Shang-Fu Yeh
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Shang-Fu Yeh has a total of 53 co-authors. Here are the top co-authors by number of shared publications:
- Calvin Yi-Ping Chao
Taiwan Semiconductor Manufacturing Company (TSMC), Hsinchu 30077, Taiwan
8 shared publications Chih-Cheng Hsieh
Department of Electrical Engineering, National Tsing Hua University, Hsinchu, Taiwan
8 shared publications- Chih-Lin Lee
Taiwan Semiconductor Manufacturing Company (TSMC), Hsinchu 30077, Taiwan
4 shared publications - Honyih Tu
Taiwan Semiconductor Manufacturing Company (TSMC), Hsinchu 30077, Taiwan
4 shared publications Kuo-Yu Chou
Taiwan Semiconductor Manufacturing Company, 9, Creation Rd. 1, Hsinchu Science Park, Hsinchu 300-77, Taiwan, R.O.C.
4 shared publications- Thomas Meng-Hsiu Wu
Taiwan Semiconductor Manufacturing Company (TSMC), Hsinchu 30077, Taiwan
4 shared publications
Latest Publications
- Open Access
High-Temperature Annealing of Random Telegraph Noise in a Stacked CMOS Image Sensor After Hot-Carrier Stress
This paper studies the temperature effects on device aging, particularly the random telegraph noise (RTN) degradation and the threshold voltage () shift in a stacked CMOS image sensor (CIS) caused by hot-carrier stress (HCS). Measurements indicate that both are worse when stressed at lower temperatures. Further, the RTN traps generated by HCS can be deactivated effectively by a subsequent high-temperature annealing at 240 °C for up to 360 min. In contrast, the RTN traps in chips not stressed by hot carriers are essentially unaffected by annealing at the same temperature for the same amount of time. This suggests that the physical structure of the RTN traps caused by process-induced damage (PID) without HCS might be different from that generated by HCS. The exact microscopic nature of the differences between these two kinds of RTN traps is not clear and requires further investigation. This work also suggests that RTN degradation could be a useful indicator for device aging for reliability testing and modeling.Open AccessHigh-Temperature Annealing of Random Telegraph Noise in a Stacked CMOS Image Sensor After Hot-Carrier Stress
This paper studies the temperature effects on device aging, particularly the random telegraph noise (RTN) degradation and the threshold voltage () shift in a stacked CMOS image sensor (CIS) caused by hot-carrier stress (HCS). Measurements indicate that both are worse when stressed at lower temperatures. Further, the RTN traps generated by HCS can be deactivated effectively by a subsequent high-temperature annealing at 240 °C for up to 360 min. In contrast, the RTN traps in chips not stressed by hot carriers are essentially unaffected by annealing at the same temperature for the same amount of time. This suggests that the physical structure of the RTN traps caused by process-induced damage (PID) without HCS might be different from that generated by HCS. The exact microscopic nature of the differences between these two kinds of RTN traps is not clear and requires further investigation. This work also suggests that RTN degradation could be a useful indicator for device aging for reliability testing and modeling. - Open Access
- 1 December 2025
- Published by AIP Publishing in APL Photonics
- Vol. 10 (12)
A novel full-chip afterpulsing evaluation technique for 116 × 160 Ge-on-Si SPAD array
This Letter presents a novel on-chip afterpulsing characterization method demonstrated using gated-mode short-wave infrared single photon avalanche diode arrays, effectively overcoming the measurement limitations of conventional methods. Through this approach, we first achieve a comprehensive evaluation of the full-chip afterpulsing effect by adjusting the ratio between gated-on time and gated-off time and employing a sampling window technique. Experimentally validated using a room-temperature-operated 160 × 116 Ge-on-Si single photon avalanche diode array, the approach clearly reveals the potential pixel-to-pixel variations in afterpulsing behavior across SWIR single-photon avalanche diode arrays, highlighting the critical importance of full-chip evaluation for designing and optimizing gating period.
