Microelectronics have come a long way in the last 40 years thanks to CMOS (complementary metal-oxide semiconductor) technology. However, this platform has been difficult to integrate with semiconductors other than silicon because of lattice mismatch between different materials. Being unable to integrate other semiconductors limits the performance of CMOS-based cameras, explains co-team leader Franck Koppens. "For example, the camera in your smartphone can only see visible light as silicon only absorbs visible light."

Koppens and Gerasimov Konstantatos have now overcome this problem by monolithically integrating a CMOS integrated circuit with graphene for the first time and making a high-resolution image sensor from the ensemble. The device consists of a 388 x 288 array of graphene-quantum dot photodetectors that operate as a digital camera sensitive to both visible and short-wave IR light – something that has never been achieved before in such image sensors. It contains around 110,000 photoconductive graphene channels that are all integrated vertically and connected to the individual electronic components of a CMOS readout integrated circuit. The chip circuitry very much resembles that used in the commercial image sensors commonly found in smartphones.

Phototransistors work in the UV, VIS and IR

“We made our device by first purchasing an off-the-shelf wafer containing many chips with CMOS read-out circuits,” explains team member and first author of the study Stijn Goossens. “We then cut the wafer into single chips that each have the connections for 100,000 pixels and the electronic circuitry to read those pixels. Next, we transferred a sheet of graphene (grown in a chemical vapour deposition reactor) on top of these chips. After that, we patterned the graphene and contacted it to the underlying circuitry. Finally, we made the graphene sensitive to light by applying a layer of colloidal quantum dots using a simple spin-coating process.”

By combining graphene, which is a high-mobility material, and quantum dots, which act as the light-absorbing components, we made phototransistors that can convert light into electronic signals. “The transistors work in the UV, VIS and IR since quantum dots can be made to absorb light in all of these regions of the electromagnetic spectrum.”

Wide range of surveillance applications

The researchers say that they are able to convert incoming light into electronic signals from each of the pixels and then build up an image by reading out of all of these. “This happens automatically inside the chip,” explains Koppens.

The CMOS-integrated device might be used in a wide range of applications, from low-cost pocket and smartphone cameras, fire-control systems, passive night vison and night surveillance cameras as well as environmental monitoring. Other application areas include automotive sensor systems, medical imaging and inspecting food and pharmaceutics.

The team, reporting its work in Nature Photonics doi:10.1038/nphoton.2017.75, says that it will now be working on increasing the resolution of the image sensor so that it is on a par with current high-definition visible cameras. “We then intend to expand the spectral coverage of the device so that it can operate even deeper into the IR – the mid-IR or even possibly the long-wavelength IR,” Konstantatos tells nanotechweb.org. “Doing this is of paramount importance since, in that part of the spectrum, there is extremely important information related to the composition of even the temperature of the objects being imaged.

“Ultimately, we would like to produce a versatile, low-cost and highly sensitive hyperspectral imager that can provide the maximum amount of information concerning a scene, not simply a 2D image but something well beyond that.”