40nm BSI CMOS Image SensorPreview

imary charge collection volume of the pinned photodiode [A1]. Unlike the preceding 'Photocathode IIPX - Photo' step, which exclusively utilizes optical lithography to define the spatial boundaries of the implant window in the photoresist, this step performs the actual high-energy ballistic insertion of dopant ions into the exposed silicon [T1]. Positioned after the foundational N-Well and P-Well definitions, this implant creates the localized n-type core within the p-type epitaxial layer [P1]. This establishes the fundamental p-n junction required for photoelectric conversion and charge storage [T2]. Following this step, the photoresist is stripped, preparing the wafer for subsequent surface pinning layer implants that will fully define the pinned photodiode structure [A1]. The core mechanism of this step relies on ion implantation, where accelerated dopant ions penetrate the silicon lattice to form a highly controlled concentration profile [T1]. The spatial distribution of the implanted N-type dopants generally follows a Gaussian function, determined by the projected range and range straggle characteristic of the specific ion energy [T1]. By carefully selecting the implant energy and dose, engineers construct a specific doping gradient that creates a built-in potential within the photodiode [A1]. This built-in potential establishes a depletion region that sweeps photo-generated electrons into the potential well while holes are repelled into the p-type substrate [T2]. Furthermore, tuning the dopant concentration at the eventual p+/n junction is critical; keeping the n-type concentration relatively low reduces the local electric field [P1]. A weakened local electric field fundamentally suppresses generation-recombination mechanisms, thereby significantly reducing dark current degradation [P1]. Phosphorus is typically selected as the n-type dopant for the deep photodiode well because its lighter atomic mass allows for deeper penetration ranges compared to arsenic at equivalent implantation energies [P3]. The precise control of the implant dose directly dictates the total volume of the potential well, which fundamentally bounds the pixel's full well capacity (FWC) [P1]. However, increasing the dose to maximize FWC presents a trade-off, as an excessively high n-type concentration makes it difficult to achieve full depletion at standard operating voltages [T2]. Full depletion is essential to eliminate image lag and ensure complete charge transfer to the floating diffusion node [A2]. Additionally, the implant energy must be perfectly matched to the photoresist thickness applied in the preceding photo step; insufficient resist stopping power allows stray ions to penetrate adjacent isolated regions, compromising the spatial separation of pixels [T1]. In a nanoscale Backside Illuminated (BSI) CMOS image sensor, the physical volume available for the photodiode is severely constrained by the highly scaled pixel pitch [P1]. Consequently, the N Photocathode IIPX process must utilize multi-energy chained implants to construct a deep, vertically graded electrostatic potential well [P3]. This tailored vertical profile ensures that even deep-penetrating photons generate carriers that are efficiently drifted upward toward the transfer gate [T2]. Furthermore, the tight thermal budget of the 40nm node restricts subsequent diffusion-based profile broadening, meaning the as-implanted Gaussian distribution must closely approximate the final target profile to maintain stringent control over junction capacitances and noise characteristics [T1].