- 1 December 2025
- Published by AIP Publishing in APL Photonics
- Vol. 10 (12)
Open AccessA novel full-chip afterpulsing evaluation technique for 116 × 160 Ge-on-Si SPAD array
This Letter presents a novel on-chip afterpulsing characterization method demonstrated using gated-mode short-wave infrared single photon avalanche diode arrays, effectively overcoming the measurement limitations of conventional methods. Through this approach, we first achieve a comprehensive evaluation of the full-chip afterpulsing effect by adjusting the ratio between gated-on time and gated-off time and employing a sampling window technique. Experimentally validated using a room-temperature-operated 160 × 116 Ge-on-Si single photon avalanche diode array, the approach clearly reveals the potential pixel-to-pixel variations in afterpulsing behavior across SWIR single-photon avalanche diode arrays, highlighting the critical importance of full-chip evaluation for designing and optimizing gating period.
- Open Access
Random Telegraph Noise Degradation Caused by Hot Carrier Injection in a 0.8 μm-Pitch 8.3Mpixel Stacked CMOS Image Sensor
In this work, the degradation of the random telegraph noise (RTN) and the threshold voltage () shift of an 8.3Mpixel stacked CMOS image sensor (CIS) under hot carrier injection (HCI) stress are investigated. We report for the first time the significant statistical differences between these two device aging phenomena. The shift is relatively uniform among all the devices and gradually evolves over time. By contrast, the RTN degradation is evidently abrupt and random in nature and only happens to a small percentage of devices. The generation of new RTN traps by HCI during times of stress is demonstrated both statistically and on the individual device level. An improved method is developed to identify RTN devices with degenerate amplitude histograms.Open AccessRandom Telegraph Noise Degradation Caused by Hot Carrier Injection in a 0.8 μm-Pitch 8.3Mpixel Stacked CMOS Image Sensor
In this work, the degradation of the random telegraph noise (RTN) and the threshold voltage () shift of an 8.3Mpixel stacked CMOS image sensor (CIS) under hot carrier injection (HCI) stress are investigated. We report for the first time the significant statistical differences between these two device aging phenomena. The shift is relatively uniform among all the devices and gradually evolves over time. By contrast, the RTN degradation is evidently abrupt and random in nature and only happens to a small percentage of devices. The generation of new RTN traps by HCI during times of stress is demonstrated both statistically and on the individual device level. An improved method is developed to identify RTN devices with degenerate amplitude histograms. - Open Access
- 18 October 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of the Electron Devices Society
- Vol. 9, 972-984
Statistical Analysis of Random Telegraph Noises of MOSFET Subthreshold Currents Using a 1M Array Test Chip in a 40 nm Process
It is difficult to measure the random telegraph noises (RTN) of MOSFET subthreshold currents at the sub-pA level directly and accurately. In this work, we used a charge integration method similar to the operation of the CMOS image sensors (CIS) to characterize the RTN of subthreshold currents approximately from 1 fA to 1 nA, using a test chip of 1M cell array in a 40 nm process. We found that each RTN trap was active only within a specific window of gate voltages. The trap became less active or inactive outside the corresponding window of operations. We showed that the sets of RTN-active devices under different gate voltages were different. Furthermore, the choice of sampling frequency in measuring RTN and the number of sampled data points determined the observable range of RTN emission and capture time constants. For the data measured by sampling periods of 3.82 s, 299 ms, and 372 us, different sets of RTN traps were observed with different spans of time constants. The combined time constants range was about 7 orders of magnitude. For single-trap RTN, we found and verified a relation between the probability of trap occupancy (PTO) and the ratio of root-mean-square random noise (RN) versus the RTN amplitude.
