LCE, the crystallinity of one-dimensional translation is very common (

). In particular, the one-dimensional crystal plane may be at a certain angle to the LCE director, which breaks the symmetry, so that when strain is parallel to the director, the width and thickness are assumed to shrink equally. Adjusting this one-dimensional crystallinity may play an important role in further improving the linear driving capability of DLCEA.

In order to further clarify the advantages of elastic anisotropy, we consider the use of a simplified DE model of an approximate volume-saving elastomer model with approximate Poisson's ratio anisotropy and large Poisson's ratio anisotropy under no load (see figure S5). In this model, almost all compressive strains caused by Maxwell stress will produce tensile strains in the soft direction of the elastomer. In other words, Maxwell strain through the thickness of the material,

=

, Causing strain

And almost no pressure

. In an isotropic elastomer, the same Maxwell strain will only produce half of the strain, because the volume conservation strain will be evenly divided into

. For linear actuators, this is the first advantage of DLCEA. The excitation voltage required to reach a given strain is reduced. The second advantage of the anisotropy of linear actuators comes from the energy considerations of the same system. The elastic energy density of deformation is squared in strain. Therefore, in the simplified model presented here, there will be no energy component from strain.

direction. In addition, for a given required linear extension strain, the input electric field energy (∝

) Will also be less, because the required Maxwell strain is less than the strain in the isotropic DE. Therefore, the anisotropic DE actuator can achieve the same strain as the isotropic DE linear actuator, but with higher efficiency. Both no-load and ideal uniaxial elastic body assumptions can be relaxed, and viscoelastic effects can be added to build a more complete model.

The anisotropy of material Poisson's ratio is also an important feature for realizing programmed shape change drive. The driven (compressed) LCE expands laterally anisotropically to produce the observed shape change. Although in principle the curved shape of DLCEA is multi-stable, we only observe a single driving state for each sample (

). We hypothesized that the gravity during the test or the uneven photocrosslinking of the LCE film during the device manufacturing process may destroy the symmetry that achieves multiple stability. Nevertheless, our demonstration of local changes in Gaussian curvature shows that our method can potentially be extended to achieve various programmable shape changes (

). In addition to programming the direction of the director in the plane, the direction of the LCE director can also be programmed along the film thickness. as the picture shows. For S10, we assembled DLCEA with a twisted LCE configuration, where the guide rotates nearly 90° from the top surface to the bottom surface. When an electric field is applied, the twisted DLCEA will produce a twisting motion, the amplitude of which depends on the inherent characteristics of the material, but also on the geometry of the LCE (

Here, by combining the ideal characteristics of DE and LCE in a single material platform, we demonstrated the excellent driving performance of the electric drive DLCEA, including high energy conversion efficiency (20%), high driving speed (120% per second) and programmable The shape changes from 2D to 3D, and the out-of-plane stroke exceeds 1800%. In order to obtain greater driving force, you can choose a multilayer DLCEA stack, as shown in the LCE and DE multilayer stack (

), although this requires the development of alternative soft electrodes. In addition, even more general shape changes, ie non-local Gaussian curvatures, can be achieved by spatially programming LCE alignment and local crosslink density.

When DLCEA is combined with 3D printing, origami and origami drive strategies, and distributed control systems to create a multifunctional soft robot in a scalable manner, the active materials and top-down microfabrication technology and electric drive introduced here The insights of the integration of mechanisms may provide exciting opportunities. Material and construction costs are low. Electric drive mechanisms can also be applied to other technologies, including energy harvesting and storage, medical equipment, wearable technology, and aerospace. In addition, fast and dynamic modulation can be useful in display and optical applications.

1,5-Pentanedithiol (1,5-PDT; >99%), 1,8-diazabicycloundecaheptene (DBU), butylated hydroxytoluene (BHT) and magnesium sulfate (MgSO)

; Anhydrous powder) purchased from Sigma-Aldrich and used as is. Hydrochloric acid (HCl), dimethylformamide (DMF) and dichloromethane (DCM) were purchased from Fischer Scientific. The photoinitiator 2,2-dimethoxy-2-phenylacetophenone (DMPA) was purchased from Toronto Research Chemicals. Brilliant Yellow (BY) was purchased from Tokyo Chemical Industry. Liquid crystal monomer 1,4-bis-[4-(4-(6-acryloxy-hexoxyoxy)benzoyloxy]-2-methylbenzene (RM82; >95%) purchased from Wilshire Technologies Inc. can be used without further purification. Conductive carbon grease NyoGel 756G was purchased from Newgate Simms.

