The mechanical latching memory of an adhesive tape

TL;DR

The study investigates how a simple adhesive tape forms and retrieves multiple mechanical memories under rectified, unidirectional driving during partial peeling. By encoding turning points as distances $d$ and reading out with the peeling force $F_p$, the authors reveal nested memory states and sequence-dependent erasure, supported by a minimal latching model where tape segments act as thresholded bits. They demonstrate that memory strength can be tuned via aging at the turning point or by changing substrates, yielding persistent or partially erased memories, and show a purely mechanical one-back-like computation. The work provides a generic framework for memories under rectified driving in non-equilibrium soft matter, with implications for mechanical computation and design of tunable, eraseable memory in everyday materials.

Abstract

The storage and retrieval of mechanical imprints from past perturbations is a central theme in soft matter physics. Here we study this effect in the partial peeling of an ordinary adhesive tape, which leaves a line of strong adhesion at the stopping point. We show how this behavior can be used to mechanically store and retrieve the amplitudes of successive peeling cycles. This multiple-memory behavior resembles the well-known return-point memory found in many systems with hysteresis, but crucially the driving here is rectified: peeling is unidirectional, where each cycle begins and ends with the tape flat on the substrate. This condition means that the tape demonstrates a distinct principle for multiple memories. By considering another mechanism that was recently proposed, we establish ``latching'' as a generic principle for memories formed under rectified driving, with multiple physical realizations. We show separately that tape can be tuned to erase memories partially and also demonstrate the function of tape as a mechanical computing device that extracts features from input sequences and compares successive values.

Figures (4)

• Figure 1: Memories of turning points. (a) Strandlines (indicated by arrows) on the beach at Julia Pfeiffer Burns State Park, California, USA. (b) An incompletely peeled adhesive tape, of width 1.3 cm, retains a line of strong adhesion at the turning point where peeling stopped (indicated by the arrow), resembling a strandline in (a). "Before" and "after" pictures show the tape laid flat. Dark speckles indicate sparse contact with the substrate, corresponding to weaker adhesion. (c) Schematic of the protocol used to encode multiple turning points by peeling to successively smaller distances. After encoding we apply readout (final, blue segment), peeling past all stored values. Side panel shows the tape with visible imprints of the three encoded points. (d) Peeling force $F_p$ during readout with encoded distances at $d= 20, 22.5,$ and $25$ mm. Distinct peaks in $F_p$ correspond to the encoded $d$ values. Dotted line is from the same tape with no memories.
• Figure 2: Conditions for multiple memory retention. (a) Protocols for encoding and readout of two memories at $d = 20~mm$ and $d = 25~mm$ in decreasing (Sequence A) and increasing (Sequence B) order. (b) Peeling force $F_p$ during readouts for both sequences. Both memories are present only if written in decreasing order. (c) Peeling force during the entirety of each protocol. Pink box highlights the second write cycle. In Sequence B (bottom), the larger amplitude reads and erases the smaller value.
• Figure 3: Tuning memory strength and partial erasure. (a)--(c) Peeling force plotted during readouts for an encoded memory of $d=25~mm$. $N_R$ indicates number of repeated readout sweeps. (a) Peeling force during two consecutive readouts. The second readout shows almost no signature at $d=25~mm$, implying complete erasure by the first readout. (b) Readout after 100 s aging (schematic of protocol in inset). The memory is much stronger in the first readout, and there is a vestige in every subsequent readout. (c) Readout of tape on a smooth acrylic substrate. Even without significant aging time, the memory persists. (d) Relative memory strength $\Delta F = F_p(\mathrm{peak}) - F_p(\mathrm{baseline})$ in successive readouts. Black squares are from panel (a), and green triangles from panel (c) (see Materials and Methods). $F_p(\mathrm{baseline})$ is the value of $F_p$ estimated at $d$ in the absence of a peak, using a linear fit of the underlying peeling force. Error bars are estimated from three different samples of each tape.
• Figure 4: Minimal model for the tape's memory behavior. (a) The tape represented as an array of bits, with boundaries at the $x$ values below and above each bit. (b) Behavior of a bit under external driving. Each bit starts out in the ($-$) state. When the peeling distance $d$ crosses the lower threshold $x_{i-1}$, the bit flips to a ($+$) state, corresponding to strong adhesion. It remains "latched" in this state until $d$ crosses the upper threshold $x_i$ and resets the bit to ($-$).