SWIR Camera Lens for Semiconductor and Silicon Inspection: How to Select Optics for Wafer, Material and Hidden-Feature Imaging

Semiconductor inspection becomes much more challenging when the feature that matters is not located on an exposed surface. Alignment structures can sit behind silicon, contaminants can become trapped between bonded wafers, damage can develop below the surface, and packaging processes can conceal interfaces that a conventional visible-light inspection system cannot access optically. This is one of the areas where a properly designed SWIR camera lens for semiconductor inspection becomes technically valuable: silicon changes from strongly absorbing in the visible region to substantially more transmissive at wavelengths beyond roughly 1100 nm, allowing appropriately configured infrared imaging systems to observe features through silicon that are otherwise hidden from a visible camera.

That does not mean every 900–1700 nm inspection automatically sees through every wafer. Silicon transmission depends on wavelength, wafer thickness, doping, surface condition and the structures deposited on or within the device. Metals remain strongly opaque, highly doped regions can attenuate infrared energy, and interfaces can introduce substantial reflection. The engineering task is therefore not simply to buy an infrared lens, illuminate the wafer and expect hidden defects to appear. The optical system must be designed around the specific silicon structure, target feature, illumination wavelength, camera sensitivity, field of view, working distance, required spatial sampling and defect-contrast mechanism.

The dedicated Kyptec Automation® SWIR Camera Lens collection provides 8.5 mm, 12.5 mm, 25 mm, 35 mm and 50 mm focal-length options designed for 900–1700 nm imaging, with 2 MP, 2/3-inch, F1.4 and C-Mount specifications across the current portfolio. Kyptec Automation® specifically identifies semiconductor inspection among the intended applications of the range. For OEMs developing wafer inspection, silicon material analysis or through-silicon machine-vision equipment, this gives a useful optical platform in which focal length can be selected according to inspection geometry while retaining a common SWIR architecture.

Why Silicon Becomes Inspectable in SWIR

The optical behaviour of crystalline silicon changes sharply near its band-gap absorption edge. At visible wavelengths, silicon is essentially opaque for conventional imaging. At wavelengths above approximately 1100 nm, its absorption falls dramatically and silicon becomes increasingly transmissive. Scientific literature places the silicon band-gap energy near 1.12 eV and confirms that silicon becomes largely transparent in the near-infrared beyond roughly 1100 nm.

This creates an important inspection window inside a 900–1700 nm SWIR system.

Instead of imaging only what lies on the front surface, the machine can potentially collect information from structures beneath the surface or on the opposite side of the wafer. Industrial SWIR imaging is consequently used for through-silicon inspection, wafer-pattern observation, backside feature alignment, bonded-wafer analysis and identification of hidden defects or foreign material. Demonstrated semiconductor imaging around 1100 nm shows that internal wafer patterns can become observable in transmission.

For an OEM, the key point is that silicon transparency is wavelength-dependent rather than binary. Selecting the optical band deliberately is therefore essential.

The Inspection Target Must Be Defined Before the Lens

“Semiconductor inspection” is too broad a specification for choosing optics.

A lens for viewing a complete wafer occupies a completely different optical design space from a lens intended to inspect a small bonding region. Similarly, detecting a relatively large contaminant between wafers differs significantly from looking for narrow cracks, alignment marks or small void-like features.

Before selecting focal length, the OEM should define four quantities: the physical inspection field, the smallest feature that must be distinguished, the available camera-to-wafer distance and the silicon thickness or stack through which the feature is being observed.

Those values determine whether a wide-area, intermediate or tighter-field SWIR optical geometry is appropriate.

Through-Silicon Imaging Is Not the Same as Surface Inspection

In ordinary reflected-light inspection, image contrast is generated mainly by surface reflectance and geometry. Through-silicon imaging adds another optical path: radiation must enter the silicon, propagate through it, interact with the target feature or interface and then reach the camera.

Every stage can reduce signal.

