Application engineers don’t just use SWIR imaging systems to “see in the dark.” These tools take advantage of specific molecular absorption bands and the inherent properties of materials like silicon using indium gallium arsenide (InGaAs) chemistry.
Physics and System Design: Why SWIR?
Choosing between SWIR and other thermal or visible bands comes down to the source of the photons.
- SWIR vs. VIS/NIR: In the visible world, we see reflected light. SWIR also relies on reflected light, unlike the emissive “heat” signatures of Middle Wavelength Infrared (MWIR) or Low Wavelength Infrared (LWIR), but it interacts with molecular bonds rather than just surface color.
- SWIR vs. MWIR/LWIR: While LWIR (8–14 µm) detects the thermal glow of objects at room temperature, SWIR captures photons from high-temperature processes (greater than 250 degrees Celsius) with much higher spatial resolution and lower atmospheric interference.
The InGaAs Advantage
Most SWIR cameras utilize an InGaAs sensor lattice-matched to an InP substrate. This configuration provides high Quantum Efficiency (QE) between 900 nm and 1700 nm. “Extended SWIR” systems involve lattice-mismatched layers, which increase dark current and typically require sophisticated multi-stage thermoelectric cooling (TEC) to maintain a functional signal-to-noise ratio (SNR).
Semiconductor Inspection: Silicon Transparency and Wafer Metrology
From a physics standpoint, Silicon has a bandgap of approximately 1.12 eV. This means photons with wavelengths shorter than 1100 nm are absorbed, while wavelengths longer than 1200 nm pass through the substrate with minimal attenuation.
In semiconductor inspection and wafer testing, SWIR imaging allows better analysis of MEMS and 3D-IC stacking. SWIR cameras minimize chromatic aberration and maximize the contrast of sub-surface defects or alignment marks.
High-Speed Sorting in Food and Recycling
In high-volume industrial sorting, area-scan cameras often hit a frame-rate bottleneck. Line scanning cameras using InGaAs sensors solve this by capturing a single continuous “slice” of the conveyor at frequencies up to 40 kHz. This process is particularly useful for:
- Moisture Mapping: Water has a massive absorption peak at 1450 nm. Under SWIR, a hydrated bruise on an apple or a moisture-rich foreign object in a grain stream appears nearly black, while the dry product reflects brightly.
- Polymer Discrimination: Recycling streams rely on the “overtone” vibrations of C-H bonds. Different plastics (PET vs. HDPE) have distinct spectral signatures in the 1200–1700 nm range, allowing for chemical-grade sorting that visible sensors cannot achieve.
Extended SWIR (eSWIR) for Energy and Composites
Standard InGaAs cuts off at 1.7 µm. However, for R&D in battery technology and solar cells, the 1.9–2.5 µm range is critical.
- Battery Inspection: eSWIR can detect subtle variations in the chemical slurry coating on electrode foils, identifying “dry spots” or thickness variances that lead to thermal runaway.
- Photoluminescence (PL): For solar wafer metrology, capturing the weak emission of photons as electrons recombine requires high-sensitivity eSWIR sensors to map grain boundaries and crystal dislocations.
Additive Manufacturing and Welding Monitoring
Process monitoring in DED (Directed Energy Deposition) or PBF (Powder Bed Fusion) is plagued by “plasma blinding.” Visible cameras see only the arc or the laser spark. By using a SWIR sensor with a narrow bandpass filter (1400–1600 nm), you bypass the intense visible emissions. This allows for:
- Melt Pool Geometry: Accurate measurement of the liquid-to-solid transition.
- Planck’s Law Integration: Since the SWIR band follows the short-wavelength side of the Planck curve for metals at 500 to 1500 degrees Celsius, it provides a more stable thermal proxy than visible-light “color temperature” methods.
Free-Space Optical Communications (FSOC) and LIDAR
The atmosphere is a chaotic medium for visible light due to Rayleigh scattering. However, “atmospheric windows”—specifically around 1550 nm—offer significantly lower attenuation in SWIR applications.
- LIDAR: Moving from 905 to 1550 nm allows for “eye-safe” operation at much higher power levels (orders of magnitude higher). This is because 1550 nm light is absorbed by the eye’s vitreous humor before it can damage the retina.
- Long-Range Imaging: SWIR can “see” through haze and smoke that completely obscures visible cameras, making it the standard for long-range maritime and border surveillance.
Non-Destructive Art and Heritage Analysis
For conservation scientists, SWIR provides a non-invasive way to analyze hidden layers in paintings and other objects.
- Reflectography: Carbon-based underdrawings (charcoal) absorb SWIR heavily, while the overlying pigment layers (oils and tempera) often become transparent.
- Material ID: Application engineers use hyperspectral SWIR imaging to identify specific mineral pigments, helping to date works or identify non-original restoration efforts without removing a single flake of paint.
Which SWIR System Do You Need?
When specifying a system, the “camera” is only part of the solution. These other factors will also play a significant role in using SWIR in application engineering:
| Feature | Consideration | Why it matters |
|---|---|---|
| Lens Coating | VIS-NIR vs. SWIR-Optimized | Standard glass loses 30% transmission in SWIR. |
| Cooling | Uncooled vs. TEC1 vs. TEC3 | Determines the "floor" of your signal-to-noise ratio. |
| Pixel Pitch | 10 µm to 25 µm | Balances spatial resolution with light-gathering capability. |
| Illumination | Narrow band LED vs. halogen | Broad halogen sources produce heat; LEDs provide spectral purity. |