- 18 October 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of the Electron Devices Society
- Vol. 9, 972-984
Open AccessStatistical Analysis of Random Telegraph Noises of MOSFET Subthreshold Currents Using a 1M Array Test Chip in a 40 nm Process
It is difficult to measure the random telegraph noises (RTN) of MOSFET subthreshold currents at the sub-pA level directly and accurately. In this work, we used a charge integration method similar to the operation of the CMOS image sensors (CIS) to characterize the RTN of subthreshold currents approximately from 1 fA to 1 nA, using a test chip of 1M cell array in a 40 nm process. We found that each RTN trap was active only within a specific window of gate voltages. The trap became less active or inactive outside the corresponding window of operations. We showed that the sets of RTN-active devices under different gate voltages were different. Furthermore, the choice of sampling frequency in measuring RTN and the number of sampled data points determined the observable range of RTN emission and capture time constants. For the data measured by sampling periods of 3.82 s, 299 ms, and 372 us, different sets of RTN traps were observed with different spans of time constants. The combined time constants range was about 7 orders of magnitude. For single-trap RTN, we found and verified a relation between the probability of trap occupancy (PTO) and the ratio of root-mean-square random noise (RN) versus the RTN amplitude.
- Open Access
Random Telegraph Noises from the Source Follower, the Photodiode Dark Current, and the Gate-Induced Sense Node Leakage in CMOS Image Sensors
In this paper we present a systematic approach to sort out different types of random telegraph noises (RTN) in CMOS image sensors (CIS) by examining their dependencies on the transfer gate off-voltage, the reset gate off-voltage, the photodiode integration time, and the sense node charge retention time. Besides the well-known source follower RTN, we have identified the RTN caused by varying photodiode dark current, transfer-gate and reset-gate induced sense node leakage. These four types of RTN and the dark signal shot noises dominate the noise distribution tails of CIS and non-CIS chips under test, either with or without X-ray irradiation. The effect of correlated multiple sampling (CMS) on noise reduction is studied and a theoretical model is developed to account for the measurement results.
Open AccessRandom Telegraph Noises from the Source Follower, the Photodiode Dark Current, and the Gate-Induced Sense Node Leakage in CMOS Image Sensors
In this paper we present a systematic approach to sort out different types of random telegraph noises (RTN) in CMOS image sensors (CIS) by examining their dependencies on the transfer gate off-voltage, the reset gate off-voltage, the photodiode integration time, and the sense node charge retention time. Besides the well-known source follower RTN, we have identified the RTN caused by varying photodiode dark current, transfer-gate and reset-gate induced sense node leakage. These four types of RTN and the dark signal shot noises dominate the noise distribution tails of CIS and non-CIS chips under test, either with or without X-ray irradiation. The effect of correlated multiple sampling (CMS) on noise reduction is studied and a theoretical model is developed to account for the measurement results.
- Open Access
- 15 January 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of the Electron Devices Society
- Vol. 7 (1) , 227-238
Random Telegraph Noises in CMOS Image Sensors Caused by Variable Gate-Induced Sense Node Leakage Due to X-Ray Irradiation
- 15 January 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of the Electron Devices Society
- Vol. 7 (1) , 227-238
Open AccessRandom Telegraph Noises in CMOS Image Sensors Caused by Variable Gate-Induced Sense Node Leakage Due to X-Ray Irradiation
- 1 February 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 88-90
A 1.1μm-Pitch 13.5Mpixel 3D-stacked CMOS image sensor featuring 230fps full-high-definition and 514fps high-definition videos by reading 2 or 3 rows simultaneously using a column-switching matrix
Slow-motion video is a desirable feature for state-of-art smartphones. The effect is achieved by capturing a video at a higher frame rate and playing it back at a lower frame rate. While the still-image resolution of smartphone cameras ranges from 8MP to 25MP, standard videos are limited to 3 formats: 3840×2160 (4K2K, 2160p), 1920×1080 (Full High Definition, FHD, 1080p), and 1280×720 (High Definition, HD, 720p). CMOS image sensors using various column-parallel aDc architectures have been reported to reach high frame rates [1-5]. The single-slope (SS) ADC is an attractive choice for a balanced performance among high speed, low noise, small area, and low power consumption. However, in conventional SS ADC design, each ADC is hardwired to a column signal line. ADCs for skipped columns are left idle during the subsampling operation, and the potential to reach higher frame rate is not optimized. In this paper, we develop an approach in which all the column ADCs are fully utilized in both of the 2-to-1 and 3-to-1 subsampling modes, such that the maximum of 4x faster FHD and 9x faster HD videos are demonstrated with reference to the 1-to-1 non-subsampled 4K2K video.