Surface reflection at the silicon-air interface can be significant because silicon has a relatively high refractive index. Doping can also increase absorption, while metal layers and certain structures may block transmission almost completely.

As a result, a feature that looks strong through a thin lightly doped wafer may become much weaker through a thicker or differently processed wafer.

The inspection system should therefore be characterized using representative production wafers, not only polished silicon test coupons.

Transmission Geometry Is Often the Strongest Starting Point

For many hidden-feature applications, transmission imaging provides a direct architecture: the SWIR illumination is positioned on one side of the silicon and the camera is positioned on the other.

Regions that transmit SWIR energy differently create image contrast.

This approach can reveal internal patterns and structures that are inaccessible in visible imaging. Through-silicon transmission is also useful for examining relationships between features on opposite wafer surfaces.

Industrial demonstrations have shown silicon wafer patterns at approximately 1100 nm using transmission illumination, confirming the practical value of this configuration for internal pattern observation.

Transmission geometry should therefore be evaluated early when access to both sides of the wafer is mechanically possible.

Reflection Geometry Has a Different Role

Not every semiconductor machine can place illumination behind the wafer. In some equipment, the process tool, chuck or substrate carrier blocks the opposite side.

Reflection-mode SWIR inspection can still provide useful information, particularly for surface and near-surface structures, material differences or particular device geometries.

However, reflected signal from silicon can be strongly influenced by interface reflections, wafer angle, coatings and surface finish. Specular highlights may dominate portions of the image and obscure low-contrast defects.

For this reason, the illumination angle should be developed together with the lens geometry rather than added after the camera position is finalized.

The Useful Wavelength Should Be Chosen for the Silicon Stack

Although silicon begins becoming substantially transmissive above approximately 1100 nm, there is no single universally optimum semiconductor-inspection wavelength.

Shorter wavelengths near the transition can provide different contrast than wavelengths farther into the SWIR range. As wavelength increases, silicon absorption generally decreases, but camera sensitivity, coatings, illumination availability and the optical characteristics of surrounding materials also change.

The best development process is therefore experimental: image representative wafers at several SWIR wavelength bands and quantify the contrast of the actual target feature.

The Kyptec Automation® SWIR lens family is designed across 900–1700 nm, allowing an OEM to develop within a broad industrial SWIR window rather than selecting an optic optimized only around visible wavelengths.

Hidden-Feature Visibility Depends on Contrast, Not Simply Penetration

Being able to transmit light through silicon is only the first requirement.

A hidden defect must also produce enough optical contrast relative to its surroundings to be detected reliably.

Consider a contaminant trapped between two bonded wafers. If the contaminant absorbs or scatters SWIR radiation differently from silicon, it may produce a visible intensity difference. A void can create another type of interface response. A metal structure may appear strongly opaque against transmissive silicon. Cracks can alter scattering and transmission depending on their orientation and geometry.

Industrial SWIR semiconductor inspection has been applied to particles between bonded wafers, through-wafer structures, subsurface damage, bonding defects, alignment features and related hidden conditions.

The lens must preserve enough spatial and contrast information for those differences to survive the complete optical chain.

Wafer Inspection Should Be Designed From Required Object Sampling

One of the most consequential OEM decisions is how much wafer area to assign to each camera pixel.

Suppose an imaging sensor has 1600 horizontal pixels. If the optical system covers 160 mm of wafer width, the nominal object sampling is approximately 10 pixels/mm, or about 100 µm per pixel before optical resolution is considered. If the field is reduced to 40 mm, nominal sampling improves to approximately 25 µm per pixel.

This does not mean a 25 µm defect is automatically detectable. Reliable defect detection normally requires multiple pixels across the target feature, sufficient modulation from the lens, adequate signal-to-noise ratio and meaningful SWIR contrast.

However, this simple calculation immediately tells an OEM whether the proposed field of view is even in the correct range.

A complete wafer and a microscopic local feature often should not be inspected with exactly the same optical station.

