A dichroic glass coating is a precision thin-film interference filter applied to a specified substrate that reflects designated wavelengths and transmits others. We specify it correctly by defining spectral edges at the required angle of incidence, the substrate material and the environmental test class. These coatings sit at the heart of beam splitters, IR dichroics, sensor windows and multi-band imaging assemblies across aerospace, defence, medical and automotive programmes.
TL;DR:
- The complexity of multilayer stacks, especially for wide spectral or multi-band applications, significantly raises manufacturing costs and lead times.
- Fragile substrates like germanium require careful handling and specific process controls to prevent stress-induced damage or distortion.
- In-process optical monitoring with error self-compensation improves accuracy and consistency, enabling reliable production of narrow-band and multi-band dichroic filters.
- Standard durability tests serve as a baseline, but extended environmental and thermal cycling are critical for long-term performance in harsh conditions.
- Custom coatings are necessary when strict spectral, environmental, or mechanical specifications cannot be met with off-the-shelf filters, with costs influenced mainly by layer count, substrate fragility, and tolerance requirements.
Table of Contents
- How dichroic coatings work: optical principles, stack strategies and materials
- Substrate selection and the mechanical side of coating performance
- Manufacturing controls: how precise dichroic stacks get made reliably
- Writing a procurement-ready specification and acceptance test plan
- Where dichroic coating technology came from
- Long-term durability beyond the standard test sequence
- Dichroic coatings versus other optical coating approaches
- Cost drivers, pricing ranges and lead times to plan around
- Where dichroic coating design is heading next
- Choosing between a custom coating and an off-the-shelf filter
- Starting an RFQ for a dichroic coating with Precision Glasses
- FAQ
- Sources
How dichroic coatings work: optical principles, stack strategies and materials
A dichroic coating achieves its wavelength selectivity through interference, not absorption or pigment. Alternating layers of high and low refractive index materials are deposited at precise thicknesses, typically a quarter-wavelength of the target light, so that reflected waves from each interface reinforce at some wavelengths and cancel at others. The result is a sharp edge between a reflective band and a transmissive one, or a narrow passband surrounded by rejection on either side.
Two design philosophies dominate industrial practice. A classic quarter-wave stack gives predictable edges with a modest layer count, but it struggles with wide angular bandwidth or multiple separated bands. Needle optimisation, by contrast, lets the design algorithm insert or remove thin layers wherever the merit function improves, producing graded-index-like profiles that meet tighter spectral windows without an arbitrary cap on layer count. Suppliers favour optimisation-based design over fixed stacks because it trades a slightly more complex deposition run for far better manufacturability and yield.
Material choice follows the spectral band. Oxide layers suit visible and near-infrared work. Moving into the shortwave and mid/long-wave infrared, materials such as germanium (Ge), zinc sulphide (ZnS) and ytterbium fluoride (YbF3) become standard building blocks, each chosen for its refractive index, transparency window and mechanical compatibility with the substrate.
- High/low index pairing creates the reflective and transmissive bands that define a dichroic edge.
- Ge, ZnS and YbF3 are common choices for SWIR, MWIR and LWIR dichroic stacks.
- Needle optimisation allows designers to hit tight spectral targets without forcing an artificial layer-count limit.
Every material brings trade-offs in absorption, stress and adhesion, which is why we treat material selection as inseparable from mechanical qualification rather than a purely optical decision.
Substrate selection and the mechanical side of coating performance
The substrate is not a passive carrier. Its native refractive index and surface reflectance shape the starting point for the coating design, and its mechanical properties determine what the deposition process can safely do to it. Germanium offers excellent infrared transmission but is brittle and prone to fracture under coating stress. Sapphire is mechanically tough but its high index and birefringence add design constraints of their own. Thermal expansion mismatch between coating and substrate becomes a real risk once a stack runs to a dozen or more layers.
Coating stress can bend a flat window into a slight curve, change its surface figure and even feed distortion into mounted assemblies. That is why we request surface-profile data before and after deposition and why a rear-surface anti-reflection coating is often specified alongside the dichroic layer to balance transmission losses.
- Specify substrate material, thickness and orientation alongside the spectral target, not as an afterthought.
