HYDROGELS
Piezoionic mechanoreceptors: Force-induced current
generation in hydrogels
Yuta Dobashi1,2,3, Dickson Yao^1 , Yael Petel^4 , Tan Ngoc Nguyen1,5, Mirza Saquib Sarwar1,5,
Yacine Thabet^1 , Cliff L. W. Ng1,4, Ettore Scabeni Glitz^1 , Giao Tran Minh Nguyen^6 , Cédric Plesse^6 ,
Frédéric Vidal^6 , Carl A. Michal7,4, John D. W. Madden1,2,5*
The human somatosensory network relies on ionic currents to sense, transmit, and process tactile information.
We investigate hydrogels that similarly transduce pressure into ionic currents, forming a piezoionic skin.
As in rapid- and slow-adapting mechanoreceptors, piezoionic currents can vary widely in duration, from
milliseconds to hundreds of seconds. These currents are shown to elicit direct neuromodulation and muscle
excitation, suggesting a path toward bionic sensory interfaces. The signal magnitude and duration depend on
cationic and anionic mobility differences. Patterned hydrogel films with gradients of fixed charge provide
voltage offsets akin to cell potentials. The combined effects enable the creation of self-powered and ultrasoft
piezoionic mechanoreceptors that generate a charge density four to six orders of magnitude higher than
those of triboelectric and piezoelectric devices.
H
uman touch perception relies on soft
cutaneous mechanoreceptors that intri-
cately interface with even softer neural
tissues. In advanced prosthetics, robot-
ics, and wearable devices, the presence
of sensor arrays that mimics human skin has
led to elastomer-based piezoresistive ( 1 ), ca-
pacitive ( 2 ), and electret ( 3 ) sensors and elec-
tronic skins ( 4 – 6 ). These sensors enable soft
interactions and can overcome the device-to-
tissue mechanical mismatch that is inherent
to conventional electronic devices. Ionic skin
presents an alternative to electronic skin,
producing ultrastretchable, tough, and trans-
parent devices, including capacitive sensors,
from hydrogels ( 7 – 9 ). In this study, we used
hydrogels to directly produce ionic currents in
response to pressure gradients—a piezoionic
approach to sensing. This method comple-
ments other recent iontronic approaches such
as the use of supercapacitive gates and highly
conductive hydrogel–conducting polymer hy-
brids, enabling the large ionic charge ex-
change at the low voltages needed for neural
interfaces ( 10 , 11 ). Piezoionics are shown to
offer similarly high–charge density and low-
voltage operation. Like their biological coun-
terparts, they rely on ionic currents. Unlike
mechanoreceptors, in which a transmembrane
Donnan potential is released by mechanical
gating (Fig. 1A), piezoionic sensors operate by
pressure-driven ion flux (Fig. 1B). This occurs
either by streaming a single polarity of ions
or by one polarity of ions being preferentially
carried over the counterions through the poly-
mer matrix, creating a net charge imbalance.
Streaming and other pressure-driven voltage-
generation methods have been documented
in ion conductors such as Nafion ( 12 ), poly-
pyrrole ( 13 ), cartilage ( 14 ), and hydrogels
( 15 – 17 ). Upon deformation of such piezoionic
matrices, the open-circuit voltage gradient,
∇Vsense, is proportional to the applied pres-
sure gradient,∇P(Fig. 2A). In the simplest
case, the relationship can be described as
∇Vsense¼a∇P, whereais the piezoionic co-
efficient (see full discussion in the supplemen-
tary materials).aranges from 0.01 to 100 nV/
Pa, depending on the ionic species and matrix
properties (table S2). It is equivalent to the
ratioLo/sproposed by de Genneset al.( 12 )
to describe field generation in ionomeric
polymers exposed to a pressure gradient,
whereLorepresents the coupling between
the applied pressure and the induced ionic
current, andsis the ionic conductivity.
To explore the molecular origins of this ef-
fect and investigate applications in sensing,
we designed an indentation experiment in
which the open-circuit voltage or short-circuit
current was measured between the deformed
and undeformed portions of the gels (Fig. 2A
and fig. S1). Indentation is analogous to the
effect of a finger or other object pressing into
skin. Two voltage responses to a very light
step compression of 20 kPa are shown in Fig.
