QQ: 859818709

Thin-Film Interference Calculator

Thin-Film Interference Calculator

Thin-Film Interference Calculator (Thin-Film TMM)
Uses the transfer matrix method (TMM) to calculate the reflectance R, transmittance T and absorption loss A of any multilayer optical thin film at a given wavelength and angle of incidence. Supports s / p polarisation and unpolarised light, and complex refractive indices (including absorbing layers and metals); suitable for the design and teaching of anti-reflection (AR) coatings, high-reflection (HR / DBR) coatings and bandpass filters.
M = Πⱼ [[ cos δⱼ , i·sin δⱼ / Yⱼ ],[ i·Yⱼ·sin δⱼ , cos δⱼ ]], δⱼ = 2π·nⱼ·dⱼ·cos θⱼ / λ

Thin-film stack calculation

Fill in the material of each layer (or refractive index n, extinction coefficient k and thickness d), set the incident medium, substrate, angle of incidence and polarisation, and you will obtain R(λ)/T(λ) spectral curves and values at specified wavelengths.

Layer stack (from the incident side to the substrate, added bottom-up)
Material n k Thickness d (nm)

Incident medium / substrate

Incidence conditions

°



Wavelength scan range

nm

nm

Common presets

Result readings (at the centre wavelength)
Reflectance R(λ)

Transmittance T(λ)

Spectral data (sampled)

Core formulae and principles

1
Single-layer characteristic matrix (transfer matrix method)
Mⱼ = [[ cos δⱼ , i·sin δⱼ / Yⱼ ],[ i·Yⱼ·sin δⱼ , cos δⱼ ]]
δⱼ = 2π·nⱼ·dⱼ·cos θⱼ / λ is the phase thickness of layer j; Yⱼ is the optical admittance of that layer for the given polarisation. The total matrix of the whole stack is the product of the individual layer matrices, M = M₁·M₂·…·M_N. Systematically established by Macleod and by Born & Wolf, this method is the classic algorithm for thin-film design.
2
Polarisation-dependent admittance Y
s(TE):Yⱼ = nⱼ·cos θⱼ ;p(TM):Yⱼ = nⱼ / cos θⱼ
For oblique incidence the admittances of s polarisation (electric field perpendicular to the plane of incidence) and p polarisation (electric field parallel to the plane of incidence) differ, and therefore so do the reflectances. At normal incidence cos θⱼ = 1 and both reduce to Yⱼ = nⱼ, so s and p give identical results.
3
Reflection / transmission coefficients
r = (Y₀ − Y_in)/(Y₀ + Y_in), t = 2Y₀/(Y₀·B + C)
Y_in = (m₁₁·Yₛ + m₁₂)/(m₂₁·Yₛ + m₂₂) is the equivalent input admittance, with Y₀ and Yₛ the admittances of the incident medium and substrate. The power reflectance is R = |r|² and the transmittance T = |t|²·Re(Yₛ)/Re(Y₀). A complex refractive index n − i·k (k ≥ 0 is the extinction coefficient) allows absorbing layers and metals to be treated alike.
4
Quarter-wave (QW) and half-wave (HW) layers
QW: d = λ₀/(4n), effective index n_eff = n₁²/n₂
At the centre wavelength λ₀ a quarter-wave layer ‘transforms’ the substrate index to n₁²/n₂: choosing n₁ = √(n₀·n₂) gives zero reflectance for an air–substrate interface (the ideal single-layer AR coating). A half-wave layer (d = λ₀/2n) is ‘transparent’ at all wavelengths, being equivalent to no layer at all (doubling its thickness does not change the reflectance).
5
Anti-reflection (AR) and high-reflection (HR / DBR) coatings
AR: n₁ ≈ √(n₀·nₛ); HR: alternating λ/4 high/low-index layers
A single-layer AR coating reduces R from ~4% (BK7–air interface) to nearly zero; a multilayer stack of alternating λ/4 high- and low-index layers (a distributed Bragg reflector, DBR) can achieve nearly 100% reflectance near the centre wavelength (more layers give a wider reflection band and a higher peak). This is the core structure of laser cavity mirrors and filters.
6
Energy conservation and absorption loss
R + T + A = 1 (A > 0 when absorbing layers / metals are present)
When all media are lossless (k = 0), R + T = 1; as soon as a layer or the substrate has an extinction coefficient (k > 0, as in metal films or semiconductors), the difference is absorbed and is recorded as the absorption loss A = 1 − R − T. This tool outputs A faithfully, making it easy to assess heating and loss in metal mirrors.
Snell’s law (complex form):n₀·sin θ₀ = nⱼ·sin θⱼ. With complex refractive indices θⱼ is also complex; this tool takes the principal branch of cos θⱼ (non-negative real part), ensuring a decaying wave in absorbing layers and an evanescent wave under total reflection, which is physically self-consistent.
By default the calculation takes a single refractive index per layer (ignoring material dispersion), which suits teaching and preliminary design; for accurate spectra you should instead use a wavelength-dependent n(λ) (e.g. the Sellmeier equation) point by point.

