(Fig. 3B). The pressure sensitivity of a single
mechanoreceptor unit is 8mV/kPa (a=8×
10 −^9 V/Pa), probed with pressures of up to
360 kPa using a 0.1-Hz sinusoidal perturbation
(Fig. 3C). A 4-cm–by–4-cm sensor array was
tested under single- and multitouch condi-
tions(Fig.3,EtoHandmoviesS1andS2),
as well as in detection of a swiping gesture
(movie S3). A gentle finger press (~100 gram-
force) generates voltage changes of approxi-
mately−10 mV, which are superimposed on
the−50-mV Donnan potential. Little cross-
talk between mechanoreceptors was observed.
By changing the film thickness or polymer con-
tent or using current response instead of volt-
age, the properties of this transient response—
including the decay time, which is ~1 s with
the current dimensions—can be made faster
or slower. The desired time constant depends
on the application. If the aim is to detect
change but ignore steady forces, then a fast
time response is desired—for example, creat-
ing a situation in which the skin“forgets”
about a constant force, just as we might stop
being aware of clothing or a persistent noise.
This piezoionic skin may be suitable for soft
robotics, medical devices, or wearable appli-
cations, given its mechanical compliance and
self-powered nature. We demonstrate piezo-
ionic detection of joint flexion and extension
(fig. S27 and movie S6), for which the com-
pliant nature makes it unobtrusive. We also
detect the touch of a finger on skin and on a
toy(movieS7).Tobeeffectiveinsuchdryap-
plications over hours or longer, the sensor
needs a stretchable encapsulation layer. Eco-
flex 00-30 was employed for this purpose (see
supplementary materials), enabling up to sev-
eral days of operation, which is sufficient for
many wearable applications. Figures S29 and
S30 demonstrate 390 cycles, where some drift
is observed, returning to normal after rest-
ing, without degradation in performance.
Switching to styrene-isobutylene-styrene block
copolymers with low rates of water-vapor
transmission should allow for months or even
years of operation without dehydration ( 32 ),
as could the use of lithium salts or ionic liquids
( 33 ). Owing to the bioinert nature of many
synthetic hydrogels, it is also conceivable to
employ piezoionics in implantable devices.
Other advantages are the simplicity of the
device and its adaptable compliance. An arte-
rial phantom (fig. S28) demonstrates the sen-
sor detecting a pulsatile flow that simulates
cerebrovascular hemodynamics ( 34 ).
We performed peripheral nerve stimulation
to demonstrate the possibility of self-powered
piezoionic neuromodulation. A piezoionic sen-
sor element composed of a pAAm hydrogel
swollenin1.5MNaClwaspressedbyhandto
generate tactile-induced current pulses. Con-
nections to the sciatic or tibial nerves of rodent
models were made via conventional stainless-
steel needle electrodes (250-mm diameter)
or poly(3,4-ethylenedioxythiophene) (PEDOT)–
coated (~1mm) Pt-Ir wires (75-mm diameter),
asshowninFig.4A.ThePEDOT-coatedPt-Ir
wires were gently wrapped around the tibial
nerve, whereas the stainless-steel needles were
placed by piercing the perineurium of the
sciatic nerve. In either case, the distance be-
tween the contacts was 2 to 4 mm. Because
the charge capacity of metals is relatively
small [~0.2 F/m^2 ( 35 )], conducting polymer
PEDOT was used to increase charge injec-
tion capacity (Fig. 4B), as discussed in the
supplementary materials. An electromyogram
(EMG) was concurrently recorded by inserting
a pair of needle electrodes in the gastrocne-
mius muscle.
A series of presses was then applied to the
piezoionic sensor connected to the peripheral
nerves. Using the stainless-steel electrodes, we
did not observe a visible twitch of the hind-
limb but did note a clear EMG signal, indicat-
ing that muscle is activated (Fig. 4C) but
insufficient to attain a visual confirmation.
The EMG signal lagged the voltage generated
by the sensor by ~8 ms, as expected given the
conduction velocities. Nonetheless, amplifica-
tion of the sensor signal via analog-to-digital
conversion into biphasic pulse trains enabled
observation of hindlimb twitches (movie S4).
Use of the PEDOT-coated Pt-Ir electrodes with
the nerve, with the resulting currents in excess
of 10mA and well under 100 mV, produced
clearly visible hindlimb twitches without any
signal conditioning or amplification (Fig. 4D
and movie S5).
Piezoionics offer sensing solutions that are
soft, self-powering, and biocompatible. In re-
sponse to pressure, they can be tailored to pro-
duce a wide temporal range of transient signals
and provide much higher charges at much
lower voltages than piezo- and triboelectric
devices. Consequently, piezoionics are a good
match for neural interfacing applications.
Piezoionic devices are another iontronic tool,
part of a system that, like nature, also includes
computation, actuation, and energy storage.REFERENCES AND NOTES- S. C. B. Mannsfeldet al.,Nat. Mater. 9 , 859–864 (2010).
- M. S. Sarwaret al.,Sci. Adv. 3 , e1602200 (2017).
- S. Zhanget al.,Nano Lett. 20 , 4580–4587 (2020).
- A. Chortos, J. Liu, Z. Bao,Nat. Mater. 15 , 937–950 (2016).
- T. Someya,IEEE Spectr. 50 , 50–56 (2013).
- D.-H. Kimet al.,Science 333 , 838–843 (2011).
- C.-C. Kim, H.-H. Lee, K. H. Oh, J.-Y. Sun,Science 353 , 682– 687
(2016). - J.-Y. Sun, C. Keplinger, G. M. Whitesides, Z. Suo,Adv. Mater.
26 , 7608–7614 (2014). - X. Puet al.,Sci. Adv. 3 , e1700015 (2017).
