South Korea is building robots that delete themselves. Samsung is funding the research.

As robots move into pipes, disaster zones, industrial systems, and even the human body, engineers face the problem of retrieving them when the job is done.

South Korean researchers think the answer may be not retrieving them at all. A team at Seoul National University has developed a material that can move like a soft robot and then destroy itself on command. The work, backed by Samsung Electronics and South Korea’s National Research Foundation, points toward a future generation of autonomous machines designed to leave nothing behind.

Published in Advanced Functional Materials on May 11, 2026, the research combines movement and self-destruction in a single material. Instead of relying on separate triggers, batteries, or mechanical systems, the material changes its behavior solely in response to the type of magnetic field applied to it. Under one field, it bends and moves. Under another, it rapidly heats up and breaks itself apart.

How it works

The material is a stretchy silicone rubber embedded with microscopic iron oxide nanoparticles that give it two distinct abilities. Under a steady magnetic field, the particles cause the material to bend and flex, making it behave like a soft robot.

Switch to a rapidly oscillating high-frequency alternating field, and the same particles absorb the energy and release it as heat through a process called ferromagnetic resonance, the same principle used in induction heating, where an alternating electromagnetic field excites conductive particles to generate intense, localized heat.

The material’s temperature rises by over 200 degrees Celsius in under a second. That heat shreds the rubber’s internal chemistry, and the robot falls apart.

The key insight, says lead researcher Professor Seung-Kyun Kang of Seoul National University, is that switching between locomotion and self-destruction requires only a change in magnetic frequency, without physical contact, a secondary trigger, or additional mechanisms. A single external field controls both modes, depending on its configuration.

The retrieval problem

Earlier attempts at dissolvable robotics ran into a wall. Promising candidates such as cyclic poly(phthalaldehyde) were brittle, moisture-sensitive, and prone to premature failure. Engineers also encountered what they call the retrieval problem: in medical procedures, reconnaissance missions, or remote environmental monitoring, going back to retrieve a device afterward is rarely an option.

The security case is equally pressing, and it is where the material’s most consequential applications are likely to emerge. Military and industrial operators increasingly want hardware that can permanently disable itself before it falls into the wrong hands, a requirement that has become more acute as autonomous systems have proliferated into contested environments where recovery cannot be guaranteed.

A soft robot that navigates a space, completes its mission, and then destroys itself on command addresses that requirement without the mechanical complexity of previous approaches. What sets this material apart is that it moves, adapts, and destroys itself using the same single-component structure throughout. Here, the end was always part of the design.

Experimental findings

The composite, made from a silicone elastomer blended with magnetic nanoparticles and a fluoride-releasing compound, heats reliably under a magnetic field and completely breaks down when the field is switched to alternating mode.

Under a steady field, the composite stretched to more than 4.5 times its original length and survived 1,000 bending cycles without losing strength. Switched to a gigahertz-range alternating field, it reached roughly 200°C within a second. Spectroscopy showed the heat-free fluoride ions that severed the silicone’s molecular backbone. More than 90 percent of the material broke down within five minutes, and almost nothing solid remained after 20 minutes.

In demonstrations, a small magnetic robot rolled through a model pipe, reached a simulated mineral blockage, and triggered its own breakdown, a process that also helped clear the obstruction. A flower-shaped version folded its petals as the magnetic field shifted. A temporary circuit switch lit an LED, then was magnetically wiped, permanently disabling the connection.

The South Korean industrial context

The Seoul National University result does not exist in isolation. It arrives inside a broader South Korean and Chinese push toward soft robots engineered to vanish after use. In South Korea’s case, that push is increasingly structured as a strategic industrial program rather than a cluster of independent academic efforts.

Seoul National University’s MDM Lab has built biodegradable robotic fingers with transient electronics embedded directly in their structure. Newer designs pair magnesium needles with pressure and temperature sensors for environmental monitoring missions in which the device is meant to blend into the ecosystem.

Other elastomers developed in South Korea now stretch to sixteen times their original length while remaining biocompatible enough for implantable, bioresorbable devices, a material’s capability with direct applications in both military field medicine and long-duration autonomous sensing.

Samsung Electronics’ funding of elastomer research in this area is the signal that this is no longer purely academic. Samsung is South Korea’s dominant electronics manufacturer and one of its largest participants in the defense supply chain. Its interest in materials that self-destruct before leaking data aligns directly with the security requirements of the next generation of autonomous sensing hardware, devices deployed into environments where capture is possible and where the cost of data exposure exceeds the cost of the hardware itself.

South Korea’s National Research Foundation funding, alongside Samsung’s support, suggests a coordinated national interest in owning this capability, not just demonstrating it. That interest sits alongside South Korea’s broader ambitions for autonomous systems.

The same country that certified its own fighter jet design this month, operates the world’s most advanced robotic border surveillance systems along the DMZ, and is actively investing in autonomous ground and maritime platforms, is now funding materials science that would let those platforms leave nothing behind.

The catch

Building a soft robot that can vanish on command is far harder than making one that moves.

The materials are the first problem. Adding iron oxide nanoparticles to silicone gives it magnetic responsiveness but weakens its structural properties. Too much filler, above roughly 10% by weight, can prevent the silicone from curing at all, limiting the magnetic sensitivity that can be built into any given design.

Degradation itself creates new issues. Breaking the silicone releases fluoride ions, raising biocompatibility and environmental concerns; future designs will need methods to capture or neutralize these ions. Triggering degradation also takes a lot of heat; magnetic methods work only above roughly 250 °C, so these robots are limited to environments that can handle intense, localized heating.

Field deployment adds practical hurdles. Lab equipment for remotely triggering breakdown is often bulky, so tests don’t match real-world conditions. Efficient power delivery and long-range remote control are still unsolved. Stability is another trade-off: early degradable materials were too fragile, and while newer silicone composites are sturdier, engineers must still ensure the material remains stable during use and breaks down quickly when commanded.

The deeper challenge, researchers say, is integrating movement, packing, controlled destruction, and adaptive control into a single compact device without external triggers. Progress, they believe, will come from distributing particles more evenly throughout the material and building wireless electromagnetic systems that enable these robots to operate independently in the field.