Dual field-of-view midwave optical design and athermalization analysis

Chih-Wei Kuo,1,* Chih-Lung Lin,2 and Chien-Yuan Han3 1Electro- Section, Materials and Electro-Optics Research Division, Chung-Shan Institute of Science and Technology, Lung-Tan 325, Taiwan, China 2Department of Electronic Engineering, Hwa Hsia Institute of Technology, Taipei 235, Taiwan, China 3Department of Electro-Optical Engineering, National United University, Miao-Li 360, Taiwan, China *Corresponding author: [email protected]

Received 2 April 2010; revised 31 May 2010; accepted 2 June 2010; posted 2 June 2010 (Doc. ID 126354); published 23 June 2010

A step-zoom and reimaging structure were utilized to construct a dual field-of-view optical design for high-magnification switching in the 3–5 μm spectral band. The design has a flexible optomechanical lay- out, which means it can be utilized for multipurpose applications. The effects of the surrounding envir- onmental and axial gradient temperature are analyzed using the concept of thermal resistance, and the thermal compensation is discussed. A description of the zooming mechanism and optomechanical control is offered. © 2010 Optical Society of America OCIS codes: 110.3080, 220.3620.

1. Introduction xial analysis of mechanically compensated zoom The applications of infrared dual field-of-view can be expressed in terms of Gaussian brackets (DFOV) optical systems have been widely discussed [9,10]. The focus of the zoom can be solved by [1–5] for both the midwave infrared (MWIR) spectrum describing the relationships among focal length, lens (3–5 μm) and longwave infrared spectrum (8–12 μm). position, ray height, and direction in relation to the These optical systems utilize the wide-angle field of matrix. The zoom process can be expressed by a view (WFOV) to search the scenery and the narrow unified varifocal differential equation with a stable angle (NFOV) mode to identify the target of interest image plane being the constraint condition [11]. Ther- close up. The algorithm for assessing a surveillance mal imaging systems for target-tracking purposes, image is derived from the Johnson criteria [6] and such as rocket and missile launching applications, then used to deduce the target being detected, recog- require WFOV systems for monitoring and NFOV nized, or identified [7]. Variations in the characteris- systems for identification. DFOV infrared optical sys- tems should have FOV switching without target loss. tic size of a particular object, and the distance from They can be classified as a subset of continuous-zoom the sensor to the target plane, both result in the dif- structures that utilize only axial steps to move lenses ferent magnification ratio requirements obtained by between two extreme magnifications. A nonaxial the computation of the Johnson criteria. Previously, moving DFOV variant, called the rotating-in scheme, the Delano diagram has been used to aid in the design was developed by combining reverse-telephoto and of zoom lenses [8], composed using the height of the telephoto structures with/without two separate lens paraxial marginal and chief rays as the longitudinal groups controlled by a rotation mechanism. This is and transverse coordinates, respectively. The para- distinct from the step-zoom scheme [12]. Most infrared semiconductor detectors are cryo- 0003-6935/10/193691-10$15.00/0 genically cooled and assembled in a thermally insu- © 2010 Optical Society of America lated Dewar flask. This is necessary to achieve the