- 1 February 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 88-90
A 1.1μm-Pitch 13.5Mpixel 3D-stacked CMOS image sensor featuring 230fps full-high-definition and 514fps high-definition videos by reading 2 or 3 rows simultaneously using a column-switching matrix
Slow-motion video is a desirable feature for state-of-art smartphones. The effect is achieved by capturing a video at a higher frame rate and playing it back at a lower frame rate. While the still-image resolution of smartphone cameras ranges from 8MP to 25MP, standard videos are limited to 3 formats: 3840×2160 (4K2K, 2160p), 1920×1080 (Full High Definition, FHD, 1080p), and 1280×720 (High Definition, HD, 720p). CMOS image sensors using various column-parallel aDc architectures have been reported to reach high frame rates [1-5]. The single-slope (SS) ADC is an attractive choice for a balanced performance among high speed, low noise, small area, and low power consumption. However, in conventional SS ADC design, each ADC is hardwired to a column signal line. ADCs for skipped columns are left idle during the subsampling operation, and the potential to reach higher frame rate is not optimized. In this paper, we develop an approach in which all the column ADCs are fully utilized in both of the 2-to-1 and 3-to-1 subsampling modes, such that the maximum of 4x faster FHD and 9x faster HD videos are demonstrated with reference to the 1-to-1 non-subsampled 4K2K video.
- Open Access
Statistical Analysis of the Random Telegraph Noise in a 1.1 μm Pixel, 8.3 MP CMOS Image Sensor Using On-Chip Time Constant Extraction Method
A study of the random telegraph noise (RTN) of a 1.1 μm pitch, 8.3 Mpixel CMOS image sensor (CIS) fabricated in a 45 nm backside-illumination (BSI) technology is presented in this paper. A noise decomposition scheme is used to pinpoint the noise source. The long tail of the random noise (RN) distribution is directly linked to the RTN from the pixel source follower (SF). The full 8.3 Mpixels are classified into four categories according to the observed RTN histogram peaks. A theoretical formula describing the RTN as a function of the time difference between the two phases of the correlated double sampling (CDS) is derived and validated by measured data. An on-chip time constant extraction method is developed and applied to the RTN analysis. The effects of readout circuit bandwidth on the settling ratios of the RTN histograms are investigated and successfully accounted for in a simulation using a RTN behavior model.
Open AccessStatistical Analysis of the Random Telegraph Noise in a 1.1 μm Pixel, 8.3 MP CMOS Image Sensor Using On-Chip Time Constant Extraction Method
A study of the random telegraph noise (RTN) of a 1.1 μm pitch, 8.3 Mpixel CMOS image sensor (CIS) fabricated in a 45 nm backside-illumination (BSI) technology is presented in this paper. A noise decomposition scheme is used to pinpoint the noise source. The long tail of the random noise (RN) distribution is directly linked to the RTN from the pixel source follower (SF). The full 8.3 Mpixels are classified into four categories according to the observed RTN histogram peaks. A theoretical formula describing the RTN as a function of the time difference between the two phases of the correlated double sampling (CDS) is derived and validated by measured data. An on-chip time constant extraction method is developed and applied to the RTN analysis. The effects of readout circuit bandwidth on the settling ratios of the RTN histograms are investigated and successfully accounted for in a simulation using a RTN behavior model.
