A few years ago, MIT researchers invented a cryptographic ID tag that is several times smaller and much cheaper than traditional radio frequency tags (RFID) that are often attached to products to verify their authenticity.
These tiny tags, which offer improved security over RFID, utilize terahertz waves, which are smaller and travel much faster than radio waves. However, these terahertz tags shared major security vulnerabilities with traditional RFID. That means a counterfeiter could rip a tag off a genuine item and reattach it to a fake item, and the authentication system would be none the wiser.
Researchers have now overcome these security vulnerabilities by leveraging terahertz waves to develop tamper-resistant ID tags that offer the advantages of being small, inexpensive, and secure.
They mix tiny metal particles into the adhesive that attaches the tag to the object, then use terahertz waves to detect the unique patterns those particles form on the surface of the item. Similar to a fingerprint, this random adhesive pattern is used to authenticate items, explains Eunseok Lee, a graduate student in Electrical Engineering and Computer Science (EECS) and lead author of a paper on tamper-evident tags.
“These metal particles are essentially mirrors of terahertz waves. If you spread several pieces of mirror on a surface and then shine light on them, different reflection patterns will appear depending on the direction, size, and position of the mirrors. However, when the chip is removed and reattached, the pattern is destroyed,” adds Ruonan Han, an EECS associate professor who leads the Terahertz Integrated Electronics Group at the Institute of Electronics.
The researchers created an optically tamper-evident tag that is about 4 square millimeters in size. They also demonstrated a machine learning model that helps detect tampering by identifying similar adhesive pattern fingerprints with greater than 99% accuracy.
Because terahertz tags are very inexpensive to produce, they can be implemented across large-scale supply chains. And its small size allows tags to be attached to items that are too small for traditional RFID, such as certain medical devices.
The paper, to be presented at the IEEE Solid-State Circuits Conference, is a collaboration between Han’s group and the Energy Efficient Circuits and Systems Group of Anantha P. Chandrakasan, MIT’s chief innovation and strategy officer and dean of the MIT School of Engineering. , and Professor Vannevar Bush of EECS. Co-authors include EECS graduate students Xibi Chen, Maitryi Ashok, and Jaeyon Won.
Tamper proof
This research project was partly inspired by Han’s favorite car wash. The company placed an RFID tag on his windshield to verify his car wash membership. For added security, the tags are made of fragile paper, so dishonest customers could destroy the tags if they tried to remove them and place them on another window.
However, this is not a very reliable way to prevent tampering. For example, someone could use a solution to dissolve glue and safely remove fragile tags.
A better security solution than authenticating the tag, Han says, is authenticating the item itself. To achieve this, the researchers targeted the adhesive at the interface between the tag and the item’s surface.
Tamper-evident tags contain a series of tiny slots that allow terahertz waves to pass through the tag and attack tiny metal particles mixed into the adhesive.
Terahertz waves are small enough to detect particles, while larger radio waves are not sensitive enough to see them. Additionally, using terahertz waves with a wavelength of 1 mm allowed researchers to create chips that did not require larger off-chip antennas.
After passing through the tag and hitting the object surface, the terahertz waves are reflected or backscattered to a receiver for authentication. The way these waves are backscattered depends on the distribution of metal particles that reflect the waves.
The researchers placed multiple slots on the chip to allow waves to hit different points on the object’s surface, capturing more information about the random distribution of particles.
“As long as the adhesive interface is destroyed by counterfeiters, these reactions are impossible to replicate,” says Han.
Suppliers first read tamper-evident tags attached to items, then store that data in the cloud and use it for later verification.
AI for Authentication
But when it came time to test the tamper-evident tags, Lee faced a problem. It is very difficult and time consuming to make sufficiently precise measurements to ensure that two adhesive patterns match.
He contacted a friend at MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) and together they solved the problem using AI. They trained a machine learning model that could compare adhesive patterns and calculate similarity with over 99% accuracy.
“One downside is that we have limited data samples for this demonstration, but deploying these tags in large quantities across the supply chain will provide us with much more data samples to improve neural networks in the future,” says Lee. .
Authentication systems are also limited by the fact that sensors need to be only about 4 centimeters away from the tag to get accurate readings, as terahertz waves suffer high levels of loss during transmission. This distance is not a problem for applications such as barcode scanning, but is too short for some potential uses, such as automated highway toll booths. Additionally, the angle between the sensor and tag must be less than 10 degrees. Otherwise the terahertz signal will be degraded too much.
They plan to address these limitations in future research, and hope to inspire other researchers to be more optimistic about what can be achieved with terahertz waves, despite the many technical challenges, Han said.
“One thing we really want to show here is that applications of the terahertz spectrum can go far beyond broadband wireless. In this case, terahertz can be used for identity, security, and authentication. “There are a lot of possibilities there,” he added.
This research is supported in part by the U.S. National Science Foundation and the Korea Foundation for Advanced Study.