1 July 2010 / Vol. 49, No. 19 / APPLIED OPTICS 3691 maximum signal-to-noise ratio and avoid anomalous ble system. The methodology and process for stabiliz- images. A special baffle, called a cold stop, located in- ing the optical qualities by temperature compensa- side the Dewar flask, causes the light cone to strike tion is called athermalization. Active athermalization the focal plane array (FPA), and the FPA records the is appropriate for quick responses to overcome rapid objective space energy exclusively. It has been deter- error inputs while operating in a thermally dynamic mined that 100% cold-stop efficiency is required [13]. environment [17]. To facilitate athermalization, a mo- This can be accomplished by the aperture stop and torized lens adjusts the position and thermal sensors cold stop coinciding. The optical system is still opera- measure the temperature difference. The lens and tional if the effective focal length (EFL) is small. the sensor are controlled by a programmed micro However, to identify or classify a distant but small processor. target, a longer EFL is essential. The layout of the An advantageous feature of the DFOV infrared op- aperture stop inside the Dewar flask will make the tical system is the incorporation of motorized me- diameter of the front end of the lens larger than chanical parts that allow it to travel quickly to the entrance pupil, because of the off-axis light cone compensate for the variation of the optical group over accommodation. The only way to shrink the aperture an exact distance. In other words, the optomechani- size of the front end of the lens while maintaining cal design can ensure the functioning of the optical 100% cold-stop efficiency and offering a longer EFL design. The mechanical layout of the zoom system is is through reimaging. The extra relay lens creates an adapted to meet overall system requirements. Inno- image of the aperture stop located on the front end vations and inventions have led to the development lens that is equal in diameter to the entrance pupil. of numerous patents. For example, two coaxial lens The reimage element is the primary objective, fol- sleeves have been assembled together. With this type lowed by the intermediate focal plane, relay lens, of device, the translation of optical elements is en- cold/aperture stop, and FPA. abled by the relative rotation of the sleeves. The Stray light caused by double in the visi- cam grooves provide a predetermined locus [18]. The ble lens can create ghost images, especially in a scene barrel is made more compact and portable when not with a signal difference source against a dark back- in operation by the movement of a thrust member ground. A high transmission rate of antireflection back and forth in the optical-axis direction [19]. coating in the IR spectrum is harder than in the visi- The diameter of the whole barrel can be reduced ble one; hence, an IR system is more vulnerable to by replacing the coaxial sleeve layout with an off-axis double-reflection stray light. Besides, the cryogeni- cam attached to the lens rim [20]. The traditional cally cooled detector acts as a strong light in a dark cam groove located on the cylindrical surface is re- scene because of its low temperature compared with placed by the plane surface [21]. The cam provides the warmer surroundings of the lens barrel or the ob- a reliable lens walkway but the mechanical manufac- served target. Therefore, an infrared FPA may reflect turing process is expansive and complex. Thus, the a single reflection of its own image. This is called the lens movement path can be controlled by a pro- narcissus effect [14,15], and this retroreflected non- grammed motor utilizing a spur gear to drive the uniformity thermal image reduces the contrast of a feed bar [22]. The optomechanical alignment must dim object. In a scanning system, the narcissus effect ensure that the line of sight (LOS) remains unaf- is noticed as a retroreflected ghost image that moves fected during the change in the FOV, especially for on the focal plane as the scanning mirror rotates. The automatic target-tracking applications. The methods starring array system can negate the narcissus effect for LOS stabilization commonly used to counteract if the optical system meets the requirements of non- jitter introduced from the optical system platform uniformity calibration, a constant surrounding tem- and surroundings can be categorized in three types: software, platform, and steering [23]. perature, and no lens movement. Otherwise, the narcissus effect is monitored by an optical design process. 2. Lens System Scheme Temperature change in the operating environment A SELEX HgCdTe FPA sensor (F=4, 384 × 488 pixels can result in variation of the physical properties of an and 20 μm square pixel size) was used in this study. optical assembly (i.e., lens thickness, air space, refrac- The optical design offers a dual EFL of 20=250 mm. tive index, structural dimensions) [16]. These effects This specification can improve the broad scene while can cause the system to lose focus or elements to be- still giving high enough resolution for most infrared come misaligned. Temperature gradients may cause optical system requirements. This is a reimaging type tilting in the axial or radial directions, reversing the system that allows for reduction of the lens aperture effects of homogenization on the materials and dis- and for compact instrument space. To ensure image torting the optical surface. The quality of the final im- quality, we must strictly consider a better minimum age can become degraded as the adverse temperature resolvable temperature difference (MRTD). The poly- effects increase. Most materials for infrared refrac- chromatic modulation transfer function (MTF) within tive lenses have high rates of index variation with the 3–5 μm spectrum approaches the diffraction limit temperature, which can result in rapid focus shifts. beneath the Nyquist spatial frequency. Degradation from thermal variations in an infrared In a Gaussian design, a primary objective is imple- system is an order of magnitude higher than in a visi- mented to generate the system optical power and to

