Science - USA (2022-04-29)

andsolvent.Thisisobservedintheioniccon-
ductivity, for which that of the polymer gel is
higher than that of the bulk electrolyte at high
salt concentrations (>1.0 M), peaking near the
sensor crossover point (Fig. 2F), and the dif-

fusion coefficients in bulk solution correspond-

ingly drop below those in the matrix, crossing

over at 1 M (fig. S14). The polymer matrix

reduces ion-ion and ion-solvent interactions

( 26 , 27 ). The lower viscosity reduces drag,

whereas effective ion concentration is increased

by reducing ion pairing, hence increasing

overall conductivity. This observation means

that we cannot readily measure the hydro-

dynamic component of drag on ions within

504 29 APRIL 2022¥VOL 376 ISSUE 6592 science.orgSCIENCE

0 5 10 15 20 25

0

-500

-1000

0 5 10 15 20 25

0

-100

-200

tn

err

u

C

(nA)

Anion dominant

Cation dominant

Anion dominant

Cation dominant

B C

H

I J K L

A Pressure (Pa) D

E F G

Anion transport

Cation transport

Induced field (V/m)

o m

a

n

d n

Fig. 2. Piezoionic mechanisms and electrical responses.(A) Schematic of
a polymer gel under indentation, exhibiting differential ionic displacement and
field, with smaller red cations being carried through the green polymer chain
network faster than the blue anions, generating a charge imbalance and electric
field. In the indentation experiments, a working (sense) electrode is placed
under the indented portion, whereas a reference (ground) electrode is placed
in the undeformed portion of the gel. In this case, a preferential displacement
of the cations over anions results in a negative voltage reading. (B) Poly(acrylic
acid-co-acrylamide) hydrogel voltage response upon step compression at
20 kPa. Transient behavior slows with increasing polymer content (15 and
30% w/v). Dashed fit curves were generated using a finite element simulation.
(C) Simulation of the contributing components of ionic current at open circuit
during the 15% poly(AA) gel compression, and (D) simulated breakdown
of the ionic transport during the 30% poly(AA) gel compression, showing that
convection dies off faster in the gel with lower polymer content. (E) Ionic
conductivity of 15% w/v pAAm gel swollen with NaCl of varying concentrations
and peak generated voltages (black; data are mean ± SD,N= 3) under
10% sinusoidal compression at 0.1 Hz, showing increased signal at higher
concentrations (red; data are mean ± SD,N= 3). (F) Ionic conductivity of

the PVDF-HFP solid polymer electrolytes (blue) and bulk LiTFSI/PC solution

(green) as a function of concentration, showing that the polymer electrolyte is

higher in conductivity at high concentrations (data are mean ± SD,N= 3).

(G) Diffusion coefficients of Li+(pink) and TFSI−(blue) ions in the polymer electrolyte,

measured by pulsed-fieldÐgradient NMR, showing that the TFSI−is more diffusive

at low concentrations, and Li+at higher concentrations (data are mean ± SD,N= 3).

(H) Peak voltages generated by 0.1-Hz, 5% sinusoidal compressive strain in PVDF-

HFP solid polymer electrolytes with PC/LiTFSI at varying concentrations, showing a

reversal in sense voltage at a concentration similar to that of the diffusion crossover

point (data are mean ± SD,N= 3). (I) Piezoionic transient response under a step

compression from 15 and 30% pAAm swollen in 1.5 M NaCl. Note the faster decaying

response both at open circuit and short circuit when the polymer content is lower.

(J) Piezoionic transient response under repeated step compressions onto a 200-mm-

thick (top) and a 1-mm-thick (bottom) pAAm gel swollen with 1.5 M NaCl, exhibiting

decay times of ~50 ms and ~5 s, respectively, emulating rapid- and slow-adapting

mechanoreceptors. (K) Peak current and power transfer as a function of resistive load,

using a pAAm gel swollen with 1.5 M NaCl (data are mean ± SD,N= 3). (L) Energy

harvested and electromechanical coupling as a function of external load, using a pAAm

gel swollen with 1.5 M NaCl (data are mean ± SD,N= 3).

RESEARCH | REPORTS