Refractive index reference for common coating materials (approximate, visible ~550 nm)

Material n (550 nm) k Type Typical applications
Air 1.000 0 Medium Incident / ambient medium
Fused silica SiO₂ 1.458 0 Low-index medium Low-index layers, protective coatings
Magnesium fluoride MgF₂ 1.380 0 Low-index medium Single-layer anti-reflection coating
Aluminium oxide Al₂O₃ 1.770 0 Medium-index medium protective / high-reflection stacks
Titanium dioxide TiO₂ 2.400 0 High-index medium High-reflection coatings, AR outer layers
Tantalum pentoxide Ta₂O₅ 2.100 0 High-index medium High-reflection coatings, low loss
Hafnium oxide HfO₂ 2.000 0 High-index medium High-power laser coatings
Silicon nitride Si₃N₄ 2.000 0 High-index medium Semiconductor coatings
Zinc sulfide ZnS 2.350 0 High-index medium IR / visible coatings
BK7 glass 1.517 0 Substrate General optical substrate
Sapphire (o-ray) 1.770 0 Substrate Wear-resistant windows
Silicon Si 3.500 0 Semiconductor (near-IR) IR substrates / detectors
Aluminium Al (metal) 1.440 5.230 Metal Mirrors (broadband high reflection)
Silver Ag (metal) 0.050 3.200 Metal Visible high-reflection mirrors
Gold Au (metal) 0.380 2.400 Metal IR high-reflection mirrors
Data notes:The table above gives approximate refractive indices of common coating materials at room temperature (around visible ~550 nm). Metals and semiconductors exhibit strong dispersion in the visible / infrared (k varies sharply with wavelength), so these figures are only order-of-magnitude references. For engineering design and simulation, refer to the measured n(λ) and k(λ) curves in manufacturers’ coating handbooks (such as refractiveindex.info). This tool takes a single refractive index per layer by default and ignores dispersion, making it suitable for principle demonstration and preliminary stack design.

Frequently asked questions (FAQ)

What is the thin-film transfer matrix method (TMM)?

The transfer matrix method is the classic algorithm for calculating the optical properties of plane-parallel multilayers: a 2×2 characteristic matrix is written for each layer, the layer matrices are multiplied to obtain the total matrix of the stack, and the reflection coefficient r and transmission coefficient t are then found using the optical admittances of the incident medium and substrate, giving the reflectance R = |r|² and transmittance T. It can handle any number of layers, any angle of incidence, s/p polarisation and complex refractive indices with absorption.

How does a single-layer anti-reflection coating reduce reflection?

