Skip to main content

Combining graphene with quantum dots makes the most broadband CMOS image sensor

By Xiwen Gong August 11, 2017
Complementary metal-oxide–semiconductor (CMOS) integration of chemical vapor deposited (CVD) graphene with 388 × 288 pixel image sensor read-out circuit. DAC, digital-to-analog converter. Credit: Nature Photonics

An image sensor, which converts light signals into electrical signals, is the key component of a digital camera. Sensors based on complementary metal-oxide-semiconductors (CMOS) are commonly used in cameras for their high efficiency and superior performance in the high-speed mode. Now researchers have introduced a graphene-quantum-dots-integrated CMOS image sensor which greatly expands the detection range.

While silicon is currently the only material that can be integrated in CMOS, its relatively wide bandgap (1.1 eV) dramatically limits the bandwidth for sensing infrared light. Researchers from Institut de Ciencies Fotoniques (ICFO), Graphenea SA, and Institució Catalana de Recerça i Estudis Avançats (ICREA), in work recently published in Nature Photonics, offer a prototype digital camera made from the graphene-quantum dot photodetectors that can operate from the ultraviolet to the visible to the near-infrared range.

“This is the first demonstration of graphene integrated on a CMOS integrated circuit, in this case operating as a focal plane array,” says Frank Koppens of ICFO and ICREA, the group leader of one of the research teams. “The graphene-integrated CMOS image sensor can operate from 300 nm to 2000 nm. So far, there has not been any commercial technology that covers the wavelength range that we have achieved. For many applications, different cameras need to be combined together, while we provide a technology that requires only one camera.”

This broadband image sensor has two key components: the graphene as a high-mobility carrier transporting material, and quantum dots as the light absorbing material. The incident light is first absorbed by the quantum dots, which generate the charge carriers that then transfer them to the graphene sheet. The charge injection from the quantum dots modulates the conductivity of graphene, by which means the incoming light signal is converted into electronic signals. Benefited from the ultrahigh mobility of graphene, the sensitivity of the sensor is high (combining the high responsivity with low noise), which means that less amount of light is needed to generate a strong signal.

The researchers fabricated more than 100,000 of these phototransistors on a small chip (see b-d in the Figure), which is similar to the chip inside the digital camera in a smart phone. These photodetectors are then connected to the CMOS electronics (containing thousands of transistors) inside this chip. In this way, the incoming light is converted into electronic signals for each pixel, which then builds up an image by reading out all of them, driven automatically by the chip.

“Even more important, the cost is lower than the current existing infrared imaging system,” says Gerasimos Konstantatos of ICFO and ICREA, the group leader of the team focusing on infrared quantum dots in this study. “Our graphene-quantum-dot technology is based on monolithic integration with Si-CMOS, which will lower the cost of the system.”

Sjoerd Hoogland, an expert in low-dimensional materials for optoelectronics at the University of Toronto, explains the importance of this technological advance: “Although graphene/quantum dot hybrid photodetectors have been demonstrated before, the implementation of this concept in a CMOS architecture breaks an important technological barrier. Achieving uniform, large-area and high-quality graphene over an entire CMOS read-out circuit is not trivial; attaining a good sensitivity, reproducible across the entire sensor array after the integration with the quantum dots, is also very challenging. Dr. Stijn Goossens and colleagues have successfully solved these problems, bringing this powerful device architecture into something tangible: a sensitive broadband infrared camera.”

For the next step, the research team will try to make a mature prototype by designing an image sensor circuit with the sensitivity improved by a factor of 100 at least. The researchers are also working on a production process that is wafer-scale. Koppens says, “To achieve this goal still requires solving quite some challenges related to the large-area transfer of graphene, and processing in a CMOS-pilot line while controlling the contamination. Those integration challenges are also a big activity within the European Graphene Flagship program, and it’s expected that volume production will be possible in the coming years.”

Read the abstract in Nature Photonics.