3692 APPLIED OPTICS / Vol. 49, No. 19 / 1 July 2010 focus on the intermediate focal plane under infinite conjugate conditions. This is reimaged by a relay lens to the final focal plane (i.e., the cryogenic detector ar- ray). This layout must ensure that the system aper- ture stop coincides with the Dewar’s cold stop for 100% cold-shading efficiency. The NFOV mode is re- quired for the layout of the telephoto objective lens system. The unit power relay lens helps make for eco- nomical saving of the axial length. Adjustment of both the focal length of the telephoto objective sys- tem and the relay lens allows the entrance pupil to be located on the primary objective, which can Fig. 1. (Color online) Gaussian arrangement. minimize the primary lens diameter. Wandering of the pupil, caused by off-axis beam accommodation, can be eliminated [12,13]. These calculations were f system ¼ðϕ1 þ ϕ2 þ ϕ3 − t1ϕ1ϕ2 − t1ϕ1ϕ3 − t2ϕ1ϕ3 carried out with a Gaussian imaging equation and − t ϕ ϕ þ t t ϕ ϕ ϕ Þ−1 m ; ð Þ paraxial yu ray tracing: 2 2 3 1 2 1 2 3 × relay 3 0 0 njuj ¼ njuj − yjϕj; ð1Þ where f system is the EFL of the optical system, and mrelay is relay lens magnification. Theoretically, 0 switching the EFL can be accomplished by move- yjþ1 ¼ yj þ tju ; ð2Þ j ment of the varying and compensating groups. where u is the tangent value of the angle between the optical axis and the ray pencil; y is the contact height 3. Optimization of Image Quality of the ray; t is the thickness of the air spacing; and ϕ Many Gaussian solutions, including optical power is the optical power of the lens. The subscript symbol and air space, can be found before aberrations and fea- represents the optical surface number, and the super- sibility are taken into consideration. The reasonable script prime represents the ray after . One initial arrangement of optical power and air space for component of the lenses composing the telephoto ob- each group and the yu ray-tracing data are listed in jective system is able to move to change the magni- Table 1. In NFOV mode, the first lens diameter is fication and the other component to eliminate the equal to the entrance pupil, and the wandering of focal shift. Consequently, we were able to construct the pupil is not existent. The material properties a WFOV with a reverse-telephoto layout. This DFOV and shape factors of the lens are set accordingly. Sili- optical system was comprised of four sequential con is used primarily for its high , groups: the focusing group, variation group, compen- which is advantageous with respect to aberration con- sation group, and relay group. The lens power of each trol. Germanium and silicon are applied to create an group was positive, negative, positive, and positive, achromatic air-spaced doublet, because their Abbe respectively [24]. The Gaussian arrangement is numbers have great difference. The total monochro- shown in Fig. 1. The EFL of the optical system can matic Seidel aberrations and lateral/axial chromatic be determined by the following equation: aberrations in lenses were evaluated. This evaluation

Table 1. NFOV and WFOV Paraxial Data

NFOV Paraxial Data Surf Type Power (mm−1) Thickness (mm) Y Marginal (mm) U Marginal OBJ Standard Infinity 0.000 0.000 1 Paraxial 0.004 170.0 31.250 −0:125 2 Paraxial −0:025 80.0 10.000 0.125 3 Paraxial 0.0125 160.0 20.000 −0:125 4 Standard 80.0 0.000 −0:125 5 Paraxial 0.025 80.0 −10:000 0.125 IMA Standard 0.000 0.125

WFOV Paraxial Data Surf Type Power (mm−1) Thickness (mm) Y Marginal (mm) U Marginal OBJ Standard Infinity 0.000 0.000 STO Paraxial 0.004 38.57 2.500 −0:010 2 Paraxial −0:025 264.00 2.114 0.043 3 Paraxial 0.0125 107.43 13.429 −0:125 4 Standard 80.00 0.000 −0:125 5 Paraxial 0.025 80.00 −10:000 0.125 IMA Standard 0.000 0.125

