Danish Team Creates 'Ionic' Tactile Sensor: Generates Power Without Electronics, Produces Voltage by Pressing

“

Tactile sensors have been around for years, but this time someone decided to change the underlying logic: not relying on electronics, but on ions.

Pressure resistance, piezoelectric, capacitive, and triboelectric—these are the four main technical routes for most tactile sensors today. They each have their strengths, but they all share one common characteristic:

the signal is essentially the movement of electrons.

While the human body works differently. The mechanoreceptors in the skin convert deformation into electrical signals through the movement of ions: when the mechanosensitive ion channels open, ions flow in, the membrane potential rises, and once it crosses the threshold, an action potential is generated, traveling all the way to the central nervous system.

On one side is electronics, on the other is ions. This "physical foundation" difference may be one of the reasons why artificial tactile sensing has yet to truly replicate human touch. For

rehabilitation robots, active prosthetics, and next-generation human-machine interfaces

realistic and credible tactile feedback is precisely the most critical part—

dexterous operation, safe human-machine interaction, and the "embodiment" of prosthetics

all depend on it.

Now, a research team from Aarhus University in Denmark has proposed a new concept—

Artificial Ionic Mechanoreceptor (Artificial Ionic Mechanoreceptor, AIM)

—the study recently appeared in the journal "Advanced Functional Materials."

Its core is simple:

a microchannel filled with NaCl solution, which generates voltage by itself when pressed.

01.

The human body uses ions to transmit signals, so the sensor does the same.

Let's look at the biological prototype first.

The research team referred to the

Pacinian corpuscle (Pacinian corpuscle)

a fast-adapting receptor located in the deeper subcutaneous layer of the skin, responsible for sensing high-frequency vibrations. It looks like an onion: a layered structure wraps around the central nerve ending, with the layers filled with extracellular fluid rich in NaCl, while the inside of the nerve ending membrane contains potassium ions and negatively charged proteins.

Because of this ion distribution,

the resting membrane potential difference inside and outside the membrane is −70 mV

. The stretch-gated sodium channels on the membrane are closed at this time. Once mechanical stimulation is applied, the deformation of the layers transmits pressure to the plasma membrane, opening the sodium channels, allowing sodium ions to flow in along the electrochemical gradient, causing a sudden rise in the electrical potential around the nerve ending—

once it exceeds the −50 mV threshold, an action potential is generated, and the signal is sent to the brain

，随后钠钾泵把细胞拉回静息态。

Structure is complex, but the principle can be abstracted into one sentence:

Using mechanical perturbation to change ion distribution, thereby changing the potential.

AIM is designed based on this sentence: two electrolyte reservoirs embedded in a flexible elastomer, connected by a microchannel, the whole is filled with NaCl solution, each reservoir is sealed by a thin PDMS membrane at the top—the thin film is responsible for "receiving touch."

The key lies in the interface. The PDMS surface itself is charged, thus forming a

double electric layer (EDL)

, ion distribution becomes uneven, divided into two parts: the

Stern layer

, tightly bound to the wall and basically immobile, and the

diffusion layer

。

, where electrostatic interaction is weaker and ions can still move freely. Pressing one side of the membrane causes the liquid inside the cavity to be pressurized, creating a pressure difference between the two cavities, and the electrolyte flows along the microchannel. The flowing liquid will "shear" the double electric layer, dragging the mobile counterions in the diffusion layer along, forming a

flowing current (streaming current)

; ions accumulate in the opposite cavity, and the gold electrode at the bottom detects the potential change—this is the

flowing potential (streaming potential)

, following the classical Helmholtz–Smoluchowski relationship, its magnitude depends on surface charge density, channel geometry, ion concentration, and flow velocity.

After releasing the pressure, the membrane on the other side is pushed back by its own elasticity, causing reverse flow in the channel, and the signal also reverses.

The entire process requires no external voltage or excitation source, it is a fully self-powered sensing mechanism.

02.

442.5 Hz, 5 mM, 0.05 mV/g

The device is fabricated using soft lithography: the bottom layer is a PDMS film about 100 µm thick with patterned gold electrodes (electron beam evaporation about 45 nm, measured about 429 Å), the top layer is made using a two-step SU-8 master mold casting, creating the reservoirs, microchannel, and inlet, then bonded through HMDS treatment and oxygen plasma activation.

