Norbert Schuster

Passive athermalization of two-lens designs in 8-12micron waveband

Passive athermalization has become a key-technology for automotive and other outdoor applications using modern uncooled 25 and 17 micron bolometer arrays. For high volume applications, passive athermalized optical designs with only two lenses reduce costs. A two lens solution requires a careful choice of lens and housing materials. A first order approach to thermal drift uses the RAYLEIGH criteria for depth of focus. It can be seen that narrow field of view lenses are the most sensitive to defocus with temperature. The different methods used to achieve stable performance over the required Temperature Range can be compared, namely passive optical athermalization and passive mechanical athermalization. GASIR® possesses inherent properties enabling optical passive athermalization. High resolution, two element designs for different field angles are presented. Each lens category is present: Super Wide Angle, Wide Angle, Standard, Tele and Super Tele. All examples are designed for 17micron VGA-detectors. These designs use aspheres and diffractive structures. The impact of temperature on all these parameters can only be determined by ray tracing. The proposed metric is the average of the tangential and sagittal MTF versus image height at Nyquist frequency. A very nonlinear impact of temperature on MTFA at different image heights is clearly visible. Examples are shown. An MTF based criteria for judging athermalization is proposed. It contains two values: the admissible MTF-drop ▵MTF in % and the resulting Temperature Range ▵T in Kelvin. The procedure to get these values is demonstrated. Values of 9 lens assemblies are listed. A comparison with results of first order approach shows limitations of this approach. A general quantification of athermalization is proposed. The pair of values (▵MTF, ▵T) is independent of other lens indexes. The limitations of this method are discussed.

Challenges, constraints and results of lens design in 8-12micron waveband for bolometer-FPAs having a pixel pitch 12micron

In the 8-12 micron waveband Focal Plane Arrays (FPA) are available with a pixel pitch of 12 microns or less. High resolution FPAs with VGA, XGA and SXGA resolution should become available at a reasonable price. These will require new lens designs to give the required fields of view. The challenge for the Optical Designer is to design lenses when the pixel pitch of the detector is the same as the wavelength of the light imaged. The lens specification will need to give more thought to the resolution required by the system. A smaller pixel pitch detector defines a requirement for a shorter focal length to give the same field of view. This will have a number of effects upon the lens design. Geometrical aberrations decrease proportionally with the focal length. Reverse telephoto layouts will become more common, particularly when the system has a shutter. The increase in pixel count will require wide field of view lenses which present particular challenges. The impact of diffraction effects on the lens design is considerably increased. The fast F-number causes an increase in the diffraction limit of the system, but also increases geometric aberrations by a cube law. Therefore the balance between the diffraction limited and the aberration limited performance becomes more difficult. The first approach of the designer is to re-use proven designs originally intended for use with 17micron detectors. Some of these designs will have adequate performance at the Nyquist limit of the 12 micron detectors. Even smaller detector pitches, such as 10 micron, will demand new approaches to Infra Red lens design. The traditional approach will quickly increase the number of elements to 3 or even more. This could lead to the lenses with medium fields of view driving the system cost. A close cooperation between the camera developer and lens designer will become necessary in order to explore alternate approaches, such as wavefront coding, in order to reach the most cost effective solution.

Two-lens designs for modern uncooled and cooled IR imaging devices

In recent years, thermal detectors with a 17 μm pixel pitch have become well-established for use in various applications, such as thermal imaging in cars. This has allowed the civilian infrared market to steadily mature. The main cost for these lens designs comes from the number of lenses used. The development of thermal detectors, which are less sensitive than quantum detectors, has compelled camera manufacturers to demand very fast F-numbers such as f/1.2 or faster. This also minimizes the impact of diffraction in the 8-12 μmm waveband. The freedom afforded by the choice of the stop position in these designs has been used to create high-resolution lenses that operate near the diffraction limit. Based on GASIR®1, a chalcogenide glass, two-lens designs have been developed for all pixel counts and fields of view. Additionally, all these designs have been passively athermalized, either optically or mechanically. Lenses for cooled quantum detectors have a defined stop position called the cold stop (CS) near the FPA-plane. The solid angle defined by the CS fixes not only the F-number (which is less fast than for thermal detectors), but determines also the required resolution. The main cost driver of these designs is the lens diameter. Lenses must be sufficiently large to avoid any vignetting of ray bundles intended to reach the cooled detector. This paper studies the transfer of approved lens design principles for thermal detectors to lenses for cooled quantum detectors with CS for same pixel count at three horizontal fields of view: a 28° medium field lens, an 8° narrow field lens, and a 90° wide field lens. The lens arrangements found for each category have similar lens costs.