We made LCE film through a two-step process recently developed by some authors (

). Before preparing the LCE film, oligomers were synthesized through the thiol-acrylate click reaction. The reactive liquid crystal monomer RM82 is chain extended by Michael addition with 1,5-PDT. In a typical synthesis, 12.5 g RM82 was mixed with 5.06 g 1,5-PDT in 120 ml DCM and three drops of DBU catalyst. After stirring at room temperature for 16 hours, the solution was washed sequentially with 1M HCl, 0.1M HCl and deionized water in a separatory funnel. The DCM product mixture was then dried with 25g MgSO 4

30 minutes, then filter. Before rotary evaporation and direct vacuum, BHT (50 mg) was added to the clear DCM and product mixture until a thick white oligomer remained. The oligomer can be stored at -30°C for up to 2 months.

Deionized water, isopropanol and acetone are usually used in an ultrasonic bath to clean glass slides that are usually 5 cm x 5 cm and 8 cm x 10 cm. Next, the glass slide was dried with nitrogen gas and then treated with oxygen plasma. A mixture of 1 wt% BY dissolved in DMF was spin-coated on a glass slide, and then dried on a hot plate at 120°C. Place a gasket with a thickness of 65 or 75μm, cut from polyimide or polyester film plastic along the edge of the BY coating side of the glass sheet; then, place two slides so that they are coated with BY The coated faces face each other. The large paper clip holds the slide with the polarizing film placed on one side. A customized 447 nm light-emitting diode (LED) light source is used to illuminate the BY-coated glass through a polarizing film, thereby optically programming the orientation of the BY molecules. In order to program the locally varying Gaussian curvature, the polarizer was cut into pixels and then reassembled by hand on a glass slide with the desired orientation (see

). The thin layer of BY molecules is rearranged perpendicular to the incident light polarization to create a spatial light programming command surface, and then the LCE director (

The previously prepared oligomer was melted with additional RM82 and a small amount of photoinitiator to crosslink the oligomer chain into the elastomer network. In more detail, assuming that the oligomer is only composed of a single-unit length RM82 chain with both ends blocked by 1,5-PDT, the LCE oligomer and other RM82 LCE monomers are melted at a molar ratio of 1:1. together(

). Therefore, the mixture is composed of an excess of thiol groups, which may be a large part of the final LCE viscosity loss (see the uniaxial alignment DLCEA characteristics section), but this sparse cross-linking also facilitates the need for greater driving strain The softness. The melt was mixed at 120°C for only 2-3 minutes, and then degassed in a vacuum oven at 90°C for about 3 minutes. Add 1% by weight of DMPA and stir carefully to avoid introducing bubbles again.

Then the isotropic LCE melt was poured onto the BY-coated glass at 80°C, and then carefully clamped on the second hot BY-coated glass substrate. The BY-LCE-BY sandwich is cooled to an oriented (nematic) phase of about 73°C, and then gradually cooled to room temperature, during which time it is consistent with the spatial programming imparted by the BY coating, and the defects are annealed from the phase change. Once the LCE has cooled to room temperature, it can be cured under UV light using the OmniCure S2000 arc source. After being exposed to ultraviolet rays to polymerize LCE in a programmed state, we immersed the BY-LCE-BY sandwich panel in water to release LCE from the BY-coated glass substrate.

The final LCE film thickness (described below) was confirmed from the area cut into the actuator by confocal laser profilometry. Good alignment of LCE and almost no defects are the basic characteristics of the film, which can give maximum elastic anisotropy and obtain the best material properties. The high contrast between the LCE directions between the crossed polarizers can be seen in Figure 5. S1.

After separating the LCE sheets from the glass substrate, they were rinsed with water to remove residual BY, and dried with nitrogen. Put the LCE sheet back on the glass substrate and carefully inspect to identify defects and bubble-free areas to manufacture DLCEA devices. For the single-axis DLCEA device, the most clearly identified area is cut into rectangular slices usually 14 mm by 34 mm, with a typical weight of 35 mg. The membrane of this size was chosen for ease of handling and the electrical actuation constraints described below. The smaller adjacent area (20mm by 5mm) was originally used to characterize the stress-strain behavior and the larger strain behavior of the LCE.