Inserting a thin film of refractive index n₁ between air (n₀ = 1) and glass (nₛ ≈ 1.5), when the film has a quarter-wave optical thickness (d = λ₀/4n₁), transforms the substrate index equivalently to n₁²/nₛ. Choosing n₁ = √(n₀·nₛ) ≈ 1.23 (MgF₂ is actually 1.38, which is close) matches the equivalent index on the air side to the substrate, giving nearly zero reflection at λ₀. Deviating from this thickness or index reduces the anti-reflection effect.

What is the difference between a quarter-wave and a half-wave layer?

A quarter-wave layer (QW, d = λ₀/4n) changes the effective refractive index at the centre wavelength and is the basic building block of AR and HR coatings; a half-wave layer (HW, d = λ₀/2n) is ‘transparent’ at all wavelengths — it has no effect on reflectance (equivalent to doubling a layer’s thickness) and is often used to ‘move’ a layer to another position without changing the overall optical performance, as well as to compensate the phase of a stack.

How is a high-reflection coating (mirror) realised?

By depositing alternating λ/4-thick layers of high-index (e.g. TiO₂, n ≈ 2.4) and low-index (e.g. SiO₂, n ≈ 1.46) materials into a multilayer stack (a distributed Bragg reflector, DBR), each pair of interfaces produces in-phase reflection at the centre wavelength and the contributions add approximately constructively; more layers give higher reflectance and a wider reflection band. Four pairs of TiO₂/SiO₂ already exceed 95% reflectance at 550 nm.

Why does R + T not equal 1 in my results?

R + T = 1 holds only when all media (layers and substrate) are lossless (extinction coefficient k = 0). As soon as absorption is present (k > 0, as in aluminium metal mirrors, semiconductors or deliberately doped absorbing layers), the difference is the energy absorbed by the stack, recorded as A = 1 − R − T. A metal mirror reflecting about 90% with the rest absorbed is precisely a case of A > 0. This tool reports the absorption loss A faithfully.

Why do s and p polarisations have different reflectances?

At oblique incidence the optical admittances of s (TE, electric field perpendicular to the plane of incidence) and p (TM, electric field parallel to the plane of incidence) differ: Y_s = n·cosθ and Y_p = n/cosθ, so their reflectances differ. At normal incidence cosθ = 1 and both reduce to Y = n, giving identical results for s and p. When designing oblique-incidence filters or polarisation components, the two polarisations must be considered separately.

Does the wavelength dependence of refractive index (dispersion) affect the results?

Yes. This tool takes a single refractive index per layer by default (ignoring dispersion), so the positions of interference fringes are approximate — adequate for qualitative demonstration, teaching and preliminary stack design, but accurate spectra (especially broadband) require using a wavelength-dependent n(λ) (e.g. the Sellmeier equation) at each wavelength. The stronger the material’s dispersion (as with TiO₂ or semiconductors), the more pronounced the deviation over a broad band.

How do I enter metal and dielectric films in this tool?

Metals (such as aluminium, silver and gold) have a complex refractive index n − i·k with a large k; entering that complex index directly in the incident medium / substrate fields nₛ, kₛ (you may leave the layers empty and only change the substrate) gives the metal reflectance. Alternatively, treat the metal as a layer (fill in n, k and thickness) to calculate the reflection and absorption of a metallic thin film. Use the table above as a guide and consult a coating handbook for precise values.

Can this calculator design bandpass / cutoff filters?

It can demonstrate the principle: under a wavelength sweep, observe the R(λ)/T(λ) curves — a DBR high-reflection stack shows a high-reflection band near the centre wavelength (reflection) and high transmission outside it; inserting a λ/2 spacer between two high-reflection stacks forms a Fabry–Pérot (FP) cavity, yielding narrowband transmission peaks (a bandpass filter). This tool provides R/T spectra and a data table for analysing the passband position and bandwidth of such structures.

What is the relationship between TMM and the Fresnel reflection formula?

TMM is the generalisation of the Fresnel formula to multilayers: when the number of layers is zero, TMM reduces to Fresnel reflection at a single interface (at normal incidence R = ((n₀−nₛ)/(n₀+nₛ))²), so with no film this tool agrees with the Fresnel result.

