Sony Image Sensor Business Re-Org

Sony Semiconductor Solutions Japan announces a re-org:

• Establish an in-vehicle product division.
• Reorganize the wafer line technology division and the Kumamoto, Nagasaki, Yamagata, and Oita production divisions by function, and establish a new wafer production division, manufacturing technology division, assembly production division, and quality control division.
• The head of the mobile sector will be newly established as a position to manage all production activities (design, development, and manufacturing) related to the mobile image sensor business at the corporate level.
• Establish a new head of IS / Automotive Sector as a position to manage at the corporate level overall production activities (design, development, and manufacturing) related to the image sensor business other than for mobile devices and the automotive image sensor business.
• A new 300-cm production center manager will be established to manage the entire production of 300-mm wafers at the Kumamoto, Nagasaki, Yamagata, and Oita Technology Centers.

Sony Imaging and Sensing Technology page presents the company's vision of its key advantages:

KAIST Presents InGaAs on SOI Integration

SemiconductorToday notes KAIST paper in Electron Device Letters "Monolithic 3D Integration of InGaAs Photodetectors on Si MOSFETs Using Sequential Fabrication Process" by Dae-Myeong Geum, Seong Kwang Kim, Subin Lee, Donghwan Lim, Hyung-Jun Kim, Chang Hwan Choi, and Sang-Hyeon Kim proposing a stacked process for SWIR pixel:

"We demonstrated the monolithic 3D (M3D) integration of InGaAs photodetectors (PDs) on silicon-on-insulator metal-oxide-semiconductor field-effect transistors (SOI-MOSFETs) by the sequential process of InGaAs PDs on pre-fabricated SOI-MOSFETs. InGaAs PDs and SOI MOSFETs showed their original performances after integration and sequential process thanks to the low-temperature process. In addition, the integrated devices successfully performed the fundamental readout circuit operation such as direct injection and source follower per detector (SFD) by connecting transistors (Trs) and PDs. By illuminating 1550 nm laser on InGaAs PDs, the different behaviors of output voltages were clearly obtained according to the Tr’s on/off state. From these results, we believe that this monolithic 3D integration method could be a feasible approach toward high-resolution multicolor imaging systems."

Camera Remains the Most Used Feature in 5G Phones

StrategyAnalytics statistics of Samsung Galaxy Note 10+ 5G users shows that photography is the single most used feature of the device taking more than 20% of the use time.

Correction: The 20.1% part is probably games. Photography is minuscule 1.3%, unfortunately.

Omnivision's Production Restored to 70%

IFNews quotes Laoyaoba.com site saying that Omnivision plans to restore its production capacity to about 70% by the end of February by "optimizing the overtime mode of production line employees." The company expects to return to 100% capacity in March.

Omnivision manufactures 400 to 500 million CMOS sensors per year. The company is said to consumes 200,000 12-inch wafers each year.

EMVA 1288 Explained

EMVA Tube publishes a number of webinars on image sensor imperfections in EMVA 1288 standard, presented by Prof. Bernd Jähne from Heidelberg University, Germany:

Tong Hsing to Double CIS Backend Capacity

Digitimes reports that Taiwan-based CIS backend services company Tong Hsing Electronic Industries has disclosed plans to double monthly production capacity to 150,000-160,000 wafers by end-2020, according to the company's chairman Pierre Chen.

Samsung 0.7um Pixel Features in Techinsights Teardown

Techinsights starts publishing Samsung Galaxy S20 Ultra 5G teardown analysis, including its 43.7MP GH1 sensor:

"The GH1, which uses ISOCELL Plus technology, has a reported resolution of 7968 x 5840 (43.7 MP). The Tetracell color filter array is in use and manipulating the contrast on our optical images we can see an artifact of the Super-PD phase detection autofocus (PDAF) pixel pattern. We have documented Samsung using masked and 2x1 on-chip lens (OCL) PDAF on different 0.8 µm generation sensors. Here at the 0.7 µm generation we expect only 2x1 OCL PDAF will be used, as on the GH1.