- 10 November 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Solid-State Circuits
- Vol. 53 (2) , 527-537
A 0.66erms− Temporal-Readout-Noise 3-D-Stacked CMOS Image Sensor With Conditional Correlated Multiple Sampling Technique
This paper presents a sub-electron temporal readout noise, 8.3 Mpixel and 1.1-μm pixel pitch 3-D-stacked CMOS image sensor (CIS). A conditional correlated multiple sampling (CMS) technique is introduced to selectively reduce the dark pixel noise by using a full-range ramp and a small-range ramp. In this way, a sub-electron temporal readout noise CIS is achieved without degrading the frame rate dramatically, compared to the conventional CMS method. A column-parallel single slope ADC with dark pixel detection function is proposed as well. A dynamic-dark-signal-region detection technique is used to mitigate differential nonlinearity (DNL) errors due to ramp slope mismatch. The implemented prototype in 45-nm CIS/65-nm CMOS occupies an area of 35.89 mm². This paper achieves a 0.66e-rms with 5-time sampling at a frame rate of 7.2 frames/s, which corresponds to a sample-rate frequency of 36.1 kHz for the column ADC. The DNL (11 b) is improved from +0.98 LSB/-0.94 LSB to +0.29 LSB/-0.39 LSB by using dynamic dark-signal region technique. The figure of merit of this paper is 2.02 nVrms/Hz.
- 10 November 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Solid-State Circuits
- Vol. 53 (2) , 527-537
A 0.66erms− Temporal-Readout-Noise 3-D-Stacked CMOS Image Sensor With Conditional Correlated Multiple Sampling Technique
This paper presents a sub-electron temporal readout noise, 8.3 Mpixel and 1.1-μm pixel pitch 3-D-stacked CMOS image sensor (CIS). A conditional correlated multiple sampling (CMS) technique is introduced to selectively reduce the dark pixel noise by using a full-range ramp and a small-range ramp. In this way, a sub-electron temporal readout noise CIS is achieved without degrading the frame rate dramatically, compared to the conventional CMS method. A column-parallel single slope ADC with dark pixel detection function is proposed as well. A dynamic-dark-signal-region detection technique is used to mitigate differential nonlinearity (DNL) errors due to ramp slope mismatch. The implemented prototype in 45-nm CIS/65-nm CMOS occupies an area of 35.89 mm². This paper achieves a 0.66e-rms with 5-time sampling at a frame rate of 7.2 frames/s, which corresponds to a sample-rate frequency of 36.1 kHz for the column ADC. The DNL (11 b) is improved from +0.98 LSB/-0.94 LSB to +0.29 LSB/-0.39 LSB by using dynamic dark-signal region technique. The figure of merit of this paper is 2.02 nVrms/Hz.
- 1 June 2015
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 21585601,p. C84-C85
A 0.66e−rms temporal-readout-noise 3D-stacked CMOS image sensor with conditional correlated multiple sampling (CCMS) technique
A conditional correlated multiple sampling (CCMS) technique for low noise CMOS image sensor (CIS) is proposed to reduce noise and address low frame rate issue caused by the conventional correlated multiple sampling (CMS) technique. An 8Mpixel 3D-stacked CIS with 1.1um pixel pitch is designed and verified. Measurement results show this technique can achieve 0.66e - rms at 36.1 kHz A/D sampling rate per pixel with analog gain at 16 and 5-times multiple sampling. The resulting DNL is within -0.49/+0.45LSB.
- 1 June 2015
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 21585601,p. C84-C85
A 0.66e−rms temporal-readout-noise 3D-stacked CMOS image sensor with conditional correlated multiple sampling (CCMS) technique
A conditional correlated multiple sampling (CCMS) technique for low noise CMOS image sensor (CIS) is proposed to reduce noise and address low frame rate issue caused by the conventional correlated multiple sampling (CMS) technique. An 8Mpixel 3D-stacked CIS with 1.1um pixel pitch is designed and verified. Measurement results show this technique can achieve 0.66e - rms at 36.1 kHz A/D sampling rate per pixel with analog gain at 16 and 5-times multiple sampling. The resulting DNL is within -0.49/+0.45LSB.