Page 193 - Read Online

Page 8 of 12 Wu et al. Soft Sci 2023;3:35 https://dx.doi.org/10.20517/ss.2023.26

During this period, the strain-sensing circuit exhibited good stability, with an acceptable change in
normalized resistance of ~1% during one bending-recovery cycle [Figure 3F]. Finally, we heat the strain-
sensing circuit to several specified temperatures to verify its tolerance to high temperatures. When the
temperature rose by 100 °C, the strain-sensing circuit still showed good conductivity, and the temperature-
induced resistance increase was totally acceptable [Figure 3G]. Notably, with supercooling effects , sensing
[48]
circuits based on liquid metal slurries can still maintain good performance for the strain sensors even at
0 °C. In characterization tests, the sensing circuit exhibits different responses in resistance to three different
parameter changes: resistance increase caused by tensile deformation, resistance decrease caused by
compression, and a gradual resistance increase with low slope caused by heating, which can be
distinguished from resistance change curves. All these characterization experiments verified that the liquid
metal sensing circuit fabricated by the scraping method can meet the fundamental functionality and
durability required for the strain-sensing glove.

Application demonstrations of strain-sensing gloves
Compared with commercial strain sensors, sensing units of the strain-sensing glove are in situ patterned on
the nitrile glove instead of being mounted on gloves using tapes or adhesives, as seen in typical commercial
strain sensors. This design possesses better conformability with finger motion. All components of the strain-
sensing glove, including the substrate, encapsulation, and conductive material, are intrinsically soft and
stretchable, showing better compliance than commercial sensors. To verify the application performance of
the strain-sensing glove fabricated by scraping the liquid metal slurry, we demonstrated it from two aspects:
monitoring and manipulating. First, a volunteer wore this strain-sensing glove on the left hand, as shown in
Figure 4A. The exceptional wearing comfort made the whole process as convenient as wearing a regular
glove without extra help. Subsequently, the volunteer performed various hand gestures with the left hand,
and we measured the resistance changes of five strain sensors on the corresponding fingers in real time
throughout this process. The strain-sensing glove is integrated, meaning the motion of any joint on the
finger will contribute to the tensile deformation of strain-sensing units on the nitrile glove. This enables the
maximum variation range in resistance when the finger bends to make a fist. During this demonstration, six
different gestures were sequentially performed, and the normalized resistance change curves corresponding
to the five strain sensors are shown in Figure 4B. From the variation curves in Figure 4B, the strain-sensing
glove shows good sensitivity, and even small changes in the process of gesture switching can be reflected in
the curves.

Furthermore, benefiting from wearing comfort and excellent responsiveness, the strain-sensing glove shows
huge application potential in human-machine interaction. For individuals who have lost limbs due to
accidents or for some operations in daily life that require the coordination of left and right hands (such as
snatching), it is not easy to synchronize a prosthetic limb with normal limbs. Therefore, for some simple
occasions that require limb coordination, it is also beneficial to manipulate prosthetics with normal limbs to
complete collaborative tasks. To this end, based on the strain-sensing glove, we demonstrated the human-
machine interaction by mirroring the hand gestures of the left hand to operate the right machine hand.
Figure 4C shows the schematic diagram of the human-machine interaction circuit, where the strain-sensing
glove serves as a soft controller to manipulate the machine hand. The strain-sensing unit on each finger is
connected in series with a fixed value resistor (R ) so that the resistance changes of strain-sensing units
fix
caused by finger motion can be converted into changes in voltage, which can be read by a microcontroller
unit (MCU, Arduino MEGA 2560, Supplementary Figure 6). The 10-bit analog-to-digital converter (ADC)
in the MCU can convert the read analog voltage signals into digital signals with an input voltage (Vin) of
1.1 V; the minimum voltage change that the 10-bit ADC can detect is ~1 mV. Besides, we chose suitable R
fix
in series with strain-sensing units to maximize the variation ranges of voltage signals. Next, the MCU
outputs corresponding pulse width modulation (PWM) signals on the pin according to the collected digital