The interactions with touchscreens rely heavily on vision: The virtual buttons and virtual sliders on a touchscreen provide no mechanical sense of the object they seek to represent. This work presents PopTouch: a 500 µm thick flexible haptic display that creates pressable physical buttons on demand. PopTouch can be mounted directly on touchscreens or any other smooth surface, flat, or curved. The buttons of PopTouch are independently controlled hydraulically amplified electrostatic zipping taxels (tactile pixels) that generate 1.5 mm of out of plane displacement. When pressed by the user, the buttons provide intuitive mechanical feedback thanks to a snap‐through characteristic in their force‐displacement profile. The snap‐through threshold can be as high as 4 N, and is tuned by design and actuation parameters. This work presents two versions of PopTouch: a transparent PopTouch for integration on Touchscreens with built‐in touch sensing, such as smartphones and a sensorized PopTouch, with embedded thin‐film piezoelectric sensors on each taxel, for integration on substrates without built‐in touch sensing, such as a steering wheel. PopTouch adds static and vibrating button‐like haptics to any device thanks to its thin profile, flexibility, low power consumption (6 mW per button), rapid refresh rate (2 Hz), and freely configured array format.

Most flexible piezoelectric transducers have a vertical setup with top and bottom electrodes, which does not enable the selective, directional detection of mechanical stimuli. Here we present a paradigm shift in the design of such transducers by placing the electrodes in a single layer and fully embedding them in a ferroelectric layer. This approach enables a selective detection of in-plane strains with a linear, orientation-dependent response. Our transducers feature microstructured, densely interdigitated electrodes embedded in the ferroelectric copolymer P(VDF-TrFE) and show an in-plane strain sensitivity of up to 8.3 nC % ⁻¹ (21.3 V % ⁻¹ ), while being 23 times less sensitive to transversal loading, compared to vertical setup devices. The embedded electrodes cause a strong anisotropy for in-plane strain coupling and make it possible to distinguish both the bending orientation and the bending intensity on a time-variable curvilinear surface. A high power density of 2.3 mW cm ⁻³ was achieved during a periodic bending movement at 90 Hz. In addition to a photolithography and electroplating-assisted method, we present an alignment-free, elegant microcapillary force-based process for scalable fabrication of embedded electrodes. The presented transducers have a high potential for application as energy-autonomous sensors for condition monitoring, robotics and wearables.

Light guiding is used to direct light from an emitting source to a different location. It is frequently realized through a clad–core structure with a difference in the refractive index of the materials. This paper explores the possibility of combining a fluoropolymer (THV) film of low refractive index, serving as a cladding layer, with a polycarbonate (PC) core, via injection molding. Pristine THV lacks adherence to the PC. However, when treated with ${\mathrm{O}}_2$ plasma prior to overmolding, bonding can be established that was quantified in peel tests. The effect of this surface treatment was further investigated by adjusting the plasma treatment duration and time to overmolding. Furthermore, parameter studies comprising the four molding parameters, namely packing pressure, injection speed, melt temperature, and mold temperature, were performed. Numerical injection molding simulations assessed the prevailing temperatures at the PC–THV boundary. Consequently, the temperature–time integral could be calculated and linked with the measured peel strengths by fitting a proportionality constant. While the plasma treatment duration showed minor influence, the activation diminished with time, halving the measured peel loads within 24 h. The adhesion was experimentally found to increase with a lower packing pressure, faster injection speed, and higher melt and mold temperature. Those same molding relations influencing the peel loads were also found with the temperature–time integral when scaled by the proportionality constant in the simulations (${\mathrm{R}}^2=85\%$). Apparently, adhesion is added by molding settings which promote higher interface temperatures that prevail for longer. Hereby, the faster injection speed increases the melt temperature through shear heating. A higher packing pressure, in contrast, presumably increases the heat transfer at the PC–THV interface, accelerating the cooling. The measured peel loads were 0.3–1.6 N/mm for plasma-treated samples and nearly zero for pristine THV.