Then check the edge of the laser cut area on the larger film with a laser confocal interferometer (KEYENCE VK-X210) to confirm the processing height of the LCE film. We found that Kapton of nominal 65 microns can produce LCE film of approximately 70 microns, and Mylar of nominal 75 microns can produce LCE film of approximately 83 microns. The thickness of the entire produced LCE board may vary by ±10% (Figure S2).

In the next step of DLCEA manufacturing (Figure S4), we use conductive carbon grease NyoGel 756G to connect the compliant electrodes to both sides of the LCE membrane. This grease is often used in other DE systems (

). In order to apply the carbon grease, the LCE was first clamped in a 3D printed plastic clip with copper tape leads designed to facilitate the connection of the equipment to the test equipment described below and the "Uniaxial Alignment DLCEA Characteristics" section. A certain degree of misalignment during clamping is inevitable. The tailored LCE is fixed in a laser-cut plexiglass assembly jig and covered with a low-viscosity removable tape, which is placed around the edge of the LCE film. The masking tape forms a boundary area on the edge of the LCE without electrode grease, which prevents short circuits when driving the device under high voltage. It was found that a gap of 2mm around the edge was sufficient to prevent a short circuit under the test voltage (see

). Use a cotton swab applicator to apply grease, and use a straight edge to remove excess grease. Before and after using the grease electrode, weigh the entire Plexiglas fixture with LCE film to find the grease weight. For the two electrodes of DLCEA, the grease weight usually totals 30 mg. Other high-conductivity electrode materials can achieve better performance while adding smaller weight and cross-sectional area (

); Future research on these actuators will study alternative electrode materials.

Throughout the work process, we only tested and activated DLCEA at room temperature under linear conditions where the strain did not cause the LCE guide to reorient. Usually, we find that the onset time of soft mold deformation (pointing to reorientation) is 45% to 50%

As shown. S3.

We performed mechanical and electrical characterization of the laser-cut single-axis DLCEA (

And figure. S2). The tensile test was performed in TA Instruments DHR3, and the capacitance measurement was performed simultaneously using a Hameg 8118 LCR meter. Generally, a uniaxial DLCEA made of 65 μm spacers and electrodes with an area of ​​1 cm x 3 cm coated on both sides has a zero strain capacitance of about 300 pF. We observe the correlation between the capacitance growth rate and the tensile strain direction of the LCE film relative to the director. The capacitance of DLCEA with strain perpendicular to the director grows faster than the capacitance of DLCEA with strain parallel to the director. We can model how strain affects the DLCEA capacitance.

The capacitance of the parallel plate capacitor (or DLCEA) is

ϵ

Is the permittivity of free space, ϵ

Is the relative dielectric constant perpendicular to the liquid crystal director (note that the reactive mesogen RM82 used in this work has negative dielectric anisotropy, ie

> ϵ

). The rectangular area covered by the electrode is

X

, The film thickness is

(For the schematic and coordinate system, please refer to Figure S5). When DLCEA was

(Perpendicular to the director), the thickness decreases

=-

And along the width

cut back

=-

. The thickness and area become (1 +

)

=(1 −

And (1 +

(1 +

=(1 +

) (1 −

, Respectively. Therefore, the capacitance becomes

Next, we normalize by the capacitance under zero strain, and then Taylor expands for small strain, that is, only the linear term

Insert this equation from the relationship between the symmetry of the system and the mechanical anisotropy,

/ν

And assuming

= 0.5, we get

≈0.04 and

≈0.84. Take the modulus of elasticity (

),

, The stiffness tensor is fully defined. These values ​​indicate that LCE is unexpectedly compressible. However, this is impossible, and due to at least the following three reasons: Capacitance

At 20% strain, the coefficient decreases from 32 to 22, and it will inevitably be slightly pre-strained when measuring the modulus.

, Namely LCE. These are in addition to the possibility of partial crystalline order (smectic C phase) mentioned in this article. These factors together lead to errors that may lead to apparent compressibility.