Why Full-Wafer Inspection and Local Defect Inspection May Need Different Optics

If the goal is to inspect a broad region or establish wafer-level alignment, relatively wide coverage is useful. If the goal is to detect smaller hidden structures, concentrating the camera pixels over a smaller field becomes more valuable.

This creates a natural two-stage architecture for some machines: a wider field for localization and a narrower optical station for detailed inspection.

An OEM should resist the temptation to make one focal length perform every task. Excessive FOV reduces effective object sampling, while an excessively narrow field increases inspection time and positioning requirements.

The dedicated focal-length progression within the Kyptec Automation® SWIR Camera Lens collection allows these geometries to be evaluated without moving outside the same SWIR lens category.

Kyptec Automation® KL-1412 for Intermediate Semiconductor Inspection Geometry

For semiconductor systems requiring a moderate inspection field rather than full wide-area coverage, the Kyptec Automation® KL-1412 25 MM SWIR Camera Lens provides a useful intermediate focal length.

A 25 mm geometry can be evaluated for localized wafer inspection, silicon material analysis and stations where the camera should devote a greater proportion of its pixels to the inspection region while preserving practical working distance.

Its F1.4 aperture also provides useful optical throughput when transmission through the silicon stack reduces the signal reaching the camera.

The final operating aperture should nevertheless be established from actual feature contrast and depth-of-field requirements rather than automatically left fully open.

Kyptec Automation® KL-1414 for Tighter Wafer and Bonding Inspection

The Kyptec Automation® KL-1414 35 MM SWIR Camera Lens provides a narrower field class within the portfolio and is specifically identified for semiconductor inspection among its SWIR applications. Its published product specifications include 35 mm focal length, F1.4 aperture, 2 MP, 2/3-inch format, C-Mount and 900–1700 nm operation.

This type of focal length is particularly useful to evaluate for bonding regions, die-level inspection, smaller wafer regions and hidden-feature imaging where the complete wafer does not need to occupy the image.

By removing unnecessary surrounding area from the FOV, the system can devote more of the camera's spatial sampling capability to the feature that actually determines pass or fail.

Kyptec Automation® KL-1416 for Narrow Inspection Fields and Greater Stand-Off

A semiconductor machine sometimes needs a narrow inspection field while keeping the camera farther from the wafer because of process hardware, illumination clearance, motion stages or protective structures.

The Kyptec Automation® KL-1416 50 MM SWIR Camera Lens gives the portfolio its longest focal-length geometry and can be evaluated for these tighter inspection stations. Its longer focal length makes it useful where greater stand-off and a more concentrated field are required.

The 50 mm option should not be selected merely because a longer lens produces greater apparent magnification. The machine still needs enough field margin for stage tolerance, wafer placement and feature localization.

For automated equipment, a tightly framed laboratory image that loses the target after a small positioning error is not a production-ready optical design.

Positioning Accuracy Becomes More Important as FOV Narrows

This is one of the main differences between laboratory imaging and automated semiconductor equipment.

A technician can manually center a feature under a camera. A production tool must find it repeatedly despite wafer-loading tolerance, stage repeatability and part variation.

As focal length increases and FOV becomes narrower, the allowable positioning error decreases.

OEMs should therefore reserve sufficient image margin around the target feature and include stage tolerance in the FOV calculation.

The technically optimum lens is not necessarily the one that fills 100% of the frame with the region of interest. It is the one that provides enough sampling and enough positional tolerance for stable production.

Defect Size Should Be Specified Before Resolution Requirements

A statement such as “we need a 2 MP SWIR system” is incomplete.

The meaningful specification is the smallest physical defect that must be detected across a defined field.

If the machine must identify a 100 µm feature, the engineer should determine how many image pixels are required across that feature for robust classification, then work backward to the maximum acceptable FOV.

If the target is an alignment structure several millimetres wide, a much broader field may be permissible.