- W. Leeet al.,Adv. Mater. 28 , 9722–9728 (2016).
- Y. Liuet al.,Nat. Biomed. Eng. 3 , 58–68 (2019).
- P. G. de Gennes, K. Okumura, M. Shahinpoor, K. J. Kim,
Europhys. Lett. 50 , 513–518 (2000). - W. Takashima, K. Hayasi, K. Kaneto,Electrochem. Commun. 9 ,
2056 – 2061 (2007). - E.H.Frank,A.J.Grodzinsky,J. Biomech. 20 , 615– 627
(1987).
15. A. Fiumefreddo, M. Utz,Macromolecules 43 , 5814– 5819
(2010).
16. K. Prudnikova, M. Utz,Macromolecules 43 , 511–517 (2010).
17. K. Sawahata, J. P. Gong, Y. Osada,Macromol. Rapid Commun.
16 , 713–716 (1995).
18. Y. Liuet al.,Small 12 , 5074–5080 (2016).
19. S. M. Villaet al.,ACS Macro Lett. 8 , 414–420 (2019).
20. V. Woehlinget al.,Multifunctional Materials 2 , 045002 (2019).
21. M. S. us Sarwaret al., inSPIE Smart Structures and Materials +
Nondestructive Evaluation and Health Monitoring, Y. Bar-Cohen,
Ed. (International Society for Optics and Photonics, 2015),
p. 943026; https://proceedings.spiedigitallibrary.org/proceeding.
aspx?articleid=2239762.
22. Y. Dobashi, G. Allegretto, M. S. Sarwar, E. Cretu,
J. D. W. Madden,MRS Adv. 1 , 63–68 (2016).
23. V. Kapur, J. Charkoudian, J. L. Anderson,J. Membr. Sci. 131 ,
143 – 153 (1997).
24. K. B. Kosto, W. M. Deen,Biophys. J. 88 , 277–286 (2005).
25. H. Jianget al.,Phys.Chem.Chem.Phys. 21 , 10417– 10422
(2019).
26. L. Othman, K. B. Md Isa, Z. Osman, R. Yahya,Mater. Today Proc.
4 , 5122–5129 (2017).
27. S. S. Sekhon,Bull. Mater. Sci. 26 , 321–328 (2003).
28. P. Delmas, J. Hao, L. Rodat-Despoix,Nat. Rev. Neurosci. 12 ,
139 – 153 (2011).
29. M. Mendelson, W. R. Loewenstein,Science 144 , 554– 555
(1964).
30. C. Dagdevirenet al.,Proc. Natl. Acad. Sci. U.S.A. 111 ,
1927 – 1932 (2014).
31. D. B. McCreery, W. F. Agnew, T. G. H. Yuen, L. Bullara,IEEE
Trans. Biomed. Eng. 37 , 996–1001 (1990).
32. S. Ebrahimi Takallooet al.,Robotics 8 , 60 (2019).
33. C. Keplingeret al.,Science 341 , 984–987 (2013).
34. L. Zarrinkoobet al.,J. Cereb. Blood Flow Metab. 35 , 648– 654
(2015).
35. A. J. Bard, L. R. Faulkner,Electrochemical Methods, Fundamentals
and Applications(Wiley, 1980).
36. Y. Dobashiet al.,“Surface sensor arrays using ionically
conducting material,”US patent 10,401,241 (2017).
37. J. D. W. Maddenet al.,“Method and sensor for pressure sensing
based on electrical signal generated by redistribution of mobile ions
in piezo-ionic layer,”US patent 10,663,358 (2018).
ACKNOWLEDGMENTS
This work greatly benefitted from discussions with V. X. D. Yang and
J. Ku on the use of piezoionics in neural interfacing and monitoring.
Peripheral neuromodulation studies were carried out with technical and
methodological guidance from S. McErlane and R. Hildebrandt. We
thank the reviewers for their insightful and constructive feedback.
Funding:This study was supported by a Natural Sciences and
Engineering Research Council of Canada (NSERC) Discovery Grant and
a Collaborative Research and Development Grant. J.D.W.M. was
supported by the Institute of Advanced Studies at CY Cergy Paris
Université through the Visiting Scholar program.Author contributions:
Y.D. and J.D.W.M. conceived the experiments and concepts. Y.D. and
D.Y. designed and fabricated the equipment for indentation experiments.
Y.D., D.Y., T.N.N., M.S.S., Y.T., C.L.W.N., and E.S.G. synthesized the
hydrogels and performed the indentation experiments. Y.D. and
D.Y. processed and analyzed the data from the indentation experiments.
Y.P. and C.A.M. constructed custom equipment and performed and
analyzed the NMR measurements. J.D.W.M. and Y.D. interpreted the
experimental data and conceived and developed the analytical and
numerical models, with input from G.T.M.N., C.P., and F.V. Y.D. and D.Y.
fabricated and characterized the sensor array. Y.D. and D.Y. fabricated
the equipment and performed the rodent neuromodulation experiment.
Y.D. and J.D.W.M. wrote the manuscript. All authors have critiqued and
contributed to the revisions.Competing interests:Two US patents
related to this work have been granted (36, 37).Data availability:All
data are available in the manuscript or the supplementary materials.SUPPLEMENTARY MATERIALS
science.org/doi/10.1126/science.aaw1974
Materials and Methods
Supplementary Text
Figs. S1 to S30
Tables S1 to S6
References ( 38 Ð 58 )
Movies S1 to S1315 March 2020; resubmitted 29 June 2021
Accepted 24 March 2022
10.1126/science.aaw1974SCIENCEscience.org 29 APRIL 2022•VOL 376 ISSUE 6592 507
RESEARCH | REPORTS