† Corresponding author. E-mail:
Project supported by the National Key Research and Development Program of China (Grant No. NKRDP 2016YFA0202101).
Photo-generated carriers may diffuse into the adjacent cells to form the electrical crosstalk, which is especially noticeable after the pixel cell size has been scaled down. The electrical crosstalk strongly depends on the structure and electrical properties of the photosensitive areas. In this work, time-dependent crosstalk effects considering different isolation structures are investigated. According to the different depths of photo-diode (PD) and isolation structure, the transport of photo-generated carriers is analyzed with different regions in the pixel cell. The evaluation of crosstalk is influenced by exposure time. Crosstalk can be suppressed by reducing the exposure time. However, the sensitivity and dynamic range of the image sensor need to be considered as well.
Image sensors are widely used in military, medical, automotive, and mobile devices.[1] With the continuous increase of application areas, the requirement for imaging quality becomes higher and the size of a pixel cell keeps on scaling down.[2–6] Optical or electrical crosstalk may occur between adjacent pixel cells.[7,8] Photons enter into the adjacent pixel cells before they are absorbed by the photosensitive areas, thus producing optical crosstalk. The optical crosstalk can be suppressed by optimizing the interconnected structures above the devices or using the backside illumination structures.[9–11] The photons absorbed by the photosensitive areas and converted into photo-generated carriers may diffuse into the adjacent cells to form the electrical crosstalk. The optimization of photosensitive areas and the introduction of isolation structures could restrain the electrical crosstalk.[7,8,12,13] The crosstalk would reduce the quantum efficiency indirectly and affect the imaging quality. This crosstalk is especially noticeable after the pixel cell size has been scaled down to approximately one micrometer.
The electrical crosstalk strongly depends on the structure and electrical properties of the photosensitive area. Deep trench isolation (DTI)[7,8,13,14] and heavy doping isolation[15,16] are employed to resist the electrical crosstalk currently. In smaller pixel cells, the width of the isolation structure also requires a corresponding reduction, which becomes a greater challenge to the process. Therefore, the influence of crosstalk needs to be systematically studied for designing and optimizing the pixel structures with different isolation structures. Different methods have been employed to study the crosstalk between pixel cells. Some researchers directly measured the crosstalk experimentally.[7,8] Due to the small number of experimental samples, the influences of pixel size and isolation structures on crosstalk may not be systematically studied. In addition, the exposure time is usually fixed experimentally. Hence the time-dependent crosstalk effect cannot be studied. On the other hand, some researches have been conducted using models or simulations.[17–20] However, most of the models did not consider isolation structures. For small pixel cells, the large aspect ratio of the photosensitive area makes the isolation structure indispensable, and the crosstalk models should consider the influence of the isolation structure. Moreover, most of the crosstalk models and simulations are the steady-state solutions of drift-diffusion equations. However, in the actual image sensor, the electrical crosstalk is the time-dependent diffusion process of photo-generated carriers. Especially for the case of low light intensity, the steady state cannot be reached within the exposure time. The influence of exposure time should be considered in the study of crosstalk.
In this work, time-dependent crosstalk effects considering different isolation structures are proposed. According to the different depths of PD and isolation structures, the transport of photo-generated carrier is modeled with different regions in the pixel cell. By using the technology computer aided design (TCAD) tool, the time-dependent crosstalk effects are evaluated for the carriers’ transport with incident light of different wavelengths, and the isolation structures with different DTI depths. The results can help to design and optimize the image sensor especially for scaling down the pixel size.
The main area of a pixel cell is the photosensitive area. Therefore, a photosensitive structure with adjacent cells is analyzed as shown in Fig.
According to the different depths of pinning layer (PIN), PD and isolation structure, the pixel cell can be partitioned into several regions as shown in Figs.
Figure
Electrons in e1: most recombined in pinning layer.
Electrons in e2: diffusion to the depletion zone under the influence of concentration gradient.