2B for acrylamide (AAm)–based charged co-
polymers that contain acrylic acid (AA) and
proton counterions. This ultra-absorbent and
biocompatible hydrogel, which is ubiquitous
in commercial uses, generates several milli-
volts as a result of streaming potentials ( 15 ).
Transient response time depends on polymer
content, with the lower polymer content (15%by weight, black curve) producing a faster rise
(1 s) and decay (15 s) than the higher polymer
content version (30%, blue curve), which ex-
hibits a slow rise (15 s) and little decay. The
halving of the polymer content changes com-
pressibility, leading to much larger relative
stress relaxation (fig. S5A). This suggests that
the faster response in the lower-content hy-
drogel is due to a greater permeability of the
hydrogel matrix, increasing within-pore flow
that carries protons and, in turn, producing a
faster rise. The decay in response is also faster
as the flow completes sooner.
A Poisson-Nernst-Planck model was com-
bined with poroelastic mechanics (see sup-
plementary materials) to help interpret the
transient behavior (purple and red model-fit
curves in Fig. 2B). The model indicates that
the application of a pressure gradient creates
an initial ionic current, driven by the internal
convection of water and protons within the
pores. This movement of charge creates con-
centration and voltage gradients that, in turn,
lead to currents that oppose the convection-
driven currents (Fig. 2, C and D; movies S8 to
S10; and fig. S19). Peak voltage is reached
when the back current, composed of diffu-
sive and electrophoretic fluxes, balances the
convective current. When electrolyte flow
subsides, the transient separation of charge
decays. The decay is much slower in the den-
ser polymer (Fig. 2B), owing to the higher
flow resistance and longer poroelastic time
constant. This time constant is given by
tp¼ðÞ^1 ^2 nhL2
21 ðÞ nGk, wherehis the fluid viscosity
within the pores of the hydrogel,Lis the
indenter size or characteristic length scale
of deformation,Gis the shear modulus,nis
Poisson’sratio,andkis the matrix perme-
ability.Gincreases andkdecreases with
increasing polymer content, with both being
very strong functions of polymer volume frac-
tion. Overall, the time constant of flow gets
longer as polymer content increases. The
simulation fits with the expectation that dif-
ferences in matrix permeability and stiffness
account for variations in rise and decay be-
haviors (model response in Fig. 2B, dashed
lines). This understanding can be exploited to
create a family of decaying or nearly steady
responses of either positive or negative po-
larity, as simulated in fig. S25. The polarity
is reversed by changing the sign of the mo-
bile ion—for example, by replacing AA with
[2-(methacryloyloxy)ethyl]trimethylammonium
chloride (MAETAC) (fig. S4). A rich set of
voltage responses is obtained as displace-
ments and forces are increased (figs. S3, S4,
and S7), with voltage amplitude increasing
and transient response time changing. A fully
comprehensive set of reasons for the force
dependence of the decay is unclear and will
require careful characterization of the changes
in permeability and viscoelastic response as a502 29 APRIL 2022•VOL 376 ISSUE 6592 science.orgSCIENCE
(^1) Advanced Materials and Process Engineering Laboratory,
University of British Columbia, Vancouver, BC, Canada.
(^2) School of Biomedical Engineering, Faculty of Applied
Science, University of British Columbia, Vancouver, BC,
Canada.^3 Institute of Medical Science, University of Toronto,
Toronto, ON, Canada.^4 Department of Chemistry, University
of British Columbia, Vancouver, BC, Canada.^5 Department of
Electrical and Computer Engineering, University of British
Columbia, Vancouver, BC, Canada.^6 CY Cergy Paris
Université, CY Advanced Studies, LPPI, 5 mail Gay Lussac,
Neuville sur Oise, F-95031 Cergy-Pontoise Cedex, France.
(^7) Department of Physics and Astronomy, University of British
Columbia, Vancouver, BC, Canada.
*Corresponding author. Email: [email protected]
RESEARCH | REPORTS