Core-free rolled actuators for Braille displays using P(VDF–TrFE–CFE)

Figure 1 shows a schematic diagram of the P(VDF–TrFE–CFE) cylindrical actuators developed in this paper. They consist of an electroded bilayer film that is rolled into a tube. An external electric field causes the film to contract through the thickness and extend longitudinally in the actuation direction. The positive and negative electrodes are offset, allowing an electrical contact at each end of the actuator. The number of layers, diameters, and length of the actuator are designed to avoid buckling under maximum blocking force. This means that the actuator does not need an internal core, setting it apart from previous devices. The P(VDF–TrFE–CFE) ferroelectric polymer used in this application provides large electrostrictive strain [17] (3.5% at 100 MV m⁻¹) and high modulus (∼10⁹ N m⁻²) for an easily read display.

Zoom In Zoom Out Reset image size

The dimensions of the roll actuator are based on a standard Braille cell [18], shown in figure 1(b). The center-to-center distance between two adjacent dots is 2.5 mm, and the tip displacement must exceed 0.5 mm while withstanding a minimum force of 0.5 N. The blocking force and free stroke of the actuator at the maximum operating voltage are designed to satisfy these specifications.

The actuator displacement

$$\begin{eqnarray} &&\Delta =l \varepsilon , \end{eqnarray} \tag{ 1 }$$

where l is the active length of the actuator and ε is the strain at the applied voltage. The blocking force is

$$\begin{eqnarray} &&P=2 E t W \varepsilon , \end{eqnarray} \tag{ 2 }$$

where t and W are the thickness and rolling width of the single layer film, respectively, and E is the Young's modulus of the material in the axial direction (0.8 GPa).

The actuation voltage is chosen to provide an electric field of 100 MV m⁻¹ and a corresponding strain of 3.5% without causing film breakdown. The average thickness of the PVDF films is 6 μm so the actuation voltage is 600 V. With 3.5% strain an active length of 15 mm provides the required 0.5 mm pin displacement. To account for manufacturing and material variability, a length of 30 mm is used. To enable a simple, core-free design, the actuator must be structurally stable under the applied blocking force. Due to the large length/diameter ratio, Euler buckling is a likely failure mode so lateral support is provided in the Braille cell design.

The polymer films are created using a solution blending and casting technique. PVDF powder mixed with CFE and TrFE is dissolved in a heated solvent solution. The solution is poured onto a glass substrate, left at rest for 12 h, and then placed in a conventional oven and then a vacuum oven above room temperature (∼80 °C) to eliminate any remaining solvent. Finally, the film is removed from the glass substrate using deionized water. This method is very cost effective and simple but it creates film with randomly oriented polymer chains. To maximize actuation strain, the polymer chains are aligned using a stretching process.

After the casting process, the P(VDF–TrFE–CFE) film thickness is typically between 50 and 70 μm. To reduce the actuation voltage the samples are stretched down to a thickness of 6 μm. Stretching also aligns the polymer chains in the actuation direction, improving strain performance. Using either a linear or roll-to-roll stretching machine, the film, pulled in tension at both ends, is locally heated with a hot wire, allowing the film to stretch according to a user defined ratio.

The abrupt change in temperature near the hot wire creates residual stresses in the stretched film. Annealing relaxes the film by heating close to the melting temperature followed by slow cooling. Annealing also increases the crystalline phase content, which contributes to the electrostrictive strain and improves the elastic modulus.

For the electrode deposition step, the film is tensioned onto 7.5 cm × 10 cm metal frames using a repositionable spray adhesive as shown in figure 2. A mask is temporarily placed on top of the frame, delimiting an electrode pattern of 30 mm × 75 mm. A conductive polymer solution (CPS—Clevos™FE-T) was used as the electrode (blue area on the film in figure 2). CPS is deposited onto the film using an ultrasonic nozzle (Sonaer 130K50T). The electrode performance (conductivity and capacitance) increases with increasing electrode thickness but the strain performance degrades due to the additional passive constraint on displacement. The CPS spray time is adjusted to balance these opposing objectives. A thin gold layer is also deposited on the edge of the electrode pattern, to improve electrical contact at the ends of the actuator. For some of the actuators, aluminum electrodes are evaporated onto the polymer film instead of CPS.