What unit is used for the thickness d?

The thickness in the input boxes is in nanometres (nm). At 550 nm a quarter-wave layer is about 99.6 nm for MgF₂ (n = 1.38) and about 57.3 nm for TiO₂ (n = 2.4). Note that d is the physical thickness, not the optical thickness.

What is the maximum angle of incidence that can be calculated?

In principle it can be calculated up to the critical angle; when the angle of incidence exceeds the critical angle for total internal reflection within a layer, an evanescent wave appears in that layer. This tool handles this self-consistently using complex Snell’s law (the principal branch of cosθ with a positive imaginary part), still giving the correct complex reflectance, although T then tends to zero.

How accurate is the calculation?

The tool’s core is pure complex matrix arithmetic and agrees with an independent Python reference implementation to within a relative error of 1e-9 (cross-validated on more than ten test cases including no film, single / multilayers, oblique incidence, absorbing layers and metals). Physical approximation comes from the fixed refractive index (ignoring dispersion), not from algorithmic error.

Can it be used to find the optimal layer count and materials for an anti-reflection (AR) coating?

You can manually try different layer counts and material combinations and observe the R spectrum; a common two-layer AR coating (e.g. MgF₂ + Al₂O₃ or a low/high-index λ/4–λ/4 pair) can suppress broadband reflection below 0.5%. Automatic optimisation (stack inversion) is outside this tool’s scope, but you can iterate manually using the results.

What is the use of the absorption loss A in the results?

A = 1 − R − T reflects the fraction of power absorbed by the stack. For laser high-reflection mirrors and low-loss filters, a smaller A is better; for metal mirrors A directly determines the heat load. When assessing high-power laser coatings, A is a more critical figure than R alone.

Can it handle non-parallel films or curved substrates?

This tool is designed for plane-parallel stacks (plane waves, uniform layers). Curved substrates, graded refractive indices, wedge errors and the like require more sophisticated vector or rigorous coupled-wave analysis (RCWA) and are outside the scope of the TMM monochromatic plane-wave model.

Why is BK7 (n ≈ 1.517) recommended as the default substrate?

BK7 is the most common visible-light optical glass, with about 4.2% reflectance at its air interface. Using it as the substrate makes it easiest to see visually how a single MgF₂ AR layer reduces reflection to below ~1%, which helps verify that the tool is working correctly.

Is the tool free? Do I need to register?

Completely free and with no registration required; it can be used directly in mobile and desktop browsers. It is provided by Soar Laser for coating design and teaching by laser and optical engineers.

Can the calculation results be used directly for production coating processes?

This tool is suitable for principle verification, teaching and preliminary design. Real coating work must consider dispersion n(λ), layer stress and adhesion, monitoring methods (optical / quartz crystal), refractive-index drift with process conditions and so on; professional coating design software and coating handbooks should be authoritative.

Why do thin films show coloured interference fringes?

When the film thickness is comparable to the wavelength, different wavelengths reflect with different phases, so some wavelengths interfere destructively (dark) and others constructively (bright); under white light this produces rainbow colours that vary with thickness and angle (as in camera lenses or the colours of an oil film). The R(λ)/T(λ) curves in this tool are precisely this interference spectrum.

How should I set things when the incident medium is not air?

Simply change the incident medium’s n₀ and k₀ to the actual refractive index of the medium, e.g. n₀ = 1.33 in water or about 1.5 in oil. This changes the baseline Fresnel reflection at the interface and the quarter-wave matching condition; the tool recalculates using the new n₀.

How does it compare with optical thin-film design software such as Essential Macleod?

Professional software offers dispersion libraries, automatic stack optimisation, tolerance analysis and monitoring and process simulation. This tool is a lightweight, ready-to-use forward TMM engine emphasising explainability and intuitive teaching; it does not include automatic inversion or process modules.


中文EN