Electrons in e3, e4, and e6: accumulated in the depletion zone edge under the influence of electric field.
Electrons in e5 and e7: diffusion upward, downward, and to the depletion zone under the influence of concentration gradient.
Electrons in e8: diffusion around under the influence of concentration gradient.
The diffusion of electrons in e8 to adjacent pixel cells results in crosstalk.
Figure
Electrons in e1: most recombined in the pinning layer.
Electrons in e2 and e6: diffusion to the depletion zone under the influence of concentration gradient.
Electrons in e3, e4, e7, and e9: accumulated in the depletion zone edge under the influence of electric field.
Electrons in e5: diffusion upward, downward, and to the depletion zone under the influence of concentration gradient.
Electrons in e8 and e10: diffusion around and to the depletion zone under the influence of concentration gradient.
The diffusion of electrons in e8 and e10 to adjacent pixel cells results in crosstalk.
Combining photon absorption rates and drift-diffusion equation (Eq. (
The Sentaurus TCAD tool[21] is used to simulate the electrical and optical characteristics. The input optical signals are introduced by the raytracing method, and the photons absorbed by the photosensitive areas are converted into photo-generated carriers. The number of electrons increasing after illumination in each pixel cell is integrated to obtain the number of photo-carriers accumulated in the pixel cell. The ratio of the average number of photo-carriers of adjacent pixel cells to the number of photo-carriers of the illuminated cells is defined as the crosstalk value. The light intensities at different wavelengths are finely adjusted such that the incident photon numbers with different wavelengths are equal. The physical models include the drift-diffusion model, the SRH combination model, and the high field saturation model. Due to the fact that the study focuses on electrical crosstalk of photo-generated carriers, neither the diffraction nor the crosstalk of light is considered in this work.
As shown in Fig.
At the beginning of exposure, the longer the incident wavelength of light, the more photo-generated carriers are absorbed in the substrate as shown in Figs.
At the time of accumulation as shown in Figs.
When the steady state is reached, the carriers in the illumination cell diffuse mainly to the substrate. In the case of the same number of incident photons, due to the fact that the partial photons are not absorbed by the photosensitive region with larger wavelength, the number of accumulated electric carriers and the electron current density are smaller under steady state as shown in Figs.
When considering the time-dependent crosstalk effect, the trends of crosstalk versus wavelength are different for different exposure times. The human eyes are more sensitive to the green wavelength range, which is not the most serious case of crosstalk at the beginning of exposure and stable state. However, in the exposure process, the crosstalk is the most serious. Evaluation of crosstalk using only a stable condition would not give the actual trend for different exposure times.
As described in Section
When the DTI depth is less than the PD depth as the case in Fig.
When the DTI depth is greater than the PD depth as the case in Fig.
As the image sensor will use a suitable exposure time for better sensitivity and appropriate dynamic range in actual use, the stabilization of photo-carriers accumulation will not be reached for a common incident light intensity. Therefore, the crosstalk evaluation of the pixel cell and the isolation structure is a time-dependent photo-carriers’ diffusion problem. On the other hand, in some special applications, shortening the exposure time can effectively reduce the crosstalk caused by photo-carriers’ diffusion.
In this work, a time-dependent crosstalk effect has been investigated. According to the different depths of PD and isolation structure, the pixel cell is partitioned into several regions. Photo-generated carriers have different motions in different regions. Only the carriers that have spread to adjacent cells can cause the crosstalk. The evaluation of crosstalk is influenced by the exposure time. The crosstalk can be reduced by reducing the exposure time but will lose the sensitivity and dynamic range. The results can help to design and optimize the image sensor especially for scaling down the pixel size.
[1] | |
[2] | |
[3] | |
[4] | |
[5] | |
[6] | |
[7] | |
[8] | |
[9] | |
[10] | |
[11] | |
[12] | |
[13] | |
[14] | |
[15] | |
[16] | |
[17] | |
[18] | |
[19] | |
[20] | |
[21] |