Quantify Passive Athermalization in Infrared Imaging Lens Systems

Passive athermalization has become a key-technology for automotive and other outdoor applications using modern uncooled 25 and 17 micron bolometer arrays. For high volume applications, passively athermalized optical designs with a minimum of lenses reduce costs and require a careful choice of lens and housing materials. But, up to now, metrics of athermal properties of these lenses are seldom published. Metrics for athermalization are mentioned in two categories: MTF-based to describe application limits under environmental conditions, and first order relations which are helpful in the optical and mechanical design process. Correlation between both categories is analyzed on several GASIR®-lens designs. The allowable degradation of MTF in the Temperature Range depends on the lens application. The MTF-approach proposed to quantify passive athermalization considers different metrics: Several Through-Focus-MTF-graphs at interesting temperatures for optical design, the MTF-versus-field-graph at interesting temperatures offers the complete customer information; the On-Axis-MTF versus temperature shows the typical thermal drift. The most effective way to describe the athermalization status is the value pair of Temperature Range and of percentage in MTF-loss for on-axis point. This pair of values is applicable for all IR-imaging lenses, closely related to lens application, and independent of the camera detector. First order relations identify the most critical influences on athermalization. Different lens materials are discussed whereby the achromatic correction by diffractive structures reduces also the effective Thermal Glass Constant. GASIR® possesses inherent passive athermalization properties. Known first order relations are expanded to two group lens systems. This new relation gives a good overview on where the most effective place for the PMA-mechanism in the lens assembly is and how to arrange it. A narrow field of view example shows different kinds of movement: first group only, second group only and both groups together. It will be seen that the shortest compensation mechanism depends on power and distance of groups.

Challenges, constraints, and results of lens design for 17 micron-bolometer focal plane arrays in 8-12 micron waveband

In the 8-12 micron waveband Focal Plane Arrays (FPA) are available with a 17 micron pixel pitch in different arrays sizes (e.g. 512 x 480 pixels and 320 x 240 pixels) and with excellent electrical properties. Many applications become possible using this new type of IR-detector which will become the future standard in uncooled technology. Lenses with an f-number faster than f/1.5 minimize the diffraction impact on the spatial resolution and guarantee a high thermal resolution for uncooled cameras. Both effects will be quantified. The distinction between Traditional f-number (TF) and Radiometric f-number (RF) is discussed. Lenses with different focal lengths are required for applications in a variety of markets. They are classified by their Horizontal field of view (HFOV). Respecting the requirements for high volume markets, several two lens solutions will be discussed. A commonly accepted parameter of spatial resolution is the Modulation Transfer Function (MTF)-value at the Nyquist frequency of the detector (here 30cy/mm). This parameter of resolution will be presented versus field of view. Wide Angle and Super Wide Angle lenses are susceptible to low relative illumination in the corner of the detector. Measures to reduce this drop to an acceptable value are presented.

Using material advances in chalcogenide glasses to improve imaging lenses in the 8-14 μm waveband

Changes in the position of best focus over temperature are a major source of contrast degradation in the long-wave infrared. The prime sources of this focus shift are the difference between thermal expansion coefficients of lens material and housing material, and the change in refractive index over temperature ∂n/∂T. These parameters, combined with the limited depth of focus when using lenses for uncooled detectors, can rapidly degrade performance with changing temperature. Firstorder paraxial calculations to model these changes are discussed, with a demonstration of its application to single-element imaging systems. The model is then expanded to include two-element systems where both elements are made of the same optical material, or the more general case where different materials are combined. It is shown how a chalcogenide glasses are well suited for athermalization, and how a combination of material choice and optical prescription can lead to an improved passive optical athermalization scheme, i.e. stable performance over temperature with no moving components. The limits of the used model are discussed and examples given for various focal lengths.