This object-space calculation should precede final focal-length selection.

Focus Through Silicon Must Be Qualified at the Actual SWIR Wavelength

A lens can appear sharply focused under visible setup illumination and still produce an unacceptable SWIR image.

The production focus should therefore be established using the actual SWIR band employed for inspection.

This becomes especially important when the target lies beneath the silicon surface. The optical path through silicon differs from focusing directly on an exposed object in air, and the apparent position of an internal feature can shift because of refraction.

For through-silicon imaging, the correct focus is the focus that maximizes contrast at the hidden target plane—not necessarily the apparent physical surface of the wafer.

Wafer Thickness Changes the Optical Problem

A thin wafer and a thick silicon substrate should not automatically use identical exposure and illumination settings.

Greater silicon thickness increases the optical path inside the material. Even at wavelengths where silicon is relatively transmissive, absorption and interface losses can reduce the available signal.

Doping and other material properties can further modify transmission.

An OEM building equipment for several wafer thicknesses should therefore test the extreme ends of the production range. A configuration optimized only for the thinnest sample can become signal-starved on a thicker substrate.

Bonded Wafers Add Interfaces That Can Create Both Information and Artefacts

Bonded structures are especially valuable targets for SWIR because hidden interfaces may contain particles, voids, bubbles, delamination or alignment structures. SWIR inspection is used industrially for through-silicon alignment and detection of hidden defects in bonded or stacked semiconductor structures.

Those same interfaces can also create unwanted reflections.

An OEM should therefore assess whether a suspicious contrast feature tracks with the physical defect or changes dramatically with illumination angle.

If contrast disappears when the lighting geometry changes slightly, the system may be detecting interface reflection rather than the defect itself.

Backside Alignment Requires Stable Geometry, Not Merely Visibility

One major benefit of through-silicon imaging is the ability to relate a feature on one side of a wafer to a feature on the opposite side.

For alignment systems, simply seeing both structures is insufficient.

The optical system must maintain geometric stability across the working field. Lens distortion, camera mounting, wafer height, stage positioning and focus repeatability all influence registration accuracy.

If the SWIR image will be used quantitatively for alignment rather than only qualitative inspection, machine calibration should therefore account for the complete lens-camera-stage geometry.

A tighter field using an intermediate or longer focal length may improve pixel utilization, but calibration remains necessary.

F1.4 Helps When the Silicon Stack Reduces Available Signal

Transmission through wafers, filters and additional interfaces can reduce the optical energy reaching the sensor. A larger available aperture therefore provides useful exposure margin.

The F1.4 specification across the Kyptec Automation® SWIR family is valuable in this context because it gives the designer the option of stronger light collection.

However, semiconductor inspection can also demand substantial depth of field or strong edge-to-edge image quality. The lens may therefore perform best at a smaller production aperture.

The correct aperture should be determined experimentally by measuring hidden-feature contrast, image sharpness and exposure requirement at several settings.

Reflections Should Be Managed Before Increasing Software Complexity

Polished silicon can create strong reflections. If these reflections saturate part of the image, software may struggle to recover the hidden information underneath.

The first solution should normally be optical rather than algorithmic.

Change the illumination incidence angle, evaluate transmitted rather than reflected geometry where possible, avoid placing bright source images directly into the lens, and stabilize the wafer orientation.

A cleaner optical signal reduces the burden on downstream image processing and tends to produce a more transferable OEM system.

Hidden-Feature Inspection Needs Known-Good and Known-Bad Samples

Semiconductor defects are often low contrast, and some legitimate process variation can resemble defects.

A production qualification should therefore include representative samples containing confirmed:

acceptable structures;

known hidden particles;

bonding anomalies;

alignment offsets;

subsurface or edge damage;

and relevant normal process variation.

The important metric is not whether the defect is visible in a selected demonstration image. It is whether the optical configuration maintains measurable separation between good and bad conditions across the expected manufacturing range.