1 July 2010 / Vol. 49, No. 19 / APPLIED OPTICS 3693 was conducted with the ZEMAX commercial software, after the variables, zoom positions, and system con- straints were assigned. The weighting functions of the FOV were set equally at the on axis, 0.5, 0.707, 0.866, and at the full FOV. The MWIR spectra (3, 4, and 5 μm), were distributed as the same weighting functions. The location of the Dewar cold stop was de- fined by the system aperture. The layouts of the opti- cal system and aperture/cold stop position are shown in Fig. 2. All lenses are housed in one barrel exclu- sively, and the other elements (window, filter, stop, and FPA) are assembled as a radiometry detector unit. The requirements of optical manufacturing and mechanical assembly were translated into the pro- gram’s merit function editor. The ZEMAX software utilized the damped least-square algorithm to mini- mize the merit function [25]: X 2 MF ¼ WjðVj − TjÞ ; ð4Þ j where W is the weighting function; V is the current value; and T is the target value. The subscript symbol represents the operand item number. As a starting point for optimization, the rms spot radius is selected as a merit function. During the final stage of optimi- zation, wavefront error was selected as system perfor- mance approached the diffraction limits. The ray fans are shown in Figs. 3(a) and 3(b). Backward ray tracing from the cryogenically cooled detector can determine the surface of the contributing retroreflection ghost. However, this procedure is time

Fig. 3. (Color online) (a) NFOV, ray fan plot. (b) WFOV, ray fan plot.

consuming and labor intensive. It is more efficient and of greater benefit to simultaneously design an op- tical lens to control the narcissus effect. The ratio of the detector area (A) multiplied by the average emis- sivity (ε) to the area of the ghost defines the narcissus intensity of a single surface, and this ratio is also named as cold return (CR ¼ εA=ðπ½4yniðF=#Þ2Þ, de- tail equation and figure referring [14]). Summing all surface ratios, CR represents a system’s narcissus. The product of three geometric optical parameters, yni, is proportional to the ghost area. Here, F/# is the numerical aperture, n is the refractive index, and i is the reflective marginal ray angle of the inci- dence surface. If the value of yni was larger, the inten- sity of the narcissus was lower. A narcissus intensity for the whole detector area and ghost variations across every pixel can be effi- ciently controlled by constraining the absolute values of yni (Table 2) using an optical design program. How- ever, these restrictions confined the development of the lens shape factor. This could be a problem for the required diffraction-limited performances. There- Fig. 2. (Color online) (a) NFOV, EFL ¼ 250 mm. (b) WFOV, fore, a trade-off is needed. In addition, a lens with a EFL ¼ 20 mm. (c) Position of aperture/cold stop. high-efficiency antireflection thin film coating (the

