FDTI / BDTI Pixel Simulator

Build a virtual deep trench isolation CMOS image sensor pixel and inspect how FDTI/BDTI choice, trench width, BDTI depth, liner material, wavelength, and chief ray angle affect optical confinement.

FDTI / BDTI CMOS Pixel Optical Simulator

Build a virtual deep trench isolation pixel and inspect the approximate optical field, collected signal, and crosstalk trend.

Preset

Pixel Geometry

Pixel pitch 1.00 um

Silicon thickness 3.2 um

Microlens height 0.52 um

Isolation

Trench type

Trench depth full silicon 3.20 um

FDTI is drawn as a full-depth trench through silicon.

Trench width 0.12 um

Fill / liner

Illumination

Wavelength 550 nm

Chief ray angle 7°

Lens offset 0.04 um

Virtual DTI structure

Approximate optical field

Central signal

82.4%

Neighbor crosstalk

0.1%

DTI suppression

7.1 dB

Optical loss

17.5%

Visual QA scenarios

Current designPASS

No DTI controlPASS

Shallow BDTIPASS

Full FDTI + obliquePASS

Model note: this browser view is a deterministic paraxial/ray approximation for design review, not a replacement for RCWA or FDTD sign-off.

Model scope

Use this browser tool for intuition, relative trends, and design-space exploration. Its local simplified model is not a substitute for RCWA/FDTD sign-off, silicon calibration, or vendor process data.

Learn more

Pixel Optical Effects · DTI Crosstalk recipe

Physics Notes

DTI Optical Confinement and Crosstalk

Plain-English Intuition

Picture each pixel as a small room. Without walls, light leaking sideways from one room shows up as colour bleeding in the neighbour — that is optical crosstalk. DTI (Deep Trench Isolation) builds tiny vertical walls between pixels, reflecting or blocking that sideways light so each pixel mostly sees its own photons.

Deep trench isolation reduces optical crosstalk by redirecting or blocking lateral light leakage between neighboring pixels.

Assumptions

DTI is represented by compact geometric confinement and leakage proxies, not a full electromagnetic trench solve.
Trench depth, fill material, and geometry are treated as idealized parameters without etch profile roughness.
Optical crosstalk indicators are qualitative unless calibrated against RCWA/FDTD or silicon measurements.

Outputs

FDTI/BDTI geometry comparison, confinement proxy, crosstalk trend, absorption path-length proxy, and field sketch.
A design-space view of how trench depth and refractive-index contrast trade sensitivity against isolation.

Validation Example

Increasing trench depth or lowering trench index should generally improve confinement and reduce crosstalk indicators.
If isolation improves without any optical cost in every case, treat the result as a qualitative proxy and verify with full-wave simulation.

Core Equations

Critical angle

$$\theta_c = \arcsin\left( \frac{n_{\text{trench}}}{n_{\text{Si}}} \right)$$

$\theta_c$: Angle for Total Internal Reflection (TIR)
$n_{\text{trench}}$: Refractive index of trench filling

A low-index trench can confine silicon-guided rays by total internal reflection.

Optical crosstalk proxy

$$XT = \frac{P_{\text{neighbor}}}{P_{\text{center}} + P_{\text{neighbor}}}$$

$P$: Optical power in pixel

The simulator reports a structural optical trend, not carrier diffusion crosstalk.

Absorption path benefit

$$A = 1 - e^{-\alpha L_{\text{eff}}}$$

$L_{\text{eff}}$: Effective absorption path length

DTI and backside structures can increase effective path length in silicon.

Model Interpretation

FDTI and BDTI differ in how much of the silicon depth is isolated.
Metal-filled trenches improve shielding but introduce absorption and process complexity.
Electrical isolation and optical isolation are related but not identical metrics.

Optical Confinement Logic

DTI works by interrupting lateral propagation paths before photons reach a neighboring photodiode.
Total internal reflection is possible when rays in silicon meet a lower-index trench beyond $\theta_c=\arcsin(n_{\text{trench}}/n_{\text{Si}})$.
Absorbing or metal-filled trenches can block leakage but may also introduce parasitic absorption and process complexity.

Design Checks

Compare center-pixel absorption and neighbor-pixel absorption; crosstalk is a ratio, not an absolute photon count alone.
Sweep trench depth and width because shallow trenches may miss long red/NIR absorption paths.
Check color dependence: blue may be surface-limited while red/NIR is more sensitive to deep lateral paths.

Known Missing Physics

Simple DTI metrics do not include waveguide modes, corner diffraction, roughness, or polarization-dependent trench response.
Electrical isolation and optical isolation are related but not identical; carrier diffusion can still cause crosstalk.
Final validation should use 3D EM fields and charge-collection modeling for advanced pixels.