FilmOptima logoFilmOptima®

Features

Dispersion Formulas

FilmOptima provides a range of dispersion models for both the real and imaginary parts of the refractive index. If none of the analytical models are suitable, tabulated data can be used instead, with intermediate values interpolated using cubic splines.

Real
  • Constant: n=A0n=A_0
  • Cauchy: n(λ0)=A0+∑i=1nBiλ0Cin(\lambda_0)=A_0+\sum_{i=1}^nB_i\lambda_0^{Ci}
  • Conrady: n(λ0)=A0+A1λ0+A2λ03.5n(\lambda_0)=A_0+\frac{A_1}{\lambda_0}+\frac{A_2}{\lambda_0^{3.5}}
  • Exotic: n2(λ0)=A0+A1λ02−A2+A3(λ0−A4)(λ0−A4)2+A5n^2(\lambda_0)=A_0+\frac{A_1}{\lambda_0^2-A_2}+\frac{A_3(\lambda_0-A_4)}{(\lambda_0-A_4)^2+A_5}
  • Gases: n(λ0)−1=A0+∑i=0nBiCi−λ0−2n(\lambda_0)-1=A_0+\sum_{i=0}^n\frac{B_i}{C_i-\lambda_0^{-2}}
  • Lorentz-Drude: Re⁡(ϵr(ω))=ϵr(∞)+ωp2∑i=1Mfiω0,i2−ω2+ȷωΓi\operatorname{Re}(\epsilon_r(\omega))=\epsilon_r(\infty)+\omega_p^2\sum_{i=1}^M\frac{f_i}{\omega_{0,i}^2-\omega^2+\jmath\omega\Gamma_i}
  • Hartmann: n(λ0)=A0+A1λ0−A2n(\lambda_0)=A_0+\frac{A_1}{\lambda_0-A_2}
  • Hartmann-Modified: n(λ0)=A0+A1(λ0−A2)2n(\lambda_0)=A_0+\frac{A_1}{(\lambda_0-A_2)^2}
  • Herzberger: n(λ0)=A0+A1λ02−0.028+A2(1λ02−0.028)2+A3λ02+A4λ04+A5λ06n(\lambda_0)=A_0+\frac{A_1}{\lambda_0^2-0.028}+A_2\left(\frac{1}{\lambda_0^2-0.028}\right)^2+A_3\lambda_0^2+A_4\lambda_0^4+A_5\lambda_0^6
  • Schoitt-Briot: n2(λ0)=A0+A1λ02+∑i=1nBiλ02in^2(\lambda_0)=A_0+A_1\lambda_0^2+\sum_{i=1}^n\frac{B_i}{\lambda_0^{2i}}
  • Sellmeier: n2(λ0)=A0+∑i=0nBiλ02λ02−Cin^2(\lambda_0)=A_0+\sum_{i=0}^n\frac{B_i\lambda_0^2}{\lambda_0^2-C_i}
  • Sellmeier-Modified: n2(λ0)=A0+∑i=0nBiλ02λ02−Ci2n^2(\lambda_0)=A_0+\sum_{i=0}^n\frac{B_i\lambda_0^2}{\lambda_0^2-C_i^2}
  • Polynomial: n2(λ0)=A0+∑i=1nBiλ0Cin^2(\lambda_0)=A_0+\sum_{i=1}^nB_i\lambda_0^{C_i}
  • RefractiveIndexInfo: n2(λ0)=A0+B1λ0C1λ02−B2C2+B3λ0C3λ02−B4C4+B5λ0C5+B6λ0C6+B7λ0C7+B8λ0C8n^2(\lambda_0)=A_0+\frac{B_1\lambda_0^{C_1}}{\lambda_0^2-B_2^{C_2}}+\frac{B_3\lambda_0^{C_3}}{\lambda_0^2-B_4^{C_4}}+B_5\lambda_0^{C_5}+B_6\lambda_0^{C_6}+B_7\lambda_0^{C_7}+B_8\lambda_0^{C_8}
  • Datatable: Datatable\text{Datatable}
Imaginary
  • Non-Absorptive: k=0k=0
  • Constant: k=A0k=A_0
  • Cauchy: k(λ0)=A0eA1(1λ0−1A2)k(\lambda_0)=A_0e^{A_1(\frac{1}{\lambda_0}-\frac{1}{A_2})}
  • Exponential: k(λ0)=A0eA1λ0+A2λ0k(\lambda_0)=A_0e^{\frac{A_1}{\lambda_0}+A_2\lambda_0}
  • Lorentz-Drude: Im⁡(ϵr(ω))=ϵr(∞)+ωp2∑i=1Mfiω0,i2−ω2+ȷωΓi\operatorname{Im}(\epsilon_r(\omega))=\epsilon_r(\infty)+\omega_p^2\sum_{i=1}^M\frac{f_i}{\omega_{0,i}^2-\omega^2+\jmath\omega\Gamma_i}
  • Polynomial: k(λ0)=A0+∑i=1nBiλ0Cik(\lambda_0)=A_0+\sum_{i=1}^nB_i\lambda_0^{Ci}
  • Sellmeier: k(λ0)=A0A1A2+A3λ0+1λ03k(\lambda_0)=\frac{A_0}{A_1A_2+\frac{A_3}{\lambda_0}+\frac{1}{\lambda_0^3}}
  • Datatable: Datatable\text{Datatable}