3694 APPLIED OPTICS / Vol. 49, No. 19 / 1 July 2010 Table 2. Absolute YNI Value technique, used in both serial and reentrant layouts Surf YNI (NFOV) YNI (WFOV) so that the thermal variation of the barrel matches the optics [26]. In infrared spectral imaging applica- 1 4.55 0.03 tions, thermal defocusing is almost larger than the 2 4.92 0.03 barrel expansion allows. Therefore, appropriate me- 3 4.97 0.03 4 1.02 0.01 tal selection is vital. However, passive compensation 5 0.58 0.09 is only suitable for steady temperature states. In a 6 0.15 0.08 thermally dynamic environment, a motorized com- 7 0.02 1.21 pensation lens can be used to slightly displace the 8 0.11 1.49 original position. Compensation is calculated with 9 0.09 1.58 a specific algorithm, which can satisfy both steady 10 0.05 0.40 and transient thermal states. This is called active 11 0.05 0.30 compensation. 12 0.23 0.23 Most infrared optical system applications require 13 0.94 0.94 higher target resolutions; therefore, a longer EFL is 14 0.96 0.97 15 0.31 0.31 essential. The total length of the infrared optical as- 16 0.80 0.79 sembly is obviously longer than that of most commer- 17 0.05 0.05 cial visible systems, even those adopting the telephoto 18 0.33 0.33 structure. The temperature difference between the 19 0.32 0.32 optical system’s front and bottom sides cannot be ne- 20 0.30 0.30 glected in an infrared lens integrated into a gimbals 21 0.30 0.29 pod or vehicle turret, where only the first lens surface is exposed to the outside environment. The axial tem- perature gradient must be taken into consideration in ghost strength depending on the residual reflection at cases where the optical system suffers from variation the offending lens surface) and sharp edge spectrum in environmental temperature. Thermal effects not bandpass filter inside the Dewar flask are helpful to only arise from shifts in the soaking temperature, relax this anomalous image. but also from the existence of the axial temperature gradient. One-dimensional thermal resistance analy- 4. Athermalization Analysis sis can be utilized [27] to evaluate axial temperature distribution: The material properties and the characteristic config- urations of optical elements and mechanical parts can lj change with variations in the surrounding tempera- Rj ¼ ; ð5Þ ture. A methodology must be developed to compen- Ajkj sate for the optical system performance and to recover the degradation of the image quality. Most ΔT H ¼ ; ð Þ lenses are made of a brittle material, so the surface ΣR 6 radius departs from the original design value. How- j ever, the major problem caused by temperature var- where R is the thermal resistance, l is the character- iation is the term for differentiating the refractive istic length, A is the heat flux crossing area, k is the index with temperature (dn=dT). This is particularly thermal conductivity, ΔT is the temperature differ- true in an infrared system but less so in a visible sys- ence, and H is the heat flux. The subscript symbol re- tem. Reference [17] lists the dn=dT of most visible op- presents the item number of the thermal resistor. In −6 tical materials as falling between 3 × 10 =°C and this study,we employed a silicon (monocrystalline) al- −6 −6 10 × 10 =°C (such as BK7, dn=dT ¼ 3:6 × 10 =°C), loy for the lens and a germanium (monocrystalline) and that of infrared materials as between 30 × and aluminum alloy (No. 6061) for the barrel. The −6 −6 10 =°C and 400 × 10 =°C (such as germanium, thermal conductivities of these materials are shown −6 dn=dT ¼ 396 × 10 =°C). Generally, the difference of in Table 3 (polycrystalline and the other aluminum dn=dT in infrared materials is an order of magnitude alloy numbers will be different tabulated values). higher than in visible materials. Therefore, it is neces- The thermal resistance diagram is shown in Fig. 4. sary to consider thermal compensation in infrared The axial heat flux is primarily conducted by the bar- spectrum optical designs. rel parts because the thermal resistance of air is sev- There are two distinct thermal compensation eral orders higher than that of aluminum alloys, methodologies: passive and active. In most visible germanium, or silicon. The axial systems, as the temperature rises, the elongation of the barrel length is greater than the growth of the EFL. The simplest way to compensate for this is to Table 3. Material’s Thermal Conductivity choose a mechanical material that allows the ther- Aluminum Germanium Silicon Air mal expansion of the barrel to be close to the optical Thermal Conductivity 180 59 163 0.026 thermal defocusing. This is considered passive com- (W/mK) pensation. Another passive technique is the bi-metal

1 July 2010 / Vol. 49, No. 19 / APPLIED OPTICS 3695 function. Finally, these data are input to the optical software for ray tracing for the whole optical system. The performance of the optical system given the axial temperature gradient can be analyzed. Fig. 4. Thermal resistance diagram. In the present study, the reference temperature is always 20 °C, and the other soaking temperature va- of the barrel can be determined by calculating the ax- lues are compared with this one. In the beginning, the ial temperature distribution with the following equa- MTF curves for the soaking temperature (0 °C, 20 °C, tion: 40 °C, and 60 °C) are plotted in Figs. 5(a) and 5(b). The off-axis curves smoothly degrade as temperature de- T − T parts from 20 °C, for the thermal variations of optical TðzÞ¼T þ max min z; ð7Þ min L and mechanical materials are getting larger as the temperature changes. The on-axis MTF is discussed, where TðzÞ is the axial temperature distribution, Tmin because the optical axis is boresighted for aiming. and Tmax are the temperature gradients on both sides, There is little thermal effect to the on-axis MTF L is the length of the barrel, and z is the coordinate for curves [Figs. 5(c) and 5(d)] even at higher frequency. the axial direction. Moreover, each lens rim is This could be the paraxial results being affected mounted between the cell and the retainer. The air slightly by the soaking temperature. The on-axis surrounding the lens provides better thermal insula- curve of 40 °C seems a little bit higher than the others, tion. Therefore, the lens radial temperature distribu- for weighting functions of all FOVare the same during ’ tion is assumed to be uniform and equal to the cell s optimization. However, the tangential and sagittal temperature. For the thermal setup, each lens surface curves spread out as the temperature ranges farther radius, thickness, and index at a specific temperature away from 20 °C. The distortion and field curvature can be calculated using ZEMAX’s multiconfiguration

