# Thermal Camera Window: Germanium, AR Coating and Readings Why glass and most plastics block long-wave infrared, what a germanium window and its coatings need, how a window shifts readings, and how to size it. Source: https://thermal.torsuntech.com/resources/thermal-camera-protective-window/ [Resources](https://thermal.torsuntech.com/resources/) / Protective window # Putting a thermal core behind a window: why glass fails, what germanium needs, and how the window changes your readings Most products built around a thermal core need something between the lens and the outside world: rain, dust, cleaning cloths, fingers. The window is easy to get wrong, because the material you would use on a visible camera does not work, and on a radiometric product the window becomes part of the measurement. Here is what to check. Core figures come from the 640 × 512 radiometric module we sell; the window physics applies to any long-wave core. ## Why glass and most plastics do not work Uncooled cores see the long-wave infrared band, 8–14 μm. Ordinary glass, polycarbonate and acrylic absorb strongly there, so behind a glass window the core sees the glass, not the scene. The window has to be an infrared material. Germanium is the most common choice for this band; zinc selenide and chalcogenide glasses are also used. ## Why the coatings matter as much as the material Germanium has a refractive index of about 4, so each uncoated surface reflects about 36% and an uncoated plate passes under 50% (calculated). Anti-reflection (AR) coatings for the long-wave band fix that: one optics catalogue lists 90% or more transmission from 8 to 11.5 μm for an AR-coated germanium window (published figure). On an exposed outer face, a hard diamond-like carbon (DLC) coating protects against abrasion and cleaning. Ask for the transmission curve across 8–14 μm, not one peak value: some coatings are specified only up to 11.5 or 12 μm. ## How the window changes a reading Through a window the core sees three things at once: the target, reduced by the window’s transmission; the window’s own emission at the window’s temperature; and reflections of whatever is behind the window, such as the inside of the housing. Without correction, the reading is pulled towards the window and housing temperature. A simple example, calculated for a black target, the 8–14 μm band and a window that transmits 90% and absorbs the rest: - Window at 25°C: a 60°C target reads about 57°C, 100°C about 94°C, 150°C about 141°C - Window warmed by the sun to 45°C: a 25°C target reads about 27°C The error grows with the difference between target and window temperature, and it shifts whenever the window temperature shifts. ## Correct it, or calibrate through it Two approaches work: - Correct for the window in your measurement chain. You need its transmission and reflectance in your band, and its temperature. If sun or a heater changes the window temperature, measure it instead of assuming it. - Check the finished assembly against a reference with the window fitted: a blackbody if you have one, or a matte black surface measured with a good contact probe, at the temperatures your product measures. Do one of these before you trust absolute numbers. The accuracy figure of the core in this example, ±2°C or ±2% (typical, lab conditions), applies to the core alone, not to the core behind your window. ## Size and position The window must not clip the field of view. Its clear aperture has to cover the front of the lens plus 2 × spacing × tan(half the diagonal field). For every millimetre between lens and window, the aperture grows by about 1.15 mm for a 9.1 mm lens on a 640 × 512 core (48.7° × 38.6°) and by about 1.6 mm for a 4 mm lens on a 384 × 288 core (64.1° × 48.8°) (calculated). Wide lenses therefore need the window very close. Tilting the window a few degrees keeps reflections of the lens and the housing interior out of the image; allow for the extra aperture this needs. ## Outdoors - A bare core has no ingress rating; the housing and its window seal provide it. Fixed outdoor thermal cameras on the market are commonly rated IP66 (published ratings). - Water, ice and dirt on the window absorb long-wave infrared. A wet window gives wrong readings. - Measurement is specified for a narrower ambient range than imaging: −10°C to +70°C against −40°C to +80°C on this core. In cold climates plan a heater, and remember that it changes the window temperature too. - The manual for this core says not to point it at the sun. Mount the camera so the sun stays out of the field of view. ## Before you order windows - Material, and an AR coating band that covers 8–14 μm - Transmission and reflectance curves at your angle of incidence - Clear aperture at your lens-to-window spacing - Outer coating for your environment and cleaning method - How you will correct for the window or calibrate through it, and how you will know its temperature [View MINI 640-2 Radiometric](https://thermal.torsuntech.com/products/mini-640-2-radiometric/) ## Planning a window for your product? Send the lens, the window spacing and what the product measures. We reply within 24 hours on working days. [Discuss your requirements](https://thermal.torsuntech.com/contact/) --- TORSUN TECH · CATC Building 2, Shipai Town, Dongguan, Guangdong, China · info@torsuntech.com · Tel / WeChat / WhatsApp +86 180 2518 5996 · Enquiry: https://thermal.torsuntech.com/contact/