One other topic to mention on this world’s first 0.7 µm pixel sensor is the stacked chip interconnect strategy. The through silicon via (TSV) arrays visible in our die photograph show us immediately that Samsung has not yet adopted a Cu-Cu direct bond interconnect strategy as we have observed in Sony and TSMC fabricated imagers."

Quantum Dot Future

IEEE Spectrum: "It’s time for another revolution in imaging technology. This one will be brought to you by the quantum dot, a nanometer-size particle of semiconductor material, which acts much differently from its bulk counterpart."

Anitoa CMOS Sensor Powers Rapid Nucleic Acid Test for New Coronavirus

PRNewswire: Anitoa has demonstrated a solution for rapid on-site nucleic acid test of the 2019 novel coronavirus (2019-nCoV). Anitoa combines a high performance ultra-portable Real Time Polymerase Chain Reaction (qPCR) instrument called Maverick, with an innovative "one-step" reverse transcription qPCR (RT-qPCR) reagent for testing 2019-nCoV.

Maverick is powered by a unique CMOS biosensor chip developed by Anitoa. The CMOS biosensor integrates many sensing and processing functions necessary for qPCR in a single silicon chip, thus significantly reducing the footprint and cost of the device.

Having a R&D center in Hangzhou China, Anitoa tested its solution in hospitals and CDC-labs of China with actual patient samples. Early test results with Maverick solution showed a 97% accuracy, with 99% specificity.

"We are encouraged by the very promising performance results of the Maverick solution in detecting the 2019 new coronavirus. We want to contribute everything we have to help fight and prevent the spread of this terrible and deadly disease," says Zhimin Ding, CEO of Anitoa Systems.

FDSOI Pixel Thesis

CEA-LETI publishes Lina Kadura's PhD Thesis "New FDSOI-based integrated circuit architectures sensitive to light for imaging applications."

"A new type of light sensor called FDPix, composed of one transistor (1T) per pixel is investigated. It consists in co-integrating an FDSOI (Fully-Depleted Silicon-On-Insulator) transistor with a photodiode to enable light sensing through optical back biasing. The absorption of photons and resulting photogenerated charges in the diode will result in a Light Induced VT Shift (LIVS). The LIVS is due to a capacitive coupling between the front and back gate of the FDSOI transistor and represents the key performance metric to be extracted and optimized. In this work, the device behavior in dc and transient domains was thoroughly investigated and modeled. Although not limited to this node, all the devices tested were fabricated using 28nm node FDSOI technology. By means of TCAD simulations and opto-electrical characterization, the device parameters such as Body Factor (BF) and junction profile were optimized to improve its performance. It was found that the FDPix is in fact a dual response sensor. It exhibits a linear response at low light intensity which results in high sensitivity, and a logarithmic response at higher intensities that ensures a high dynamic range (DR) of more than 120dB. The dedicated developed model is implemented in SPICE environment for circuit design. New pixel circuit in analog and digital domain, based on the FDPix were designed, fabricated, and tested. The results obtained and presented in this work, shows the potential of using the FDPix sensor for smart, highly embedded, low power image sensors for More-than-Moore applications."

imec Pixel Technology

imec SPIE AR/VR/MR presentation has a number of interesting slides about its imaging technology:

poLight Reports Multiple Design Wins in Smartwatches

poLight announces that it has received a purchase order of approximately NOK 250,000 of TLens and ASIC drivers. The purchase order is related to the two smartwatch phone design-wins announced earlier in 2020.

“We are still in the early days since the market launch of the two smartwacthes and therefore very pleased to see the customer returning to refill its stock of TLens,” said Øyvind Isaksen, the CEO of poLight.

Following this order, an accumulated order intake of NOK 2,3 millions for TLens and ASIC drivers has so far been ordered related to the two smartwatch phone releases.