Fig. 5. (Color online) (a) NFOV and MTF results at 0 °C, 20 °C, 40 °C, and 60 °C soaking temperature. (b) WFOV and MTF results at 0 °C, 20 °C, 40 °C, and 60 °C soaking temperature. (c) NFOV and MTF on-axis results for 0 °C, 20 °C, 40 °C, and 60 °C soaking temperature. (d) WFOV and MTF on-axis results 0 °C, 20 °C, 40 °C and 60 °C soaking temperature.

3696 APPLIED OPTICS / Vol. 49, No. 19 / 1 July 2010 depending on FOV show insignificant changes at dif- act the MTF degradation introduced by the axial ferent soak . temperature distribution, as shown in Figs. 7. These The temperature gradient is defined by the tem- curves exhibit the athermalized benefit of the com- perature at the front side of the barrel minus the tem- pensating group. Although efficient for the WFOV perature at the bottom of the barrel. The lens barrel mode, the NFOV mode can still tolerate some resi- bottom side temperature is always 20 °C [as Fig. 2(c) dual unsolved thermal defocusing. Because of the indicates, the window, filter, stop, and cryogenic FPA chromatic aberration in the NFOV mode, the ther- are evacuative as a stand-along unit, which is sepa- mal defocus cannot be eliminated completely by the rated from the lens barrel by air space.). The MTF motion of the compensation group. The curves of on- curves for axial gradient temperatures of −20 °C, axis polychromatic rms spot radius plotted against þ20 °C, and þ40 °C are shown in Figs. 6. Thermal gradient temperature are shown in Fig. 8. It can be degradation was manifest in both the NFOV mode seen that the polychromatic spot radii in the NFOV and the WFOV mode, which was especially sensitive mode were larger than those in the WFOV mode. It to the existence of the axial temperature gradient. indicates that the NFOV mode suffered temperature The variations of distortion and field curvature de- gradient effect more than the WFOV mode. During pending on FOV are negligible as gradient tempera- this EFL 20=250 DFOV application, the WFOV helps ture alterations. to search one object (assuming its image spatial fre- As discussed above, this optical system is consti- quency f image), and the NFOV identifies the same ob- tuted of four groups, one of which is the compensat- ject (image spatial frequency being f image divided by ing group, counterbalancing any defocus originating the zoom ratio, 12.5). This optical design can offer a from variation group. Consequentially, the neutrali- better modulation at f image=12:5 in NFOV than that zation of thermal defocusing can be analyzed. The at f image in WFOV (cf. Fig. 7), even suffering the ad- displacement of the compensating group can counter- verse temperature changing effects. The chromatic

Fig. 6. (Color online) (a) NFOV and MTF results at −20 °C, þ20 °C, and þ40 °C temperature gradient. (b) WFOV and MTF results at −20 °C, þ20 °C, and þ40 °C temperature gradient. (c) NFOV and MTF on-axis results at −20 °C, þ20 °C, and þ40 °C temperature gradient.

1 July 2010 / Vol. 49, No. 19 / APPLIED OPTICS 3697 Fig. 7. (Color online) (a) NFOV and compensated MTF results at −20 °C, þ20 °C, and þ40 °C temperature gradient. (b) WFOV and com- pensated MTF results at −20 °C, þ20 °C, and þ40 °C temperature gradient. (c) NFOV and compensated on-axis MTF results at −20 °C, þ20 °C, and þ40 °C temperature gradient. (d) WFOV and compensated on-axis MTF results at −20 °C, þ20 °C, and þ40 °C temperature gradient. aberration coming from thermally dependent index temperature gradient could be interpreted by the mi- variation is residual somewhat. However, there are croprocessor. To counteract thermal defocus, the dis- a limited number of lens materials that are used placement of the compensation group was carried out in MWIR spectral bands, and the common solution is a silicon–germanium air-spaced achromatic doub- let. The image quality has been satisfied in most ap- plications.