Evaluate depth of field limits of fixed focus lens arrangements in thermal infrared

More and more modern thermal imaging systems use uncooled detectors. High volume applications work with detectors that have a reduced pixel count (typically between 200x150 and 640x480). This reduces the usefulness of modern image treatment procedures such as wave front coding. On the other hand, uncooled detectors demand lenses with fast fnumbers, near f/1.0, which reduces the expected Depth of Field (DoF). What are the limits on resolution if the target changes distance to the camera system? The desire to implement lens arrangements without a focusing mechanism demands a deeper quantification of the DoF problem. A new approach avoids the classic “accepted image blur circle” and quantifies the expected DoF by the Through Focus MTF of the lens. This function is defined for a certain spatial frequency that provides a straightforward relation to the pixel pitch of imaging device. A certain minimum MTF-level is necessary so that the complete thermal imaging system can realize its basic functions, such as recognition or detection of specified targets. Very often, this technical tradeoff is approved with a certain lens. But what is the impact of changing the lens for one with a different focal length? Narrow field lenses, which give more details of targets in longer distances, tighten the DoF problem. A first orientation is given by the hyperfocal distance. It depends in a square relation on the focal length and in a linear relation on the through focus MTF of the lens. The analysis of these relations shows the contradicting requirements between higher thermal and spatial resolution, faster f-number and desired DoF. Furthermore, the hyperfocal distance defines the DoF-borders. Their relation between is such as the first order imaging formulas. A calculation methodology will be presented to transfer DoF-results from an approved combination lens and camera to another lens in combination with the initial camera. Necessary input for this prediction is the accepted DoF of the initial combination and the through focus MTFs of both lenses. The accepted DoF of the initial combination defines an application and camera related MTF-level, which must be provided also by the new lens. Examples are provided. The formula of the Diffraction-Limited-Through-Focus-MTF (DLTF) quantifies the physical limit and works without any ray trace. This relation respects the pixel pitch, the waveband and the aperture based f-number, but is independent of detector size. The DLTF has a steeper slope than the ray traced Through-Focus-MTF; its maximum is the diffraction limit. The DLTF predicts the DoF-relations quite precisely. Differences to ray trace results are discussed. Last calculations with modern detectors show that a static chosen MTF-level doesn’t reflect the reality for the DoFproblem. The MTF-level to respect depends on application, pixel pitch, IR-camera and image treatment. A value of 0.250 at the detector Nyquist frequency seems to be a reasonable starting point for uncooled FPAs with 17μm pixel pitch.

Depth of field in modern thermal imaging

Modern thermal imaging lenses for uncooled detectors are high aperture systems. Very often, their aperture based fnumber is faster than 1.2. The impact of this on the depth of field is dramatic, especially for narrow field lenses. The users would like to know how the image quality changes with and without refocusing for objects at different distances from the camera core. The Depth of Field approach presented here is based on the lens specific Through Focus MTF. It will be averaged for the detector area. The lens specific Through Focus MTF will be determined at the detector Nyquist frequency, which is defined by the pixel pitch. In this way, the specific lens and the specific FPA-geometry (pixel pitch, detector area) are considered. The condition, that the Through Focus MTF at full Nyquist must be higher than 0.25, defines a certain symmetrical depth of focus. This criterion provides a good discrimination for reasonable lens/detector combinations. The examples chosen reflect the actual development of uncooled camera cores. The symmetrical depth of focus is transferred to object space using paraxial relations. This defines a typical depth of field diagram containing three functions: Hyperfocal distance, nearest and furthest distance versus sharp distance (best focus). Pictures taken with an IR Camera illustrate the effect in the depth of field and its dependence on focal length. These pictures confirm the methodology. A separate problem is the acceptable drop of resolution in combination with a specific camera core and specific object scenes. We propose to evaluate the MTF-graph at half Nyquist frequency. This quantifies the resolution loss without refocus in accordance with the IR-picture degradation at the limits of the Depth of Field. The approach is applied to different commercially available lenses. Pictures illustrate the Depth of Field for different pixel pitches and pixel counts.