Evaluate Center, Edge and Corner Performance

A semiconductor feature should not become significantly harder to detect simply because it appears near the corner of the inspection field.

During optical qualification, move representative structures across the entire FOV and compare contrast, sharpness and registration.

If a feature is strong in the centre and weak at the edge, investigate illumination uniformity, focus plane, lens performance and wafer alignment.

For OEM acceptance, the worst qualified location matters more than the best demonstration image.

Do Not Freeze the Mechanical Design Before Optical Testing

Camera brackets, enclosure height, illumination access and wafer-stage spacing are often committed too early.

Once those dimensions are fixed, optical engineers may be forced to select a focal length that is merely compatible rather than optimum.

For a SWIR semiconductor system, prototype the camera, lens, wafer and illumination geometry as one assembly before finalizing the production enclosure.

This is particularly important if 35 mm or 50 mm optics are being considered because their preferred working geometry can differ substantially from wide-angle configurations.

A Practical OEM Qualification Sequence

Begin with a representative silicon stack and identify the exact hidden structure or defect that must be detected. Test several SWIR wavelength bands above the silicon transmission edge and determine where target-to-background contrast is strongest. Then establish the physical field of view and calculate whether the smallest feature receives enough object-side sampling.

Select a candidate focal length from the Kyptec Automation® SWIR family and establish the practical working distance. Optimize illumination geometry before tuning exposure aggressively. Focus on the actual hidden plane using the production SWIR wavelength. Then test aperture, exposure and illumination levels while monitoring target contrast rather than apparent image brightness alone.

Finally, repeat the measurement across wafer thickness, stage position, field location, process variation and representative defect classes.

Only after the optical configuration remains stable through those tests should focal length, aperture, camera height and illumination positions be frozen for production.

Why Kyptec Automation® Is a Strong SWIR Lens Platform for Semiconductor OEMs

The Kyptec Automation® SWIR Camera Lens portfolio is particularly well aligned with semiconductor-machine development because its 900–1700 nm design range covers the spectral region in which silicon becomes increasingly transmissive beyond approximately 1100 nm. The range also provides five focal lengths rather than forcing every wafer and hidden-feature application into one fixed field of view.

This lets OEMs approach semiconductor optics systematically: shorter focal lengths can address broader fields, 25 mm can support intermediate inspection regions, and 35 mm or 50 mm can concentrate the available sensor resolution on smaller wafer, die or bonding areas. The common F1.4, 2 MP, 2/3-inch and C-Mount architecture further gives machine builders a coherent specialized optical family for prototype development, machine variants and repeat production.

Frequently Asked Questions About SWIR Camera Lenses for Semiconductor and Silicon Inspection

1. At what wavelength does silicon start becoming transparent enough for SWIR inspection?

Silicon absorption decreases sharply around its approximately 1.1 µm band-gap transition, and wavelengths above roughly 1100 nm are commonly used for through-silicon imaging. The exact useful wavelength depends on silicon thickness, doping and the structure being inspected. An OEM should therefore characterize representative wafers across several wavelengths rather than assuming that every wavelength above 1100 nm will provide equal transmission or defect contrast.

2. Can SWIR imaging see completely through a silicon wafer?

It can see through many silicon wafers under suitable wavelength and illumination conditions, but “transparent” should not be interpreted as perfectly clear. Thickness, doping, surface coatings, metallization and internal structures can attenuate or block SWIR radiation. The correct test is whether the target feature retains sufficient contrast through the actual production wafer.

3. Can a SWIR camera lens reveal particles trapped between bonded wafers?

Yes, this is an established use of through-silicon SWIR inspection. When silicon transmits the selected wavelength and a trapped particle has different absorption, scattering or opacity, it can become distinguishable between bonded layers. Industrial SWIR systems are used for detecting particles and other hidden defects in wafer-bonding processes.