Zoom In Zoom Out Reset image size

Figure 3 schematically shows laminated bilayers with electrodes on three and all four sides. Four sided samples work well if the lamination is poor with gaps or air pockets. For high quality laminations, however, three sided samples exhibit higher strain because there is less passive electrode material constraining motion. Three sided samples are also easier to fabricate, requiring one less electrode deposition step. Simultaneous external pressure and heat are applied to both films for successfully lamination. The layers are compressed using a vacuum bag (∼10 cm Hg) and then placed in an oven. The temperature (100 °C) and heat exposure time (2 h) are optimized to maximize the interlayer bounding strength and equalize the measured capacitance between the three and four layer designs. The optimized lamination process gave bilayers with similar average capacitance (251 nF for 3 sided samples and 268 nF for 4 sided samples), indicating a high quality lamination. In theory, one could use only two electrodes in the bilayer if the final step in the vacuum oven provides a sufficiently high quality lamination of the rolled bilayer.

Zoom In Zoom Out Reset image size

Rolling transforms the flat bilayer to a wound roll actuator. The bilayer is removed from the metal frame. Figure 4 shows a mandrel acting as a temporary core to wind the film around. During winding, the bilayer is pulled along a flat surface by the rotating mandrel with a weight on the film providing tension to minimize the formation of wrinkles. The mandrel temporarily adheres to the film during rolling using an adhesive. At temperatures above 80 °C the adhesive liquefies, allowing the core to be easily removed from the actuator after the lamination in a vacuum oven. To prevent unraveling, another adhesive is lightly deposited on the end of the rolled sample, permanently bonding the last two layers.

Zoom In Zoom Out Reset image size

The last step of the actuator fabrication process is to heat the device at 90 °C in a vacuum oven for 90 min. The vacuum helps eliminate air pockets trapped in between layers during the rolling process and the heat bonds all the rolled layers together, improving the mechanical strength of the structure. Removing residual air in the actuator is essential to improve the voltage breakdown of the device. Lastly, steel balls are bonded on both electrode ends using conductive epoxy to serve as electromechanical contacts (see figure 5).

Zoom In Zoom Out Reset image size

Figure 6 shows the experimental set up used to analyze the performance of 19 actuators fabricated using the process described above. A Trek 609E-6 high voltage amplifier, controlled by LabView, provides the actuation voltage. The displacement and blocking force are measured by a PolyTec OFV 534 laser vibrometer and a Transducer Techniques EBB-1 cantilever beam load cell, respectively. The current going through the actuator is also measured. An acrylic block is machined with 1.9–2.4 mm dia. holes to provide lateral support for the actuator. Galden fluid is pumped into the block to minimize arcing and improve heat dissipation. The steel balls of the actuator are contacted using a magnet on the bottom (positive end) that is fixed to the acrylic block and a magnet on the top that is soldered to a very thin (0.07 mm diameter) copper wire connected to ground.

Zoom In Zoom Out Reset image size

Table 1 shows the average and standard deviation of the parameters and performance of the 19 actuators fabricated and tested. The devices average 2 mm in diameter and 43.6 mm in height formed from bilayers averaging 11 μm in thickness.

The capacitance and loss of the actuators are measured using a precision LCR meter (HP 4284A). The theoretical capacitance

$$\begin{eqnarray} &&C=\frac{{\varepsilon }_{0}{\varepsilon }_{\mathrm{r}}A}{t} \end{eqnarray} \tag{ 3 }$$

where ε₀ is the electric constant, ε_r is the relative permittivity of P(VDF–CFE–TrFE), A is the active area of the actuator, and t the thickness of the single layer film. For the average parameters in table 1 and permittivity ε_r = 40, the expected value of capacitance is 290 nF. The actual measured capacitance is less than theoretically predicted, probably due to incomplete electroding of the bilayer surfaces. The loss is also relatively high due to resistance in the CPS electrodes.