5. Optomechanical Control and Layout The optomechanical actuation of the varying and compensating groups was planned utilizing two inde- pendent feed bars driven by programmed DC motors [28]. A microprocessor, encoder, and motor, making up a closed-loop feedback control system, were utilized to propel each moving lens group to the desired position. The proportional-integral-derivative (PID) algorithm (which is part of a fully developed and mature technol- ogy) acts as the controller’s mathematical back- ground. An RS232 interface and proper protocols play the communication role. The results show that the field of view can be switched as required. Thermal sensors were attached to the front and Fig. 8. (Color online) Temperature gradient, NFOV, WFOV, and bottom sides of the barrel [29], so that the axial compensated spot size.

3698 APPLIED OPTICS / Vol. 49, No. 19 / 1 July 2010 plane is located between the first three groups and the final group. This type of reimaging can ensure that the entrance pupil matches a narrow field of view to produce a cold stop. The three previous groups execute the field-of-view switching function. Their EFL times the final group’s magnification de- termines the EFL of the whole optical system: 20=250 mm. After optimization, the MTF approaches the diffraction limit. Different soaking temperature and axial gradient temperature situations are analyzed. The concept of thermal resistance is employed to simplify the com- plex calculation. Uniform temperature changes do not lead to deterioration of optical performance when the elongation of the barrel and variation in thermal focus are similar. However, the axial temperature gradient results in a decline of quality of this optical system. To compensate for defocusing while zooming, thermal defocus is compensated for by a shift in the third optical group. To activate motion of the desired optical group for EFL switching, the PID algorithm drives the motor- ized feed bar and the encoder feedback composes a closed loop control system. The microprocessor moni- tors axial temperature distribution based on the thermal sensor for execution of the compensating movement for athermalization. To ensure that this optical design can be applied for multipurposes, the optical path can be folded with an extra planar mir- ror. And the folding mirror can also benefit lens Fig. 9. (Color online) (a) Optomechanical layout L-shape. alignment and LOS stabilization. (b) Optomechanical layout U-shape. The author acknowledges the contributions of col- leagues in the EO Section, Materials and Electro- according to the temperature distribution. And this Optics Research Division, Chung-Shan Institute of procedure was arranged by the automatic control Science and Technology. program stored in the memory of the microprocessor. Consequently, stand-alone athermalization was achieved. References The present design is intended for multipurpose 1. R. E. Aldrich, “Three elements optically compensated two- applications. The optomechanical layout [30–32] position zoom for commercial FLIRs,” Proc. SPIE 2539,87– can be straight, L,orU-shaped [Figs. 9(a) and 9(b)], 107 (1995). given the electro-optical (EO) system’s mounting re- 2. D. W. Anderson, “M1A2 tank commander’s independent ther- quirements for portable finders, vehicle turrets, or mal viewer optics: optics design perspective,” Proc. SPIE gimbal pods. The folding mirror offers optical path 1970, 128–138 (1993). 3. M. Norland and A. Rodland, “Design of high performance IR turning ability but requires careful placement to ” 2269 – avoid reimaging the mirror surface’s defect to the sensor, Proc. SPIE , 462 471 (1994). 4. M. C. del la Fuente, “A compact dual FOV objective for 3–5 μm FPA if it were placed right on the intermediate focal waveband,” Proc. SPIE 3061, 348–355 (1997). plane. The folding mirror provides an extra benefit 5. J. W. Howard and M. S. Garner, “A two positions IR zoom lens for aligning the optical axis and barrel axis during with low F-number and large format,” Proc. SPIE 1690, assembly. It can also be steered to act as a line-of- 129–136 (1992). sight stabilizer to counteract jitter arising from the 6. R. Simmons, P. A. Manning, and T. Chamberlain, “The design EO system’s platform or surrounding environment. of passively athermalised narrow and wide field of view infra- A precision gyroscope, quick response actuator, and red objectives for OBSERVER unmanned air vehicle,” Proc. complex digital signal processing algorithm are SPIE 5612, 236–248 (2004). needed to facilitate state-of-the-art performance. 7. G. C. Holst, Electro-Optical Imaging System Performance, 3rd ed. (SPIE Press, 2003), pp. 386–400. “ 6. Conclusion 8. W. Besenmatter, Designing zoom lenses aided by the Delano diagram,” Proc. SPIE 273, 242–255 (1980). A DFOV infrared optical design is described, and 9. K. Tanaka, “Paraxial analysis of mechanically compensated constructed of positive, negative, positive, and posi- zoom lenses. 1: Four-component type,” Appl. Opt. 21, tive optical power groups. An intermediate focal 2174–2183 (1982).

