1. Structural Difference as a Design Premise
In FSI sensors, incident light must pass through multiple metal wiring layers before reaching the photodiode. This architecture, while historically sufficient, inherently introduces optical obstruction, especially as pixel sizes shrink.
BSI sensors invert the silicon substrate, relocating metal interconnects to the rear of the photodiode. As a result, photons are allowed to reach the light-sensitive region with fewer losses.
This structural reorientation establishes the physical basis for the performance divergence observed between the two architectures.
2. Photon Efficiency and Low-Light Performance
Because FSI pixels are partially shadowed by front-side circuitry, their quantum efficiency degrades noticeably under low-illumination conditions. As pixel pitch decreases, this limitation becomes increasingly pronounced, necessitating stronger ISP-level noise suppression.
BSI architecture, by contrast, improves photon collection efficiency by design. Consequently, higher signal-to-noise ratios can be maintained at lower illumination levels, and image degradation is delayed rather than abruptly compensated through algorithmic intervention.
Thus, the advantage of BSI is not limited to "brighter images," but extends to more predictable signal behavior under non-ideal lighting.
3. Impact on Pixel Scaling and Optical Design
As imaging systems trend toward higher resolution and smaller module footprints, pixel size reduction becomes unavoidable. In such contexts, FSI architecture encounters a structural bottleneck: reduced effective photosensitive area and increased optical crosstalk.
BSI alleviates this constraint. By decoupling metal routing from the light path, it enables further pixel miniaturization without proportionally sacrificing sensitivity.
This characteristic directly expands optical design flexibility, allowing smaller apertures, wider fields of view, or thinner module stacks without catastrophic image quality loss.
From a camera module perspective, this translates into greater freedom in mechanical and optical co-design.
4. Manufacturing Complexity and Cost Considerations
It should be noted that BSI sensors introduce additional manufacturing steps, including wafer thinning and backside processing, which increase fabrication complexity and cost.
FSI sensors, benefiting from mature processes and higher yields, remain economically attractive for applications where lighting conditions are controlled and extreme miniaturization is not required.
Therefore, the persistence of FSI in certain market segments should not be interpreted as technological stagnation, but rather as context-appropriate engineering optimization.
5. System-Level Implications for Camera Modules
When evaluated at the camera module level:
- FSI sensors tend to perform adequately in
cost-sensitive, illumination-stable environments, where system simplicity and supply chain maturity outweigh the need for low-light robustness.
- BSI sensors demonstrate clear advantages in
compact modules, wide-angle optics, low-light scenarios, and applications requiring consistent image quality across variable conditions.
The distinction, therefore, is not absolute superiority, but architectural alignment with system intent.
Conclusion
The transition from FSI to BSI reflects a broader evolution in imaging system philosophy-from compensating for physical limitations through algorithms, to mitigating those limitations at the structural level.
For camera module designers and system integrators, understanding this architectural difference is essential. Sensor selection should not be guided solely by resolution or generation, but by how illumination architecture interacts with optics, mechanics, ISP resources, and application environments.
In this sense, BSI and FSI are not competing labels, but two responses to different historical and engineering constraints.