Equidistant testing is performed by quasi-statically increasing the voltage applied to the pre-strained sample. After the capacitance measurement, strain the DLCEA still fixed in the rheometer to a fixed amount (5%, 10%, 15%, and 20%), and then relax for a period of time until the measured stress creep is much smaller than the induced stress (Figure S6A). Starting from 500 V, the driving voltage (Heinzinger LNC-10 kV) increases by 100 V every 15 s. Samples are taken from the middle 5 s of each cycle to measure the active change in stress due to the applied voltage. Following the relationship given by Maxwell's stress equation, the logarithmic effective nominal stress reduction for all equidistant strains has a slope of 2.0 to the logarithmic voltage relationship (Figure S6B).

The equipotential test is done by first straining the DLCEA without applying a voltage, and then applying a voltage of 2 kV (Figure S7). The difference in induced stress between the 0 kV and 2 kV curves represents the expected stroke when DLCEA is used as an actuator under constant load.

In order to characterize the basic characteristics of LCE as a muscle-like actuator, we tested the DLCEA tightened by a constant gravity load.

. The weight suspended by the DLCEA causes the initial strain that makes the material thinner, which contributes to a greater driving force and thus a higher initial load. When voltage is applied to DLCEA

⊥

As the elastic response of LCE has changed, the system adopts a new length. The LCE hardens, so the weight stops when the forces are balanced. However, after the initial elastic response, DLCEA continues to creep due to the viscoelasticity of LCE. The strain gradually increases until it finally reaches a steady state. After a period of time, a short-circuit path is provided to the electrodes of DLCEA through a custom switching mechanism. Therefore, DLCEA is discharged and suddenly contracted elastically, and then continued to contract further slowly due to viscoelasticity (

And figure. S8). if

, Due to the significantly higher stiffness (movie S1), there is no significant drive along the loading direction; therefore, no further testing of this DLCEA configuration was performed.

At the same time as the startup, a high-speed camera (Vision Research v641) was manually triggered. For shrinking data

, The camera captures at 1400 frames per second. The video frame for switching high voltage is identified by a pair of LEDs triggered by the same solid state relay as the high voltage switch. Use an oscilloscope (Tektronix MDO4024C) to measure the discharge current on a resistance divider pair connected in series with DLCEA. A schematic diagram of the high-voltage switching mechanism that measures the discharge current by reading the voltage,

On known resistance

,as the picture shows. S8A. The drive also depends on the applied voltage. For each load, voltages of 2, 2.5 and 3 kV were tested on the same DLCEA (Figure S8C).

Use Tracker Video Analysis (Tracker Video Analysis (

), and then use a custom Python script for analysis. In these tests, the high voltage was switched on for about 20 s before discharging, so that DLCEA reached its active, stable resting length. Mark the initial distance manually in Tracker Analysis, then compare it with the known component size to calculate the distance, and then calculate the energy, power, and efficiency. The oscilloscope data was also analyzed using custom Python code. By measuring the discharge with the switch not connected to DLCEA, the baseline capacitance charge can be subtracted from the measured discharge. The capacitance of the high-voltage cable per meter is about 100 pF.

The mechanical work done by the actuator during discharge is calculated based on the mass of the additional load.

, And the displacement found through the high-speed video, namely

Δ

Is the acceleration due to gravity, 9.8 m/s

And Δ

Is the displacement of mass. Power into the system,

Discovered by integrating the discharge current measured as voltage

, Through the known resistance,

, And then multiply by the applied voltage (for example, 3 kV). Finally, efficiency is calculated based on the ratio of these energies,

.

The finite element simulation was carried out using the structural mechanics module of COMSOL Multiphysics 5.3a (COMSOL, 2008). Several mesh refinement steps were performed to ensure convergence of the results. For the no-load simulation of DLCEA (Figure S5B), the LCE film was modeled as an anisotropic thin plate (width 14 mm; length 30 mm) with an initial thickness of 80 μm. Use five independent elastic constants to calculate the flexibility and stiffness matrix of an anisotropic material using Voigt notation

Assuming "1" is

Nematic guide

In order to demonstrate the effect of anisotropy on the performance of the actuator, we scanned

From 1 to 20 MPa, while assuming Poisson's ratio

= 0.5 and ν

= 0.9-ν

And shear modulus

(Assuming that Young's modulus and bulk modulus are equal). The geometric boundary condition is defined as clamping on one side. By considering the electrode margin as 2 mm, a normal pressure load (100 kPa representing Maxwell stress) was applied to the top side of the LCE plate, and the roller boundary conditions were set to the bottom side of the LCE. A constant force is applied to the free edge of the LCE beam (as opposed to the clamping edge) to induce deflection and simulate gravity load.