- Request allowable surface deformation in RMS terms, measured before and after coating.
- Ask for a note on material compatibility, particularly where Ge, sapphire or other fragile substrates are involved.
Pro Tip: Ask for pre- and post-coating interferometry on a witness piece from the same lot before you sign off on a fragile substrate run.
Readers wanting more background on substrate trade-offs can see our glass substrate explained guide.
Manufacturing controls: how precise dichroic stacks get made reliably
Deposition method sets the ceiling on achievable precision. Ion-assisted deposition, thermal evaporation and sputtering each offer different combinations of layer density, stress and repeatability, and the right choice depends on the material pairing and the spectral window being targeted. As the layer count climbs into the teens, the risk of cumulative thickness error grows, which is why real-time control matters more than raw deposition technique.
Modern coating runs rely on in-process optical monitoring, where the deposition chamber measures the growing stack’s optical response and adjusts remaining layers to correct for drift, a practice known as error self-compensation. Combined with needle optimisation at the design stage, this is how suppliers consistently hit narrow-band or multi-band targets that would be unreliable with open-loop deposition alone, according to the SPIE optical coatings reference.
- Ion-assisted deposition and sputtering tend to produce denser, more stable films than simple thermal evaporation for demanding stacks.
- In-process optical monitoring with error self-compensation corrects for layer drift during the run itself.
- Monitor traces and witness coupons are the evidence a buyer should ask for, lot by lot.
A centrosymmetric multilayer stack of 15 to 19 layers can reach an extinction ratio of around 30 dB with wide angular bandwidth when properly optimised, a benchmark reported in research on thin-film polarising beam splitters, illustrating how layer count and optimisation quality directly set achievable performance.
Witness coupons and monitor trace records also underpin lot traceability, letting a buyer tie a delivered part’s measured spectral curve back to the specific deposition run that produced it.
Writing a procurement-ready specification and acceptance test plan
An RFQ for a dichroic coating should read like a test plan, not a description. Spectral performance only means something when tied to a measurement geometry, so every edge wavelength, passband ripple figure and stopband reflectance value needs its angle of incidence stated alongside it. Extinction ratio, where relevant to a polarising beam splitter, needs the same treatment.
Environmental and mechanical tests follow the same logic: specify the method, not just the pass criterion. Adhesion, abrasion, damp-heat and temperature cycling are the standard set, and for space or harsh-environment programmes, SN EN 16602-70-17 lists durability test types that cover exactly this ground.
- Define spectral metrics (edge wavelength, ripple, stopband reflectance, extinction ratio) at a named angle of incidence.
- Specify environmental and mechanical acceptance tests, referencing SN EN 16602-70-17 where the application demands it.
- Require surface-profile (RMS) measurement before and after coating, plus a rear-surface AR note if applicable.
- Request reporting format: instrument type and resolution, witness sample data and lot-level traceability records.
| Specification item | What to state in the RFQ |
|---|---|
| Spectral edge | Wavelength, tolerance band, angle of incidence |
| Passband/stopband | Ripple limit, stopband reflectance minimum |
| Surface profile | RMS limit, measured pre- and post-coating |
| Environmental class | Referenced standard (e.g. SN EN 16602-70-17), test sequence |
| Reporting | Instrument type, resolution, witness coupon data, lot number |
For worked RFQ language on transmittance targets, see our RFQ specs for optical window coatings.
Where dichroic coating technology came from
Interference coatings trace back to early twentieth-century optics work on thin-film reflection and transmission control, but dichroic filtering as an industrial capability matured alongside vacuum deposition technology in the second half of the century. Early dichroic mirrors were limited to relatively simple stacks with modest angular tolerance, suited to projection and broadcast optics where a single clean edge was enough.
The shift towards multi-band and ultra-broadband designs, now common in multispectral imaging and sensor fusion assemblies, followed directly from two parallel developments: computational design optimisation and in-process monitoring. Once needle optimisation made it practical to design stacks with uneven, non-quarter-wave layers, and once chambers could monitor and self-correct during deposition, the achievable complexity of a dichroic stack rose sharply. What once required a dozen separate single-band filters can now often be achieved in a single engineered multilayer.
