A number of years ago, MIT researchers invented a cryptographic ID tag that’s several times smaller and significantly cheaper than the normal radio frequency tags (RFIDs) which are often affixed to products to confirm their authenticity.

This tiny tag, which offers improved security over RFIDs, utilizes terahertz waves, that are smaller and travel much faster than radio waves. But this terahertz tag shared a serious security vulnerability with traditional RFIDs: A counterfeiter could peel the tag off a real item and reattach it to a fake, and the authentication system could be none the wiser.

The researchers have now surmounted this security vulnerability by leveraging terahertz waves to develop an antitampering ID tag that also offers the advantages of being tiny, low cost, and secure.

They mix microscopic metal particles into the glue that sticks the tag to an object, after which use terahertz waves to detect the unique pattern those particles form on the item’s surface. Akin to a fingerprint, this random glue pattern is used to authenticate the item, explains Eunseok Lee, an electrical engineering and computer science (EECS) graduate student and lead writer of a paper on the antitampering tag.

“These metal particles are essentially like mirrors for terahertz waves. If I spread a bunch of mirror pieces onto a surface after which shine light on that, depending on the orientation, size, and site of those mirrors, I’d get a special reflected pattern. But should you peel the chip off and reattach it, you destroy that pattern,” adds Ruonan Han, an associate professor in EECS, who leads the Terahertz Integrated Electronics Group within the Research Laboratory of Electronics.

The researchers produced a light-powered antitampering tag that’s about 4 square millimeters in size. In addition they demonstrated a machine-learning model that helps detect tampering by identifying similar glue pattern fingerprints with greater than 99 percent accuracy.

Since the terahertz tag is so low cost to supply, it could possibly be implemented throughout a large supply chain. And its tiny size enables the tag to connect to items too small for traditional RFIDs, comparable to certain medical devices.

The paper, which can be presented on 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, dean of the MIT School of Engineering, and the Vannevar Bush Professor of EECS. Co-authors include EECS graduate students Xibi Chen, Maitryi Ashok, and Jaeyeon Won.

Stopping tampering

This research project was partly inspired by Han’s favorite automotive wash. The business stuck an RFID tag onto his windshield to authenticate his automotive wash membership. For added security, the tag was created from fragile paper so it might be destroyed if a less-than-honest customer tried to peel it off and stick it on a special windshield.

But that will not be a really reliable method to prevent tampering. For example, someone could use an answer to dissolve the glue and safely remove the delicate tag.

Somewhat than authenticating the tag, a greater security solution is to authenticate the item itself, Han says. To attain this, the researchers targeted the glue on the interface between the tag and the item’s surface.

Their antitampering tag comprises a series of miniscule slots that enable terahertz waves to go through the tag and strike microscopic metal particles which have been mixed into the glue.

Terahertz waves are sufficiently small to detect the particles, whereas larger radio waves wouldn’t have enough sensitivity to see them. Also, using terahertz waves with a 1-millimeter wavelength allowed the researchers to make a chip that doesn’t need a bigger, off-chip antenna.

After passing through the tag and striking the thing’s surface, terahertz waves are reflected, or backscattered, to a receiver for authentication. How those waves are backscattered will depend on the distribution of metal particles that reflect them.

The researchers put multiple slots onto the chip so waves can strike different points on the thing’s surface, capturing more information on the random distribution of particles.

“These responses are inconceivable to duplicate, so long as the glue interface is destroyed by a counterfeiter,” Han says.

A vendor would take an initial reading of the antitampering tag once it was stuck onto an item, after which store those data within the cloud, using them later for verification.

AI for authentication

But when it got here time to check the antitampering tag, Lee bumped into an issue: It was very difficult and time-consuming to take precise enough measurements to find out whether two glue patterns are a match.

He reached out to a friend within the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL) and together they tackled the issue using AI. They trained a machine-learning model that would compare glue patterns and calculate their similarity with greater than 99 percent accuracy.

“One drawback is that we had a limited data sample for this demonstration, but we could improve the neural network in the longer term if a lot of these tags were deployed in a supply chain, giving us rather a lot more data samples,” Lee says.

The authentication system can also be limited by the proven fact that terahertz waves suffer from high levels of loss during transmission, so the sensor can only be about 4 centimeters from the tag to get an accurate reading. This distance wouldn’t be a difficulty for an application like barcode scanning, however it could be too short for some potential uses, comparable to in an automatic highway toll booth. Also, the angle between the sensor and tag must be lower than 10 degrees or the terahertz signal will degrade an excessive amount of.

They plan to deal with these limitations in future work, and hope to encourage other researchers to be more optimistic about what may be completed with terahertz waves, despite the various technical challenges, says Han.

“One thing we actually need to point out here is that the appliance of the terahertz spectrum can go well beyond broadband wireless. On this case, you need to use terahertz for ID, security, and authentication. There are a number of possibilities on the market,” he adds.

This work is supported, partly, by the U.S. National Science Foundation and the Korea Foundation for Advanced Studies.