The voltage applied during initial strain testing ranged from 200 V to approximately 700 V in 50 V increments. Two of the nineteen actuators fabricated failed immediately at 200 V. Of the remaining 17 samples, 12 survived the entire experiment and the other 5 failed at various input voltages due to film defects.

Figure 7 shows the experimental results for two example devices. The voltage signal is ramped at 100 V s⁻¹ until reaching the desired value (figure 7(a)). The signal is held constant for 60 s before being ramped back down to zero. Using a ramp input rather than a step input limits the maximum current during the charging process to about 0.1 mA (figure 7(b)). A step input of 650 V generates a current peak of approximately 2 mA, more than an order of magnitude higher. High currents are likely to cause heating at localized imperfections that ultimately lead to breakdown. Figure 7(c) shows the displacement response of the actuator peaking at more than 1.2 mm for an applied voltage of 750 V. Figure 7(d) shows the displacement of an Al-electroded actuator, in which a strain of 3% is obtained under 100 MV m⁻¹. After a quick initial transient, the actuator continues to slowly extend. This extension eventually reaches a constant steady state value after approximately 600 s. When the voltage is turned off, the actuator quickly contracts and then slowly returns to zero displacement. This 'reversible creep effect' [19] has also been observed in piezoelectric materials.

Zoom In Zoom Out Reset image size

Figure 8 shows the average and standard deviation of strain versus electric field for the 17 actuators tested. At 100 MV m⁻¹, the average strain is 2%, much lower than the performance expected from previously tested single layer P(VDF–CFE–TrFE) films [14]. The standard deviation is relatively large at medium and high strains. This variability can be explained by several factors. First, the actuators tend to move axially and transversely. The acrylic block constrains this transverse movement but does not entirely eliminate the associated reduction in the axial displacement. Second, the capacitance data also shows significant uncertainty, indicating variances in material properties and nonuniformity of the electrode deposition. On average, actuators with higher capacitance also produce higher displacement. Third, the adhesive deposited during the rolling process, although minimal, could restrain the axial displacement.

Zoom In Zoom Out Reset image size

Eight actuators are successfully tested for blocking force by positioning the actuator tip in contact with the load cell and applying the voltage time trajectory shown in figure 9(a). The voltage is ramped up, held constant for 30 s and ramped down. The blocking force results for one actuator are shown in figure 9(b). For most of the actuators, the response looks similar to this example for voltages less than 300 V. The blocking force rises until the applied voltage ramps down. The blocking force at that voltage is chosen to be the value just before the down ramp starts. For the example actuator at 400 V, however, the blocking force does not grow monotonically, peaking before the downward ramp. In this case, the blocking force produced by the device is higher than the critical Euler load of the structure and the tube bends laterally. The side walls of the acrylic block limit buckling to a certain extent but at 700 V the actuator fails and no longer withstands high loads.

Zoom In Zoom Out Reset image size

Figure 10 shows the average force and standard deviation for all tested samples with summary results in table 1. On average, the actuators are capable of producing 1 N of force at 100 MV m⁻¹. Buckling typically first occurs at electric fields near 85 MV m⁻¹ and failure is most likely around 125 MV m⁻¹. Weak bounding between the layers and geometric imperfections in the fabricated actuators may also contribute to buckling. The lower forces overall may be due to unmodeled softness in the load path or lower material stiffness caused by high temperature processing during lamination.

Zoom In Zoom Out Reset image size

Frequency response and cycle life of the devices are still in the testing phase, as more samples are needed to get sufficient data. The actuators have a much lower reliability when used in AC as compared to DC. At frequencies above 10 Hz, they tend to break down and burn within a few cycles, even at low voltages. It is believed that heat accumulation due to higher currents is causing this failure. One actuator survived 1000 cycles at 0.2 Hz and 700 V, however.