Black Phosphorus-based Intelligent Image Sensor

Seokhyeong Lee, Ruoming Peng, Changming Wu & Mo Li from U-Dub have published an article in Nature Communications titled "Programmable black phosphorus image sensor for broadband optoelectronic edge computing".

Our blog had advertised a pre-print version of this work back in November 2021: https://image-sensors-world.blogspot.com/2021/11/black-phosphorus-vision-sensor.html.

Abstract: Image sensors with internal computing capability enable in-sensor computing that can significantly reduce the communication latency and power consumption for machine vision in distributed systems and robotics. Two-dimensional semiconductors have many advantages in realizing such intelligent vision sensors because of their tunable electrical and optical properties and amenability for heterogeneous integration. Here, we report a multifunctional infrared image sensor based on an array of black phosphorous programmable phototransistors (bP-PPT). By controlling the stored charges in the gate dielectric layers electrically and optically, the bP-PPT’s electrical conductance and photoresponsivity can be locally or remotely programmed with 5-bit precision to implement an in-sensor convolutional neural network (CNN). The sensor array can receive optical images transmitted over a broad spectral range in the infrared and perform inference computation to process and recognize the images with 92% accuracy. The demonstrated bP image sensor array can be scaled up to build a more complex vision-sensory neural network, which will find many promising applications for distributed and remote multispectral sensing.

It is now peer reviewed and officially published as an open access paper: https://www.nature.com/articles/s41467-022-29171-1

Peer review report and authors' responses are also publicly available. In particular, it is interesting to see the response to some comments and about pixel non-uniformities, material stability during etching and longevity of the sensor prototype. 

Some lightly edited excerpts from the reviews and authors responses below:

Reviewer: The optical image of the exfoliated flake clearly shows regions of varying thickness. How did the authors ensure each pixel is of the same thickness? 

Authors: The mechanically exfoliated bP has several regions with different thicknesses. We fabricated all the pixels within a large region with uniform optical contrast, as outlined by the red dotted line, indicating uniform thickness. The thickness of the region is also confirmed with atomic force microscopy.

Reviewer: There is hardly any characterisation data provided for the material. How much of it is oxidised?

Authors: The oxidation of bP, it is indeed a concern. To mitigate that, we exfoliated and transferred bP in an Ar-filled glovebox. The device was immediately loaded into the atomic layer deposition (ALD) chamber to deposit the Al2O3 / HfO2 /Al2O3 (AHA) multilayers, which encapsulate the bP flake to prevent oxidation and degradation. This has been a practice reported in the literature, which generally leads to oxidation of only a few layers. Thanks to the 35 nm thick AHA encapsulation layer, our device shows long-term stability with persistent electrical and optical properties for more than 3 months after fabrication. We discuss that in the response to question 7. Furthermore, Raman spectroscopy shows no sign of Px Oy or Hx POy forming during the fabrication process. Thus, we expect that the oxidation of bP flake is no more than 3 layers (or 1.5 nm), which, if any, marginally affects the optical and electrical properties of the bP-PPT device. 

Reviewer: Why did the authors focus only on the IR range when the black phosphorus can be even more broadband into the visible at the thickness used here?

Authors: The photoresponsivity of black phosphorus certainly extends to the visible band. We have
utilized both the visible and the IR range by engineering the device with the AHA stack: IR light to input images for optoelectronic in-sensor computing; visible light to optically program the device by activating the trapped charges and process the encoded images such as pattern recognition.

Reviewer: How long do the devices keep working in a stable manner?

Authors: We agree with the reviewer that more lifetime measurement data is important to ensure the
stability of the device’s operation. We have evaluated the performance of the bP-PPT devices over a long period of time (up to 3 months) ... the gate modulation, memory window, on-off ratio, and retention time of our devices remain consistent even 3 months after they were fabricated.

In today's day and age of Twitter, it's refreshing to see how science really progresses behind the scenes --- reviewers raising genuine concerns about a new technique; authors graciously accepting limitations and suggesting improvements and alternative ways forward.