4. Can SWIR be used to align front-side and backside wafer features?

Yes. Through-silicon imaging can make structures on opposite sides of silicon accessible within the same optical workflow, supporting backside alignment and wafer-to-wafer registration. For precise alignment, however, lens distortion, stage calibration, wafer height and optical geometry must be controlled; simply making the backside mark visible does not guarantee registration accuracy.

5. Why are metal features still dark when silicon becomes transparent in SWIR?

Silicon transmission above its absorption edge does not mean every semiconductor material becomes transparent. Metal interconnects and other highly absorbing structures can remain opaque in SWIR. This difference can actually provide useful contrast because a metal feature may appear clearly against a comparatively transmissive silicon background.

6. Does silicon doping affect SWIR transmission?

Yes. Doping can alter infrared absorption, particularly in more conductive silicon. Two wafers with similar thickness but different doping characteristics may therefore produce different transmission levels. OEM qualification should include the real silicon grades used in production instead of treating one wafer sample as representative of every device.

7. Which is better for semiconductor inspection: reflected SWIR or transmitted SWIR?

Neither geometry is universally better. Transmission is particularly powerful for through-wafer patterns, hidden structures and interfaces when both sides of the wafer are optically accessible. Reflection is useful when backside illumination is mechanically impossible or when surface and near-surface information is important. The target feature and machine architecture should determine the configuration.

8. How do I choose between a 25 mm and 35 mm SWIR lens for wafer inspection?

Calculate the required field and smallest feature at the available working distance. The 35 mm option produces a tighter field and generally places more sensor pixels over a given local region, while 25 mm gives broader coverage. The Kyptec Automation® KL-1412 25 MM SWIR Camera Lens is useful for intermediate fields, while the Kyptec Automation® KL-1414 35 MM SWIR Camera Lens can be evaluated for tighter semiconductor inspection regions.

9. Is a 50 mm SWIR lens always better for detecting smaller wafer defects?

No. A 50 mm lens can provide a narrower field and greater object representation at a given geometry, but the resulting working distance and positioning tolerance must remain practical. If wafer placement variation causes the target to leave the field, the additional magnification does not improve the production system. Focal length must balance sampling with machine tolerance.

10. How many pixels should a semiconductor defect occupy?

There is no universal minimum because detectability depends on contrast, lens modulation, sensor noise and the classification algorithm. A defect represented by only one or two pixels is generally much less robust than one spanning several pixels. OEMs should therefore establish the minimum required pixel footprint experimentally using the hardest valid defect rather than relying solely on theoretical pixel size.

11. Why can an internal wafer feature look blurred even when the surface is sharply focused?

The hidden feature lies at a different optical plane, and light travelling through silicon experiences a different refractive path than light reflected from the exposed surface. The focus that makes the wafer surface sharp may therefore not maximize contrast at the internal target. Production focus should be established directly on the hidden feature under the actual SWIR wavelength.

12. Can SWIR inspection find cracks inside silicon?

SWIR can reveal certain subsurface cracks and damage because these structures can alter transmission or scattering through silicon. Detectability depends strongly on crack size, orientation, depth, illumination geometry and wavelength. The OEM should validate the smallest production-relevant crack rather than assuming that all internal cracks will appear with equal contrast.

13. Can SWIR detect bonding voids or delamination?

Potentially yes. Voids, bubbles and delamination create interface changes that can alter transmitted or reflected SWIR intensity. These applications are used in bonded and stacked semiconductor inspection, but contrast depends on interface geometry and surrounding materials. A production system should test known defects at multiple positions and wafer conditions before acceptance thresholds are established.

14. Why does wafer thickness matter when selecting exposure?

A thicker silicon path can attenuate more optical energy before it reaches the camera. Consequently, a configuration that gives excellent signal through a thin wafer can require additional exposure or illumination for a thicker substrate. Exposure should therefore be qualified at the maximum relevant wafer thickness rather than optimized only with an easy-to-image sample.

