An Easier Way to Make Reliable Glucose Sensors

Using semiconductor manufacturing methods could help produce a stable, affordable glucose sensor.

by Jeff O'Heir
April 10, 2017

Many of the world’s 387 million diabetics maintain their blood sugar levels with the help of electromechanical sensors that detect blood glucose. The sensors fall into two categories, enzymatic and nonenzymatic, each with their own set of drawbacks. Gou-Jen Wang says he has a better option.

Wang, who’s the dean of Taiwan’s National Chung Hsing University’s College of Engineering, is currently working with Phoenix Semiconductor to manufacture a new nonenzymatic chip-based biosensor that he says has a longer storage life and is easier and cheaper to manufacture than most other glucose sensors, and the research was published in ASME's 2015 Proceedings of the 9th International Conference on Micro- and Nanosystems.

The sensors are fabricated into strips, and they work the same as most commercially available glucose strips. What’s more, the new strip can be inserted into most glucose meters. Wang’s team is currently developing their own meter and an app that diabetics can use to track information on their blood glucose level over a mobile device. 

In a paper presented at ASME’s International Design Engineering Technical Conference in Boston, Wang and his team outlined problems with today’s electromechanical glucose detectors, and how his team overcame them.

Viability of Biosensors

A challenge with enzymatic biosensors is that the enzyme used to measure blood glucose, called glucose oxidase, degrades by up to 25 percent once it’s exposed to air. This limits the life of the biosensor to about one month. Manufacturing uniform enzymatic biosensors is also tricky, Wang explained.

Nonenzymatic biosensors, on the other hand, typically use gold nanoparticles that are more stable than glucose oxidase and don’t degrade as quickly. But preparing the nanoparticles is complicated, relatively expensive, and time-consuming, which limits the commercial viability of nonenzymatic biosensors.

“It’s not suitable for mass production,” Wang said.

Semiconductor Techniques

As an alternative method of preparing a nonenzymatic glucose sensor, Wang and his team turned to semiconductor fabrication techniques. They used a lithographic process in which Photoresist AZ-1518 was spin-coated onto a silicon wafer. The photoresist was then melted to form hemispheres. A thin film of gold was then sputtered onto the hemisphere array to make the sensing electrode.

The finished sensors, Wang said, are highly sensitive and degrade up to one percent, which makes them far more stable than an enzymatic sensor. Wang estimates each sensor would cost about $0.50 each, compared with about $0.80 for the enzymatic version.

The process can yield about 50 chips on a 6-inch wafer. Wang is now working with Phoenix to produce 2,000 chips on an 8-inch wafer.

“Our niche is that we use very simple semiconductor techniques and the quality over every chip is very uniform. That’s important, he said.”

Pending further testing and FDA approval, Wang said the new sensors could hit the market in two to three years. But even if the strips are approved, the technology will face another challenge on the road to the mainstream: competition. Several major companies dominate the glucose-monitoring market. Going up against them will be difficult.

“People always recognize the brand name. That’s the big problem,” Wang said. “But if a major company thinks this is a good process for them, then we would like to collaborate.”


Jeff O’Heir is a science and technology writer based in Huntington, New York.