In Petsch's work

), insert the filament heater into the aligned LCE. When heated, the device shrinks in its alignment direction. The 90% contraction reaction time reported in this work is 20 to 30 s. In the example reported in this article, they achieved a 1.85 mm stroke with a test load of 2.25 g and an input power of 430 mW. Then the stroke efficiency is

Or about 0.0005% for strokes only. In order to maintain this stroke, a constant current must be applied.

The edge of the uniaxial LCE film is constrained by a laser-cut plexiglass frame. Place the film carefully on top of the frame so as not to cause prestress or leave any slack. Coat the center square carbon grease electrode on both sides of the film through a low-viscosity removable tape mask. The in-plane length of the film grows in the soft direction, but due to the fixed boundary conditions, curved wrinkles are generated. The height of the wrinkle pattern is measured by the laser confocal profiler in the off state and at every 250 V voltage from 500 V to 2.5 kV. In the 2.5kV activation state, the out-of-plane peak-to-peak travel is 1.47 mm or 1800% of the LCE film thickness (approximately 80 μm).

In order to determine the frequency response of the uniaxial buckling DLCEA, we applied the sinusoidally varying 1 kV provided by the Physik Instrumente E-107 piezoelectric high-voltage amplifier. The input signal is generated by a function generator (Tektronix). Observe the movement of the DLCEA film with a Thorlabs Telesto optical coherence tomography microscope. First, manually find the maximum height of the DLCEA film in the DC on state, and then observe various frequencies in the same position (

And figure. S9B).

The defect array is achieved by programming the laser-cut square light polarization of a linear polarizing film that is sewn back into the desired grid on the glass slide using NOA65 ultraviolet curing glue. Due to the defects in the laser cutting step and the difficulty of manual stitching, the stitched polarizers are not perfect with each other. However, the misalignment boundary between the aligned regions in the LCE film is small and obviously does not affect the drive response.

Similar to uniaxial buckling, after manufacturing the LCE and removing it from the BY-coated glass slide, we fixed the LCE film on the laser-cut plexiglass frame and covered the programming area with conductive carbon grease. Due to the mismatch in refractive index between the isotropic and well-aligned areas, it is difficult to distinguish the programming area from the surrounding area under ambient lighting. The curved shape of the film is measured by the KEYENCE laser profiler (VK-X210), and the 10x objective lens is set to ultra-high speed

Scan at a pitch of 4μm. The observation area is a stitched image of many individual images.

For supplementary materials for this article, please visit:

Figure S1 Optical characterization of a uniaxially arranged LCE.

Figure S2 Photograph of LCE laser cutting area and measured height.

Figure S3 Stress and strain characterization of uniaxial LCE.

Figure S4. The assembly process of a typical single-axis DLCEA construction.

Figure S5 Schematic diagram of DLCEA, with coordinate axes and simulation results.

Figure S6. Isometric uniaxial DLCEA relaxation and log-log stress-voltage relationship.

Figure S7 Equipotential test of single axis DLCEA.

Figure S8. A schematic diagram of the high-voltage switching mechanism, isotonic full-cycle actuation, and isotonic actuation characteristics that vary with voltage.

Figure S9. DLCEA uniaxial buckling voltage and frequency response.

Figure S10. Drive for twisted configuration of LCE film.

Movie S1. Single axis DLCEA, director parallel to

Movie S2. Single axis DLCEA, director perpendicular to

Movie S3. Demonstration of uniaxial buckling DLCEA.

Movie S4. Demonstrates programmable shape change buckling DLCEA.

This is an open access article distributed under the following terms

, It allows use, distribution and reproduction in any medium, as long as the final use is

For commercial interest, and provide the original works appropriately cited.

Volume 5, Issue 11

November 01, 2019

Thank you for your interest in promoting the term "scientific progress".

Note: We only ask you to provide your email address so that the person you recommend to the page knows that you want them to see the email and that it is not spam. We will not capture any email addresses.

This question is used to test whether you are a visitor and prevent automatic spam submission.

Please log in to add an alert to this article.

by

: Eaay0855

Maxwell's stress-activated shape-deformable liquid crystal elastomer actuator is efficient, fast and powerful.

Volume 371, Issue 6526

©2021

. all rights reserved. The American Association for the Advancement of Science is 

, 

 with 

ISSN 2375-2548.