This evolution matters for procurement because it changes what is reasonable to ask for. A spectral requirement that would have demanded a custom research programme decades ago may now be a standard capability, provided the substrate, environmental class and measurement geometry are specified clearly. Understanding that history helps a buyer judge whether a quoted lead time and price reflect routine manufacturing or a genuinely novel design challenge.
Long-term durability beyond the standard test sequence
Standard acceptance tests such as damp-heat, adhesion and temperature cycling establish a baseline, but real deployment conditions often run longer and harder than a qualification cycle. UV exposure over years can degrade certain fluoride and sulphide layers at the margins, particularly where a coating sits exposed rather than protected within a sealed assembly. Humidity cycling over a service life measured in years, rather than the days or weeks of a standard test, can reveal slow adhesion failure that a short damp-heat test misses.
Thermal cycling in service is rarely a clean sine wave either. Automotive dashboard optics see repeated rapid swings between a cold start and a hot cabin; aerospace and defence optics may see wide swings at altitude combined with vibration. These combined-stress conditions are harder to predict from single-variable test data alone.
We recommend treating standard test certificates as a floor, not a ceiling, for any component with a multi-year service life. Where the application involves sustained UV exposure, marine salt-fog conditions, or extended thermal cycling beyond the standard test envelope, it is worth requesting extended or combined-stress test data specific to the materials in the stack, rather than assuming a pass on SN EN 16602-70-17’s standard sequence guarantees performance over a decade of field use.
Dichroic coatings versus other optical coating approaches
Dichroic interference coatings are not the only way to control which wavelengths pass through an optical component. Absorptive filters, which use dyed or doped glass to remove unwanted wavelengths, are simpler to manufacture and often cheaper, but they waste the absorbed energy as heat and cannot achieve the sharp, steep edges that interference-based designs manage. For a high-power application or one where rejected light needs to be redirected rather than absorbed, an absorptive approach is often the wrong choice.
Single-layer or simple broadband anti-reflection coatings solve a different problem entirely: reducing reflection losses rather than selecting wavelengths, so they are not a substitute for a dichroic function but often a complementary one, applied to the rear surface of a dichroic window to recover transmission.
The genuine trade-off within interference coatings themselves is spectral breadth versus manufacturability. A narrow single-band dichroic filter is relatively straightforward to produce reliably. An ultra-broadband design covering several atmospheric windows, such as the Ge-substrate beam-splitting coating reported in recent MDPI research achieving average transmittance above 91% across both 3 to 5 micron and 8 to 12 micron bands, demands tighter process control, more layers and a greater sensitivity to stress and uniformity. Choosing dichroic interference technology over an absorptive alternative is usually right when steep edges, low insertion loss or multi-band performance matter; the harder decision is how much spectral breadth to demand before cost and yield start working against the programme.
Cost drivers, pricing ranges and lead times to plan around
Three variables drive dichroic coating cost more than any others: layer count, substrate fragility and spectral tolerance. A stack optimised to 15 to 19 layers for a demanding polarising beam splitter, of the kind described in thin-film polarising beam splitter research, carries more deposition time, more monitoring overhead and more scrap risk than a simple two-band edge filter, and that shows up directly in unit price.
Substrate choice compounds this. A fragile material such as germanium adds handling cost and scrap risk at every stage from cutting through coating, and any rear-surface AR requirement adds a second deposition run. Tight spectral tolerance at a specific angle of incidence, rather than a looser general-purpose spec, pushes both design time and in-process monitoring effort upward, since the margin for drift before a part fails inspection narrows.
Lead time follows the same logic: a simple, well-characterised coating on a robust substrate can often move from design to delivery faster than a novel multi-band design on a fragile substrate, where first-article qualification, environmental testing and surface-profile verification all add schedule. We do not publish a flat price list for coated optical components, because the combination of substrate, stack complexity and test requirements changes the manufacturing path for each project; pricing for services such as coated glass and optical components is available on request once a specification is defined.
Building the spectral and environmental requirements clearly into the RFQ from the outset, rather than iterating after quotation, is the single biggest lever a buyer has over both cost and schedule.