Optical Characteristics

Conventional, phase, and layer-specific outputs available during optimization and analysis.

Conventional
  • R - Reflectance (s-pol, p-pol, avg)
  • T - Transmittance (s-pol, p-pol, avg)
  • A - Absorptance (s-pol, p-pol, avg)
Phase
  • pR - Phase Shift on Reflection (s-pol, p-pol)
  • pT - Phase Shift on Transmission (s-pol, p-pol)
  • GDR - Group Delay on Reflection (s-pol, p-pol)
  • GDT - Group Delay on Transmission (s-pol, p-pol)
  • GDDR - Group Delay Dispersion on Reflection (s-pol, p-pol)
  • GDDT - Group Delay Dispersion on Transmission (s-pol, p-pol)
Layer-Specific
  • LA - Layer Absorptance (s-pol, p-pol, avg)
  • LOA - Local Absorption (s-pol, p-pol, avg)
  • E - Electric Field (s-pol, p-pol, avg)

Merit Formulas

Objective formulations for balancing target deviation and optimization priorities.

Available Formulas
  • Lp-Norm: F=[1m∑i=1wi∣Si−Si^∣ΔTp]1/pF=\left[ \frac{1}{m}\sum_{i=1}w_i\frac{\vert S_i-\hat{S_i}\vert}{\Delta_T}^p \right]^{1/p}
  • Max: F=max⁡[wi∣Si−Si^∣ΔT:1..m]F=\max\left[w_i\frac{\vert S_i-\hat{S_i}\vert}{\Delta_T}: 1..m\right]

Coherence Models

Select the propagation model that matches your stack's physical coherence behavior.

Coherent Stack

All optical characteristics.

Incoherent Stack

R/T/A.

Hybrid (Mixed) Stack

R/T/A.

Algorithms

Explore FilmOptima's algorithm portfolio by category and jump to detailed documentation for each method.

Refinement
AdamW

Adaptive local refinement

Iteratively adjusts existing layer thicknesses using momentum and adaptive learning rates for stable, efficient convergence.

ClusterRefinement
Proprietary

Cluster-based candidate refinement

Enumerates clustered layer-thickness candidates and polishes each with AdamW to escape shallow local minima.

Synthesis
Needle

Zero-thickness synthesis

Builds designs by inserting zero-thickness layers at high-impact positions and refining the full stack after each insertion.

AnyNeedle
Proprietary

Finite-thickness synthesis

Extends Needle method by evaluating insertion depth and finite thickness candidates, then accepting the best refined stack.

Global Optimization
BranchAndBound

Bounded global search

Systematically partitions thickness space and prunes subregions using merit bounds to target globally superior solutions.

ParticleSwarmOptimization

Swarm-based exploration

Uses a population of particles that balance personal and social learning signals to discover high-quality regions.

Stochastic

Randomized global sampling

Performs repeated random initializations within bounds and refines each candidate to improve near-global search coverage.

How It Works

Datamanagement

FilmOptima employs a relational SQL database to ensure data integrity and efficient storage. This structure normalizes data, meaning common entities like general targets or materials are stored only once and referenced across multiple user designs. To preserve data consistency, the system prevents modifications to any entity that is actively referenced in other designs, preventing unintended cascading effects.

Flexible Optimization Control

FilmOptima puts you in complete control of the optimization process. Define your goals (e.g., max layers, max optimization duration) as stopping conditions, and hand-pick the algorithms to use. Our engine runs your chosen algorithms in round-robin manner, and intelligently concludes the process when no further improvements can be made.

Hands-Free Queue

FilmOptima allows you to queue multiple designs for sequential processing, perfect for running long optimizations overnight or while you focus on other tasks. You can also compare different optimization strategies on the same design by adding them to the queue.