15. Should I focus the system using visible light before switching to SWIR?

Visible light may be convenient for mechanical setup, but it should not be the final focusing reference for a through-silicon inspection. The production focus must be optimized at the actual SWIR inspection wavelength and target plane because the optical focus condition can differ. Qualification should be based on hidden-feature contrast rather than the appearance of the visible wafer surface.

16. How can I reduce reflections from polished silicon during SWIR inspection?

Start by optimizing illumination angle and separating the desired signal from direct specular reflection. Transmission illumination is worth evaluating when the machine permits access to both sides. The wafer should also be held at a controlled orientation. Optical reflections should be minimized before relying on image processing to correct saturated or highly non-uniform regions.

17. What information should an OEM provide before selecting a SWIR lens for silicon inspection?

Provide the camera sensor dimensions, required field of view, smallest hidden feature, wafer thickness, material or doping range, intended SWIR wavelength, available working distance, inspection geometry, positioning tolerance and whether the target is at the surface, backside, inside silicon or between bonded layers. These parameters allow the lens to be selected from the actual inspection problem rather than from the general phrase “semiconductor inspection.”

18. Why is Kyptec Automation® a strong option for semiconductor and through-silicon SWIR inspection?

Kyptec Automation® provides a dedicated SWIR Camera Lens family operating across 900–1700 nm, which includes the wavelength region above approximately 1100 nm where silicon becomes increasingly transmissive. The portfolio spans 8.5 mm, 12.5 mm, 25 mm, 35 mm and 50 mm focal lengths while maintaining a common 2 MP, 2/3-inch, F1.4 and C-Mount architecture. This gives semiconductor OEMs a focused optical platform that can be adapted from broader wafer inspection to tighter hidden-feature, bonding and silicon-material inspection without changing to an unrelated lens family.

Conclusion

The value of a SWIR camera lens for semiconductor and silicon inspection comes from a fundamental optical property of silicon: above roughly 1100 nm, silicon absorption decreases enough for appropriately designed imaging systems to access information that is hidden from conventional visible inspection. That capability opens important industrial possibilities including through-wafer pattern observation, backside alignment, bonded-wafer inspection, hidden-particle detection, subsurface damage analysis and inspection of structures located beneath silicon.

The strongest semiconductor inspection systems, however, are not designed around the statement that “SWIR can see through silicon.” They are engineered around a much more precise question: at what wavelength, through what silicon stack, over what field of view and with what object sampling can the required hidden feature be detected reliably across production variation? Wafer thickness, doping, metallization, interfaces, reflection, focal length, exposure, working distance and focus at the true target plane all influence the answer.

For broader wafer regions, shorter SWIR focal lengths may provide the required inspection coverage. For localized silicon structures, the Kyptec Automation® KL-1412 25 MM SWIR Camera Lens provides a useful intermediate geometry. For tighter bonding, die or hidden-feature inspection, the Kyptec Automation® KL-1414 35 MM SWIR Camera Lens and longer focal-length alternatives can concentrate more of the available sensor area on the region that actually determines inspection success. The right lens is therefore the one that provides sufficient hidden-feature contrast and spatial sampling while retaining workable camera distance and production positioning margin.

The Kyptec Automation® SWIR Camera Lens collection gives OEMs a particularly useful foundation for this engineering process because its five focal lengths share a specialized 900–1700 nm, 2 MP, 2/3-inch, F1.4 and C-Mount platform. Rather than forcing every semiconductor inspection task into one optical geometry, machine builders can choose the focal length around wafer size, hidden-feature dimensions, available stand-off and required sensor utilization.

For reliable OEM implementation, the decisive principle is to qualify the complete optical path using the real wafer and real defect at the actual production wavelength. When spectral transmission, silicon thickness, illumination geometry, focal length, field of view, focus and object-side sampling are designed together, a dedicated Kyptec Automation® SWIR Camera Lens can become a powerful component in nondestructive semiconductor inspection—revealing wafer, material and hidden-feature information that conventional surface imaging cannot reliably access.