Where dichroic coating design is heading next
Nanostructured coatings, where sub-wavelength surface textures replace or augment conventional multilayer stacks, are an active area of development aimed at achieving broadband performance with fewer discrete layers and potentially lower stress. Novel fluoride and chalcogenide materials are also being explored to extend usable spectral range further into the long-wave infrared while improving environmental resistance compared with legacy materials.
On the design side, optimisation algorithms continue to push beyond needle synthesis towards more automated, multi-objective approaches that balance spectral performance against stress, absorption and manufacturability simultaneously rather than sequentially. This matters for multispectral sensor fusion applications, where a single component increasingly needs to manage several separated reflection and transmission bands at once, the kind of ultra-broadband performance already demonstrated on germanium substrates in recent beam-splitting coating research.
For procurement teams, the practical implication is that spectral requirements considered exotic a few years ago are steadily becoming achievable within standard production capability, provided the design brief allows the supplier room to optimise rather than mandating a fixed, legacy layer structure.
Choosing between a custom coating and an off-the-shelf filter
A bespoke coated component earns its cost when the window is integrated into an assembly, needs environmental qualification, or must hit a spectral response no catalogue part offers. An off-the-shelf filter or a different optical architecture altogether can be the faster, cheaper route when tolerances are loose and timelines are tight. Weigh qualification effort, lead time and expected service life before committing either way.
— Alexandra
Starting an RFQ for a dichroic coating with Precision Glasses
We design and manufacture custom thin-film coatings for industrial optics, from prototype runs through to volume production, with documented lot traceability and quality controls aligned to ISO 9001. When a catalogue filter will not meet your spectral target, substrate constraint or environmental class, that is where a direct manufacturing relationship earns its keep: we work from your specification through to a delivered, tested component rather than leaving you to translate a datasheet into a working part.
A useful RFQ to send us includes:
- Substrate material, thickness and orientation.
- Spectral targets and tolerance at the required angle of incidence.
- Mechanical tolerances, including allowable RMS surface deformation.
- Environmental test class and any referenced standard.
- Sample quantity and delivery timeline.
A good first step is to request a capability statement along with witness coupons from a comparable run, and to ask for any prequalification test reports already on file for similar substrates or coatings. You can review our services or browse our optical components catalogue, then send your specification through for a quote.
FAQ
What is the difference between a dichroic coating and a standard AR coating?
A dichroic coating is designed to reflect specific wavelengths while transmitting others, creating a sharp spectral edge or band. A standard anti-reflection coating is designed to reduce reflection losses broadly rather than select wavelengths, and the two are often used together on the same component.
Which materials are typically used in infrared dichroic stacks?
Germanium, zinc sulphide and ytterbium fluoride are common choices for shortwave through long-wave infrared dichroic stacks, selected for their refractive index and transparency in the target band. Material choice is paired with substrate compatibility and coating stress tolerance, as described in recent beam-splitting coating research.
How many layers does a typical dichroic filter need?
Layer count depends on the spectral requirement rather than a fixed rule, and can range from a handful for a simple edge filter to the high teens for a demanding polarising beam splitter. Research on thin-film polarisers describes 15 to 19 layer designs reaching extinction ratios around 30 dB.
What should a procurement specification for a dichroic coating include?
It should state spectral edges, ripple and stopband reflectance at a defined angle of incidence, along with substrate material, allowable surface deformation and the environmental test standard. Referencing a standard such as SN EN 16602-70-17 is useful for harsh-environment or space applications.
Does Precision Glasses manufacture custom dichroic-coated components?
Yes, we design and manufacture custom thin-film coatings on specified substrates, from prototype through to volume production, with lot traceability and quality controls aligned to ISO 9001. Pricing depends on substrate, spectral complexity and test requirements and is available on request through our services page.
Sources
- Optical coatings: design, characterization, monitoring (SPIE ebook)
- Design and deposition of ultra-broadband beam-splitting coatings (MDPI, 2025)
- Thin-film polarizers and polarizing beam splitters (Optical Society paper)
- SN EN 16602-70-17 – Space product assurance – Durability testing of coatings and surface finishes