The final entire device is only about 400 µm thick.

Characterization results show that the SU-8 master mold has a size error of about 0.5%, surface roughness of about 15 nm, and the PDMS replica has a size deviation less than 1%. For the electrodes, the team tested three serpentine patterns: S, M, and L.

Under 4% tensile strain, the resistance of all three patterns gradually increases but no cracks or breaks occur, and in bending tests, the resistance change is all below 1.5%

， the final choice was the S-type with lower initial resistance.

The probe is pressed down, causing forward flow in the channel, and the signal rises to a peak; the probe remains stationary for about 50 ms, with almost no flow in the channel, but the signal slowly decreases—because ions accumulated near the electrode diffuse back to equilibrium; the probe is lifted, and the residual ion distribution is strongly disturbed by reverse flow, causing the voltage to drop rapidly and

a negative "undershoot" (undershoot) appears

; finally, the liquid is stationary, and ions redistribute uniformly, bringing the signal slowly back to baseline.

First overshoot, then undershoot, finally returning to zero

—this biphasic waveform is highly similar to the transient response logic of fast-adapting biological mechanoreceptors: using a sharp overshoot to mark "contact start", and using the undershoot and recovery segment to mark "contact end", thereby clearly separating continuous stimulation events in time.

Performing FFT on the typical waveform,

the main peak appears at approximately 442.5 Hz, which exactly falls within the working frequency range of vibration-sensitive Pacinian corpuscle.

Stability aspects:

Under 2 Hz, continuous 2000 loadings, the output remains stable, with relative standard deviation between cycles about 3.4%

; no structural failure of the thin film was observed; the output remains overall stable within the temperature range of 20–36 ℃, with signal change about 6% when heated to 28 ℃, followed by a slight decrease; when bent on different diameter cylindrical supports, the features such as the rising segment, peak, overshoot, and recovery are basically retained. In addition, under cyclic stimulation, the device has a working frequency window:

below about 0.5 Hz, the signal is too weak to be reliably detected, and above about 7 Hz, it becomes unstable because there is not enough time to complete ion redistribution

—this is quite similar to the behavior of biological receptors needing an "absolute refractory period" to neutralize charge.

When the channel is widened (60→140 µm), the peak voltage increases, but the rate of increase slows down near 140 µm; when the channel is lengthened (1.5→3.5 mm), the peak also increases, and it tends to level off near 3.0 mm.

More ions mean more charge carriers, and the flow current is expected to increase; however, higher concentration will significantly reduce zeta potential and thin the diffusion layer—based on the estimation for monovalent electrolyte,

the Debye length for 0.1 mM, 0.5 mM, 5 mM, 0.5 M, and 5 M corresponds to approximately 30.4, 13.6, 4.3, 0.43, and 0.14 nm

respectively. The result of the two opposing effects is:

there exists an optimal value, approximately 5 mM NaCl.

In terms of stimulation parameters, the film insertion depth from 30 µm to 190 µm increases the peak voltage and makes the response faster; the probe speed in the range of 2–10 mm/s is approximately linear,

the speed sensitivity is about 0.06 mV/mm, and the force sensitivity is about 0.05 mV/g;

at higher speeds, the rapid flow can wash away ions near the electrode, slowing down the amplitude increase. The team also defined a dimensionless metric

relative rising time

(signal peak time / stimulation duration) to describe the delay: the deeper the insertion, the greater the delay, while the faster the speed, the smaller the delay.

03.

Mounted on a prosthetic finger, it also felt the pulse

In the concept validation part, the researchers attached AIM to a soft prosthetic hand.

Mounted on the fingertip, it can generate clear and readable signals when in contact with a human finger;

switched to the thumb, pressed on the radial artery at the wrist, it successfully captured the periodic pulse waveform.

This is the most demanding test scenario in the paper—arterial pulsation provides the weakest mechanical stimulation, and the output signal is also the smallest.

The signal-to-noise ratio of the original signal is about 2.2, and after conventional signal processing, it is improved to about 8

and the pulse waveform is preserved, which is sufficient to reliably identify physiological signals.

However,

although the waveform shape is already very similar to biological receptors, the voltage amplitude is still insufficient

—the current output has not yet reached about 20 mV, the threshold for neuronal excitation. To directly communicate with the nervous system, either external amplification is needed, or the device itself should be improved. The team's direction is to

modify the surface chemistry of the microchannel walls to increase surface charge density, thereby enhancing zeta potential and flow potential.