An alternative approach to depth of field which avoids the blur circle and uses the pixel pitch

Modern thermal imaging systems apply more and more uncooled detectors. High volume applications work with detectors which have a reduced pixel count (typical between 200x150 and 640x480). This shrinks the application of modern image treatment procedures like wave front coding. On the other hand side, uncooled detectors demand lenses with fast F-numbers near 1.0. Which are the limits on resolution if the target to analyze changes its distance to the camera system? The aim to implement lens arrangements without any focusing mechanism demands a deeper quantification of the Depth of Field problem. The proposed Depth of Field approach avoids the classic “accepted image blur circle”. It bases on a camera specific depth of focus which is transformed in the object space by paraxial relations. The traditional RAYLEIGH’s -criterion bases on the unaberrated Point Spread Function and delivers a first order relation for the depth of focus. Hence, neither the actual lens resolution neither the detector impact is considered. The camera specific depth of focus respects a lot of camera properties: Lens aberrations at actual F-number, detector size and pixel pitch. The through focus MTF is the base of the camera specific depth of focus. It has a nearly symmetric course around the maximum of sharp imaging. The through focus MTF is considered at detector’s Nyquist frequency. The camera specific depth of focus is this the axial distance in front and behind of sharp image plane where the through focus MTF is <0.25. This camera specific depth of focus is transferred in the object space by paraxial relations. It follows a general applicable Depth of Field diagram which could be applied to lenses realizing a lateral magnification range -0.05…0. Easy to handle formulas are provided between hyperfocal distance and the borders of the Depth of Field in dependence on sharp distances. These relations are in line with the classical Depth of Field-theory. Thermal pictures, taken by different IR-camera cores, illustrate the new approach. The quite often requested graph “MTF versus distance” choses the half Nyquist frequency as reference. The paraxial transfer of the through focus MTF in object space distorts the MTF-curve: hard drop at closer distances than sharp distance, smooth drop at further distances. The formula of a general Diffraction-Limited-Through-Focus-MTF (DLTF) is deducted. Arbitrary detector-lens combinations could be discussed. Free variables in this analysis are waveband, aperture based F-number (lens) and pixel pitch (detector). The DLTF- discussion provides physical limits and technical requirements. The detector development with pixel pitches smaller than captured wavelength in the LWIR-region generates a special challenge for optical design.

Methodology for lens transmission measurement in the 8-13 micron waveband: integrating sphere versus camera-based

Transmission is a key parameter in describing an IR-lens, but is also often the subject of controversy. One reason is the misinterpretation of “transmission” in infrared camera practice. If the camera lens is replaced by an alternative one the signal will be affected by two parameters: proportional to the square of the effective aperture based F-number and linearly to the transmission. The measure to collect energy is defined as the Energy Throughput ETP, and the signal level of the IR-camera is proportional to ETP. Most published lens transmission values are based on spectrophotometric measurement of plane-parallel witness pieces obtained from coating processes. Published aperture based F-numbers derive very often from ray tracing values in the on-axis bundle. The following contribution is about transmission measurement. It highlights the bulk absorption and coating issues of infrared lenses. Two different setups are built and tested, an Integrating Sphere (IS)-based setup and a Camera-Based (CB) setup. The comparison of the two principles also clarifies the impact of the F-number. One difficulty in accurately estimating lens transmission lies in measuring the ratio between the signal of ray bundles deviated by the lens under test and the signal of non-deviated ray bundles without lens (100% transmission). There are many sources for errors and deviations in LWIR-region including: background radiation, reflection from “rough” surfaces, and unexpected transmission bands. Care is taken in the set up that measured signals with and without the lens are consistent and reproducible. Reference elements such as uncoated lenses are used for calibration of both setups. When solid angle-based radiometric relationships are included, both setups yield consistent transmission values. Setups and their calibration will be described and test results on commercially available lenses will be published.