1 July 2010 / Vol. 49, No. 19 / APPLIED OPTICS 3699 10. K. Tanaka, “Paraxial analysis of mechanically compensated 21. Y. Kayanuma, “Zoom lens device,” U.S. patent 5,140,468 (18 zoom lenses. 2: Generalization of Yamaji Type V,” Appl. August 1992). Opt. 21, 4045–4053 (1982). 22. N. Kaneda, “Lens position control device and optical appara- 11. C. K. Tao, “Design of zoom system by the varifocal differential tus with driving and controlling of first and second lenses,” equation,” Appl. Opt. 31, 2265–2273 (1992). U.S. patent 5,854,711 (29 December 1998). 12. M. N. Akram, “A design study of dual-field-of-view imaging 23. A. Kazemy, S. A. Hosseini, and M. Farrokhi, “Target-based systems for the 3–5 μm waveband utilizing focal-plane ar- line-of sight stabilization in periscopes,” in Proceedings of rays,” J. Opt. A Pure Appl. Opt. 5, 308–322 (2003). the 15th Mediterranean Conference on Control and Automa- 13. M. N. Akram, “Design of a dual field-of-view optical system for tion, T20-005 (Mediterranean Control Association, 2007). infrared focal plane arrays,” Proc. SPIE 4768,10–20 24. R. Hudyma, “Role of diffractive optical elements in the design (2002). of mid-wave infrared zoom lenses,” Proc. SPIE 1970,11–22 14. J. W. Howard and I. R. Abel, “Narcissus: reflections on (1993). retroreflections in thermal imaging systems,” Appl. Opt. 21, 25. ZEMAX, “Optical Design Program, user’s guide” (2008). 3393–3397 (1982). 26. D. Vukobratovich, “Optomechanical systems design,” in The 15. L. M. Scherr and H. J. Orlando, “Narcissus considerations in Infrared and Electro-Optical Systems Handbook, D. L. Shu- optical design for infrared starring arrays,” Proc. SPIE 2864, maker, ed. (SPIE Press, 1993), Vol. 4, Chap. 3. 442–452 (1996). 27. M. N. Ozisik, : a Basic Approach (McGraw-Hill, 16. P. R. Yoder, “Effects of temperature changes on optical compo- 1985), pp. 42–49. nent mountings,” in Mounting Optics in Optical Instruments, 28. T. A. Palmer, “Mechanical aspects of a dual field of view infra- SPIE Monograph Vol. PM134 (SPIE Press, 2002), Chap. 12. red lens,” Proc. SPIE 4444, 315–320 (2001). 17. E. Ford, “Active temperature compensation of an infrared 29. R. E. Fischer and T. U. Kampe, “Actively controlled 5∶1 afocal zoom lens,” Proc. SPIE 3129, 138–143 (1997). zoom attachment for common module FLIR,” Proc. SPIE 1690, 18. S. Kamata, “Arrangement for adjusting a member in a lens 137–152 (1992). barrel for effecting a zooming function zoom lenses,” U.S. pa- 30. G. M. Miller, “Dual field-of-view tracking lens system,” Proc. tent 4,281,907 (4 August 1981). SPIE 4820, 868–877 (2003). 19. S. Kohmoto, “Cam mechanism in zoom lens,” U.S. patent 31. R. L. Sinclair, “High magnification zoom lenses for 3–5 μm ap- 5,043,752 (27 August 1991). plications,” Proc. SPIE 3429, 11–18 (1998). 20. J. G. Capstaff and O. E. Miller, “Photographic objective,” U.S. 32. A. Mann, Infrared Optics and Zoom Lenses, SPIE Tutorial patent 2,165,341 (11 July 1939). Text in Optical Engineering Vol. TT42 (SPIE Press, 1999).

3700 APPLIED OPTICS / Vol. 49, No. 19 / 1 July 2010