Looking ahead, this "pressure-driven ion redistribution" mechanism is completely different from the mainstream solid-state transducers. It is self-powered, naturally compatible with soft materials, and sensitive to weak stimuli. Researchers believe that in addition to prosthetic fingertips, it may also be extended to distributed artificial skin, implantable pressure sensors, and

neuro-robotic hybrid interfaces

scenarios—where chemical and ionic compatibility with biological tissue is precisely the key. The long-term goal set by the research team for AIM is also clear:

to install it in future robotic prosthetics, enabling prosthetics to truly possess tactile sensation and real-time neural feedback capabilities.

Paper address:

https://doi.org/10.1002/adfm.77916

END

Industrial robot companies

Eston Automation

Efe Robot

Lawrence Robotics

Yuejiang Robotics

Jieka Robotics

Songling Robotics

Luosh Stone Robotics

Artom Robotics

Jizhi Jia

Hikvision Robotics

Yifei Technology

Ailite Robotics

Service and Specialized Robotics Enterprises

Yijiahe

Jingpin Tiesheng

Qiteng Robotics

Shihe Robotics

Pudu Robotics

Schroder Robotics

Kuma Technology MAMMOTION

Humanoid Robotics Enterprises

UBTech Robotics

Universal Robots

Star Depth

Xingdong Jiayuan

Weijing Robotics

Zhujidongli

Leju Robotics

Daxiong Robotics

Magic Atom

Zhongqing Robotics

Paxini Perception

Cyborg Robotics

Digital China

Fuliye Intelligent

Tianlian Robotics

Kepler Humanoid Robotics

Lingbao CASBOT

Qingbao Robotics

Zhejiang Humanoid Robotics Innovation Center

Dongyi Technology

Zhishen Technology

PNDbotics

Zhuoyide Robotics

Qinglang Intelligent

Galileo GALILEO

Songyan Power

Tianji Intelligent

Woa'an Robotics

Linghu Huawei

Jiace Evolution

Embodied Intelligence Enterprises

Cross Dimension Intelligent

Galaxy General

Qianxun Intelligent

Lingxin Qiaoshou

Ruierman Intelligent

Weiyi Intelligent Manufacturing

Tuisheng Technology

Zhongke Silicon Chronicle

Shu途 Technology

Dexterous Intelligent

Xingchen Intelligent

Qiongche Intelligent

Fangzhou Infinite

| iFlytek |

Beijing Humanoid Robotics Innovation Center

National and Local Co-construction Humanoid Robotics Innovation Center

Daimeng Robotics

Xibit Robotics

Star Sea Chart

Yuequan Bionics

Zero Equation Robotics

Zhongke ShenGu

Zhifang Square

Daka Robotics

Haocun Technology

Identify Intelligent

Xynova Xinnov Future

Feixi Technology

Future Power

Boden Intelligent

Qianjue Technology

Lingsheng Technology

Ji萃 Intelligent Manufacturing

Xinbeit Technology

Chenhuanshan Technology

Dexmal Original Force

Youliqi

Self-variable

Ruiyan Zhikong Dexterous Hand

启物科技

RoboScience机器科学

中科第五纪

临界点

当虹科技

桥介数物

Vbot维他动力

他山科技

具脑磐石

优艾智合机器人

智行腱

阿米奥机器人

鹿明机器人

量子智能

若愚科技

钛虎机器人

智无际

医疗机器人企业

元化智能

天智航

思哲睿智能医疗

精锋医疗

佗道医疗

真易达

术锐®机器人

罗森博特

水木东方

｜

康诺思腾

迪视医疗

上游产业链企业

绿的谐波

因时机器人

坤维科技

脉塔智能

青瞳视觉

本末科技

蓝点触控

| 鑫精诚传感器

BrainCo强脑科技

宇立仪器

极亚精机

思岚科技

神源生

非普导航科技

因克斯

巨蟹智能驱动

凌云光元客视界

璇玑动力

意优科技

| 瑞源精密 |

灵足时代

HIT华威科

星汇传感

凌迪科技

泉智博

CubeMars机器人动力

旺龙机器人乘梯

航凯微电

悟通感控