The watch targets a growing market tailored for children in China. The watch camera which is used to take pictures is based on the AF function delivered by poLight.

Huawei to Open Image Sensor Design Group in Belgium

Huawei opens a new R&D center in Belgium and one of its charters is image sensor design, as its affiliated headhunter states in his email: "Huawei is looking for .... a 3D Image Sensor Chip Circuit Architect/Manager"

We are looking for Image Sensor Chip Circuit Architect/Manager, and the responsibility includes:

• Develop a image sensor chip readout circuit architecture, decompose the chip system design specifications and perform corresponding design constraints on each readout module; at the same time, coordinate the pixel design team to complete the circuit design and backend implementation.
• Propose related key technologies of circuits to improve the core competitiveness of the chip, including readout noise reduction, power consumption reduction at high frame rate, and high dynamic range.

The job candidates should have "many experience in mass production of image sensor chip products; solves design or engineering problems in the process from sample to mass production."

IPVM: Masks Cause Major Face Recognition Problem in Security Systems

IPVM has tested four facial recognition systems and they all have difficulties with masks:



The same problem has also been reported for smartphone FaceID unlock. But, as the say, "Every Problem is an Opportunity in Disguise." A San Francisco-based company offers photo-printing on protective masks that is claimed to solve the FaceID hurdles:

Quantum Parametric Sorting Proposed to Separate between Signal and Noise

Nature paper "Noise-tolerant single photon sensitive three-dimensional imager" by Patrick Rehain, Yong Meng Sua, Shenyu Zhu, Ivan Dickson, Bharathwaj Muthuswamy, Jeevanandha Ramanathan, Amin Shahverdi, and Yu-Ping Huang from Stevens Institute of Technology, NJ, proposes to solve a fundamental problem of distinguishing between signal and and ambient photons:

"Active imagers capable of reconstructing 3-dimensional (3D) scenes in the presence of strong background noise are highly desirable for many sensing and imaging applications. A key to this capability is the time-resolving photon detection that distinguishes true signal photons from the noise. To this end, quantum parametric mode sorting (QPMS) can achieve signal to noise exceeding by far what is possible with typical linear optics filters, with outstanding performance in isolating temporally and spectrally overlapping noise. Here, we report a QPMS-based 3D imager with exceptional detection sensitivity and noise tolerance. With only 0.0006 detected signal photons per pulse, we reliably reconstruct the 3D profile of an obscured scene, despite 34-fold spectral-temporally overlapping noise photons, within the 6 ps detection window (amounting to 113,000 times noise per 20 ns detection period). Our results highlight a viable approach to suppress background noise and measurement errors of single photon imager operation in high-noise environments."

Albert Theuwissen Reviews ISSCC 2020 - Part 3

Part 3 of ISSCC review starts with Taiwan National Tsing Hua University paper "A 0.8V Multimode Vision Sensor for Motion and Saliency Detection with Ping-Pong PWM Pixel" by Tzu-Hsiang Hsu, Yen-Kai Chen, Jun-Shen Wu, Wen-Chien Ting, Cheng-Te Wang, Chen-Fu Yeh, Syuan-Hao Sie, Yi-Ren Chen, Ren-Shuo Liu, Chung-Chuan Lo, Kea-Tiong Tang, Meng-Fan Chang, and Chih-Cheng Hsieh.

Prophesee and Sony paper "A 1280 x 720 Back-Illuminated Stacked Temporal Contrast Event-based Vision Sensor with 4.86μm Pixels, 1.066GEPS Readout, Programmable Event Rate Controller and Compressive Data Formatting Pipeline" by Thomas Finateu, Atsumi Niwa, Daniel Matolin, Koya Tsuchimoto, Andrea Mascheroni, Etienne Reynaud, Pooria Mostafalu, Frederick Brady, Ludovic Chotard, Florian LeGoff, Hirotsugu Takahashi, Hayato Wakabayashi, Yusuke Oike,
and Christoph Posch presents the latest results of the cooperation: