Invited Paper

MEMS-Based Uncooled Infrared Bolometer Arrays – A Review
Frank Niklaus
KTH - Royal Institute of Technology, Microsystem Technology Lab, 10044 Stockholm, Sweden and Faun AB, SE-18361, Sweden Tel: +46 8 76 216 73 49, Fax: +46 8 10 08 58,

Christian Vieider
Acreo AB, Electrum 236, 164 40 Kista, Sweden Tel: +46 8 632 78 45, Fax: +46 8 750 54 30,

Henrik Jakobsen
Vestfold University College, Institute for Microsystem Technology, 3103 Tønsberg, Norway Tel: +47 330 37 711, Fax: +47 330 31 03,

Uncooled infrared bolometer arrays have become the technology of choice for low-cost infrared imaging systems used in applications such as thermography, firefighting, driver night vision, security and surveillance. Uncooled infrared bolometer arrays are reaching performance levels which previously only were possible with cooled infrared photon detectors. With a continuously increasing market volume (> 100 000 units per year to date), the cost for uncooled infrared imaging chips are decreasing accordingly. In this paper we give an overview of the historical development of uncooled infrared bolometer technology and present the most important bolometer performance parameters. The different technology concepts, bolometer design approaches and bolometer materials (including vanadium oxide, amorphous silicon, silicon diodes, silicon-germanium and metals) are discussed in detail. This is followed by an analysis of the current state-of-the-art infrared bolometer technologies, the status of the infrared industry and the latest technology trends. Keywords: uncooled infrared bolometer arrays, microbolometer, MEMS, infrared detector, focal plane array, IR, FPA

1. Introduction
The word bolometer originates from the Greek word bolē and means ray-meter [1]. Bolometers are thermal infrared sensors that absorb electromagnetic radiation and thus increase their temperature. The resulting temperature increase is a function of the radiant energy striking the bolometer and is measured with e.g. the thermoelectric, pyroelectric, resistive or other temperature sensing principles. In the context of uncooled infrared imaging technologies, the term “infrared bolometer” usually refers to resistive microbolometers in which the temperature increase is measured by a resistance change. This review paper focuses on resistive microbolometers and does not include other thermal infrared detectors such as e.g. thermocouples [2], pyroelectric [3] and ferroelectric [4] uncooled infrared detectors. In the first bolometer, invented by the American scientist Samuel P. Langley in 1880 [5], a Wheatstone bridge was used along with a galvanometer that produced a deflection proportional to the intensity of radiation for small deflections. A later bolometer [6] consists of four platinum gratings, each of which is made of a series of strips, inserted in the arms of a resistance bridge. A number of other resistive thin film bolometers have been proposed from 1947 to 1980 [7-11]. The first thin film resistive micobolometers were proposed by R. Hartmann [12] and K.C. Liddiard [13, 14] in 1982 and 1984 respectively. Uncooled infrared microbolometers and focal plane array technology developed at Honeywell has been published since the late 80’s [15-22]. A number of books, book chapters and review articles on infrared technology, including uncooled infrared bolometers, have been published in recent years [23-29]. Uncooled infrared bolometers have become the dominating technology for the majority of commercial and military infrared imaging applications. Some of the most common infrared imaging applications are thermography, night vision (military, commercial and automotive),
MEMS/MOEMS Technologies and Applications III, edited by Jung-Chih Chiao, Xuyuan Chen, Zhaoying Zhou, Xinxin Li, Proc. of SPIE Vol. 6836, 68360D, (2007) · 0277-786X/07/$18 · doi: 10.1117/12.755128 Proc. of SPIE Vol. 6836 68360D-1

The infrared spectrum where the infrared transmission is allowed by the atmosphere is in the 3-5 µm wavelength region (mid wave infrared. MWIR) and in the 8-14 µm wavelength region (long wave infrared. Infrared Bolometer Operation and Figures of Merit Infrared radiation is part of the electromagnetic spectrum with wavelengths above the visible spectrum. The emitted radiation of a black body with a temperature of 300 K has an intensity peak at a wavelength of about 10 µm as can be seen in Figure 1b. LWIR) as can be seen in Figure 1a. medical imaging. For the majority of infrared imaging applications the bolometers are optimized to detect radiation in the 8-14 µm wavelength region. predictive maintenance and industrial process control. y•pa CeLt Design Infrared Lens System Pitch Bolometer FPA Infrared Rays (a) (b) Figure 2: (a) Schematic drawing of a monolithically integrated infrared bolometer focal plane array (FPA) [32] and (b) placement of a focal plane array with an infrared lens system. of SPIE Vol. depending on the emissivity of the surface of the object. a temperature increase of 1 K in the object typically results in a temperature increase in the bolometer membrane on the order of 4 mK [25]. 4x0 lx 10'' Wv 2>110 I >110>1 'a (a) +4+4+++++ H±O Wavelength (microns) ca° caa 4 H±O Ca +4 4 Ca1 ID Absorbino Molecule 20 30 40 30 (b) Wavelength (Din) Figure 1: (a) Plot of atmospheric transmittance in part of the infrared region [30] and (b) radiant flux versus infrared wavelength for a black body with a temperature of 300 K [31]. The temperature change correlates to the energy of the absorbed radiation and is measured by a change of the electrical resistance of the bolometer thermistor material. Bolometers absorb the incident radiation and cause the thermally isolated bolometer membrane to increase its temperature. Figure 2a shows a typical infrared bolometer pixel and a two-dimensional bolometer focal plane array.75 µm to 1000 µm. Ensembles of infrared detectors in two-dimensional arrays are called focal plane arrays (FPAs). The radiation flux from real objects varies. For microbolometers. 2. 6836 68360D-2 . surveillance. Figure 2b shows a schematic of a typical assembly of an infrared imaging system with a FPA and an infrared lens system.mine detection. Proc. fire fighting. ranging from 0. reconnaissance. There are also bolometers that are optimized for broad-band detection in both the 3-5 µm and the 8-14 µm wavelength region at the same time [33].

Infrared imaging systems based on uncooled bolometer arrays can reach NETDs of below 25 mK with a F-number of the infrared optics of F = 1 [29. The responsivity ( ℜ ) is defined as the output signal voltage or output signal current of an infrared bolometer pixel per incident radiant power on the pixel.6) where F is the F-number of the infrared optics.5) VROIC ⎛ R ⋅R = Vamp ( xl ⋅ f i ) + + ⎜ ROIC bol ⎜R 12 ⎝ ROIC + Rbol λ 1−λ 2 VQ2 ⎞ ⎟ ⋅ I ROIC ( xl ⋅ f i ) 2 ⎟ ⎠ 2 (Eq.2) NETDJohnson ⎞ ⎛ ⎛C ⎞ T1 ⋅ Rbol ⋅ xl ⋅ f i ⋅ 1 + ω 2 ⋅ ⎜ ⎟ ⎟ ⎜ 2 F G ⎝G⎠ ⎟⋅2⋅ k ⋅ ⋅ = ⎜4⋅ ⎟ ⎜ ∆P ⎞ TCR ⋅U bias β ⋅ A ⋅ ε λ1−λ 2 ⎛ ⎟ ⎟ ⎜ φ λ 1−λ 2 ⋅ ⎜ ⎝ ∆T ⎠ λ1−λ 2 ⎠ ⎝ 2 (Eq. The detectivity provides information that is equivalent to the noise equivalent power (NEP).4) NETDROIC ⎞ ⎛C⎞ ( RROIC + Rbol ) ⋅ 1 + ω 2 ⋅ ⎜ ⎟ ⎟ 2 F G ⎝G⎠ ⎟⋅ ⋅ ⋅ VROIC ⎟ β ⋅ A⋅ε ∆P ⎞ TCR ⋅U bias ⋅ RROIC ⎛ λ 1− λ 2 ⋅⎜ ⎟ ⎟ ⎝ ∆T ⎠ λ1−λ 2 ⎠ 2 2 (Eq.e. 42]. but with the possibility to compare bolometer pixels of the same type but with different pixel areas. resulting in a signal to noise ratio of 1). the NETDJohnson from the Johnson noise of the bolometers. φ is the transmission of the infrared optics in the wavelength interval from λ1 to λ2. the NETDthermal from the thermal fluctuation noise of the bolometers including temperature fluctuation noise from radiation heat exchange. 34-40]. C is the heat capacity of the Proc. The noise equivalent power (NEP) is the incident infrared power on an infrared bolometer pixel that generates a signal output that is equal to the root-mean-square (RMS) noise output (i. For conventional infrared imaging systems based on uncooled bolometer arrays with column based read-out designs and integrated analogue-to-digital conversion (ADC). the contributing NETD parts can be estimated by equations 2 to 6 ⎛x ⋅f ⎞ K ⎛C ⎞ ⎛ ⎞ ⋅ ln⎜ l i ⎟ ⋅ 1 + ω 2 ⋅ ⎜ ⎟ ⎜ ⎟ ⎜ f ⎟ 2 υ ⎝G⎠ F G ⎝ s ⎠ ⎟⋅ NETD 1 = ⎜ 4 ⋅ ⋅ ⎜ ⎟ β ⋅ A⋅ε TCR ⎛ ∆P ⎞ f λ 1− λ 2 ⎟ ⎜ φ λ 1− λ 2 ⋅ ⎜ ⎟ ⎝ ∆T ⎠ λ1−λ 2 ⎠ ⎝ 2 (Eq. For the assumption that the bolometer temperature and the surrounding (background) temperature are equal. the responsivity ( ℜ ). 39] 2 2 2 2 (Eq. 37.1) NETD 2 = NETD + NETD + NETD + NETD 1 f Johnson thermal ROIC where the total NETD consists of the NETD1/f from the 1/f-noise of the bolometers.3) NETDthermal ⎛ ⎜ = ⎜4⋅ ⎜ ⎜ φ λ 1− λ 2 ⎝ ⎛ ⎜ = ⎜4⋅ ⎜ ⎜ φ λ 1−λ 2 ⎝ ⎞ ⎟ T F G ⎟⋅ k ⋅ ⋅ 1 ⎟ β ⋅ A ⋅ ε λ 1− λ 2 C ⎛ ∆P ⎞ ⋅⎜ ⎟ ⎟ ⎝ ∆T ⎠ λ1−λ 2 ⎠ 2 (Eq. of SPIE Vol. gives rise to a difference in signal-to-noise-ratio of 1 in the electrical outputs of the two halves of the array. the NETD including the contributing NETD parts can be expressed by [38. 41. The noise equivalent temperature difference (NETD) is one of the most important performance parameters for infrared imaging systems and is defined as the difference in temperature between two side-by-side blackbodies of large lateral extent which. 32. viewing the two blackbodies. The detecitivity (D*) is defined as the root-mean-square signal-to-noise ratio of a 1 Hz bandwidth per unit root-mean-square incident radiant power per square root of bolometer area. G is the thermal conduction between each bolometer and its surroundings. and the NETDROIC from the read-out integrated circuit (ROIC) related noise. 6836 68360D-3 . when viewed by the infrared imaging system. the noise equivalent power (NEP) and the detecitivity (D*) [24.Important figures of merit for uncooled infrared bolometer and system performances are the noise equivalent temperature difference (NETD). 25.

the space between the bolometer and the neighboring bolometer membranes. VROIC is the total noise voltage of the ROIC. Vamp is the input reference noise voltage of the ROIC (depending on the ROIC architecture. Small bolometer pixels allow the implementation of high-resolution focal plane arrays at acceptable cost. a number of design features and tradeoffs have to be considered as can be seen from equations 1 to 9. xl is the amount of bolometer pixels per column (number of bolometers that are read-out during one imaging frame). ⎛ ∆P ⎞ ⎜ ⎟ ⎝ ∆T ⎠ λ1−λ 2 contrast in the wavelength interval from λ1 to λ2.1 Geometrical and optical bolometer design To design uncooled infrared bolometer arrays with a high sensitivity (low NETD). and the vias that connect the bolometer and the read-out integrated circuit wafer as can be seen in Figure 3a. The remaining area of the bolometer pixel is consumed by the bolometer legs. Some of the most important bolometer design parameters are a low thermal conductance between the bolometer and its surrounding. f bol = 3. β is the infrared absorption rate of the is the bolometer pixel fill factor. a bolometer temperature sensing material with a high temperature coefficient of resistance (TCR) and low 1/f noise properties and a sufficiently low bolometer thermal time constant. 43. ω is the modulation frequency of the infrared signal from the image scene. At the same time it is important for commercial infrared imaging applications. The bolometer pixel fill factor defines the portion of the bolometer pixel area that is used to absorb the incident infrared radiation.9) τ= G where τ is the thermal time constant of a bolometer pixel. Rbol is the bolometer resistance at the temperature T1. K is the 1/f noise constant of the resistive bolometer material. fs is the shutter or uniformity correction frequency for the bolometer array. 44]. VQ is the input reference analogue-to-digital quantization interval. two-layer bolometer designs (umbrella Proc. Bolometer Design and Manufacturing 3.7) f r = xl ⋅ f i 1 (Eq. the bolometer legs are long. and k is the Boltzmann constant. Ubias is the bolometer bias voltage. A is the bolometer pixel area (pixel pitch). typically a function of xl ⋅ f i ). fi is the image read-out frequency (imaging frame rate).8) 4 ⋅τ C (Eq. The legs typically contain a thin metal layer to provide electrical contact between the bolometer material and the read-out electronics. υ is the volume of the resistive bolometer material. To obtain a small thermal conduction between the bolometer and its surroundings. 37]. Figure 3a shows details of a typical bolometer design. Conventional single-level infrared bolometer arrays typically have a fill factor between 60 % and 70 % [32. To increase the bolometer pixel fill factor. The cost for both the FPA chip and for the infrared optics is reduces by reducing the active area of the FPA. a high absorption of the infrared radiation including a large absorbing area. ε λ 1−λ 2 bolometer membrane in the wavelength interval from λ1 to λ2. fr is the bolometer read-out integration frequency and fbol is the thermal bolometer integration frequency. RROIC is the input impedance (resistance) of the ROIC. The above NETD equations are valid for the assumption that the bolometer readout integration frequency fr is much larger than the thermal bolometer integration frequency fbol with (Eq. TCR is the temperature coefficient of resistance of the resistive bolometer material (dependent on the bolometer temperature). 6836 68360D-4 . Thermal conduction through the bolometer legs can be as low as 3. IROIC is the current noise from the ROIC input including the bolometer bias is the temperature current source (depending on the ROIC architecture. of SPIE Vol. K In addition. The vacuum atmosphere in which the bolometers are operated is typically on the order of 0. have a small cross-sectional area and consist of materials with a low thermal conductivity. that the bolometer pixels are as small as possible with reported pixel pitches being as small as 17 µm x 17 µm [33. typically a function of xl ⋅ f i ).bolometer pixel. conventional bolometers are operated in a vacuum package to minimize the thermal conduction between the bolometers and their surroundings through the surrounding gas.01 mbar [45]. T1 is the bolometer membrane temperature.5 ⋅ 10 −8 W [32].

42]. For a targeted wavelength interval of 8 µm to 14 µm. The membrane thickness is typically set to λ x . Such umbrella designs have been implemented in bolometer FPAs with very small pixel sizes [33. 45. Antireflection Layer Bolometer Membrane Free Space Infrared Mirror λ 4 To provide a high absorption of the radiation in the bolometer membrane. the ROIC is pre-manufactured and the bolometer Proc. 6836 68360D-5 . conventional bolometers contain resonant optical cavity (Fabry-Perot) structures that are optimized for the targeted wavelength interval as depicted in Figure 4.2 Bolometer focal plane array manufacturing techniques The most commonly used manufacturing approach for uncooled infrared bolometer FPAs is monolithic integration [37. bolometer membrane is placed at a distance of λ from the mirror surface on the substrate. Most commercial bolometer FPAs make use of the optical cavity design shown in Figure 4a since it allows thinner bolometer membranes with a lower heat capacity. 41. in which the resonant optical cavity is part of the bolometer membrane. The top metal layer (e.g.5 µm [37. which in turn can be used to minimize the thermal bolometer conductance and the resulting NETD.designs) as shown in Figure 3b have been reported that reach fill factors of up to 90 % [46. 3. of SPIE Vol. — 10 05 0 (a) (b) Figure 3: (a) Details of a typical single-level bolometer design [24] and (b) example of a two-level (umbrella) design for improved pixel fill factor [33]. A second type of resonant optical cavity design is shown in Figure 4b. titanium nitride) of the Fabry-Perot cavity usually has a target sheet resistance of 377 Ω/sq to reach optimum performance of the FabryPerot cavity in the wavelength interval of 8 µm to 14 µm [48]. The bolometer legs. where λ x is the wavelength of the targeted infrared 4 radiation in the bolometer membrane material(s) [48]. and in some cases the temperature sensing material [33] are placed underneath the absorbing bolometer membrane. The mirror of this type of resonant optical cavity is placed at the lower surface of the bolometer membrane and the thickness of the bolometer membrane defines the resonant optical cavity. 49] as shown in Figure 5. The most commonly used resonant optical cavity design is shown in Figure 4a. 46]. the distance between the bolometer membrane and the mirror on the substrate is typically about 2 µm to 2. in which the infrared mirror (typically aluminium) of the resonant optical cavity is placed on the surface of the underlying substrate (the ROIC) and the Via Antireflection Layer Bolometer Membrane Infrared Mirror λx 4 Via d d Substrate Wafer (ROIC) Substrate Wafer (ROIC) (a) (b) Figure 4: Cross-sectional image of two bolometer designs with resonant optical cavities for high absorption of the incident radiation. 42. In monolithic integration. 47]. Thus. a high fraction of the 4 incident infrared radiation at a specific wavelength λ is absorbed in the bolometer membrane [37].

(b) deposition of bolometer materials. the bolometer materials are deposited on a separate handle wafer. 0. 6836 68360D-6 . The bulk micromaching processes can be implement before. In 3D bolometer integration. (b) adhesive wafer bonding. Typically. (d) via formation. 51-56]. Sacrificial Layer Bolometer Materials Via Landing pads Bolometer Vias Bolometer Membrane ROIC Wafer (a) ROIC Wafer (b) ROIC Wafer (c) ROIC Wafer (d) ROIC Wafer (e) Figure 5: Monolithic integration for uncooled infrared bolometer arrays: (a) deposition of sacrificial layer on ROIC wafer. a high-temperature stable polyimide is used as the sacrificial layer. (f) etching of sacrificial layer. The materials are then transferred from the handle wafer to Proc. Commercial infrared bolometer arrays that are manufactures using bulk micromachining techniques are diode bolometers. A disadvantage of bulk-micromachining is that the electronics for the signal readout can not be placed underneath the bolometer membranes. the polyimide layer is sacrificially removed in an oxygen plasma to obtain the free-standing. the bolometers are formed in the substrate surface of a wafer. (f) sacrificial etching of polymer adhesive. in between or after processing the wafers e. of SPIE Vol. (d) bolometer definition. One potential disadvantage of monolithic integration is that the deposition process for the temperature sensing bolometer material is limited to about 450ºC and does not allow the deposition of monocrystalline materials. thermally isolated bolometer membranes. in a CMOS line to implement the necessary electronic components.materials are subsequently deposited and patterned on the ROIC wafer. Subsequently. Heterogeneous three dimensional (3D) bolometer integration has been proposed for the integration of high-performance mono-crystalline temperature sensing bolometer materials on ROICs [48. (a) (b) Figure 6: Bulk micromachining for uncooled infrared bolometer arrays: (a) Formation of the bolometer and the electronics for signal read-out (typically side-by-side). In the final step. Monolithic integration is a cost-efficient and well-established post CMOS process in which the electronics for the signal read-out can be efficiently placed underneath and beside the bolometer membranes. This can make the optimization of the temperature sensing bolometer material difficult.g. (e) via formation. This usually reduces the bolometer fill factor. All commercially available vanadium oxide and amorphous silicon bolometer FPAs are monolithically integrated on top of the ROICs. The advantage of bulk micromachining is that the electronics and the bolometers can typically be manufactured in a standard CMOS line. (c) thinning of handle wafer. The free-etching of the bolometers can then be done in a single post CMOS process step.t i a e n-well. but has to be placed beside the bolometers. Thin-Film Resistive Thin-Film Thermistor Bolometer Material Material Handle Wafer Polymer Adhesive Etch-Stop Layer Handle Wafer Via Landing Pads Bolometer Vias ROIC Wafer (a) ROIC Wafer (b) ROIC Wafer (c) ROIC Wafer (d) ROIC Wafer (e) ROIC Wafer (f) Figure 7: Heterogeneous 3D integration for uncooled infrared bolometer arrays: (a) separate fabrication of ROIC wafer and handle wafer with resistive bolometer material. the substrate is selectively etched underneath the bolometers to thermally separate them from the rest of the substrate. (c) patterning of bolometer materials. (b) selective etching of the bulk material underneath the bolometer membrane [50]. Bulk micromachining is a second alternative to manufacture uncooled infrared bolometers [50] as shown in Figure 6. Process temperatures higher than 450ºC risk damaging the ROIC. In bulk micromachining.

3. KTH-MST and Acreo [53-55]. amorphous silicon (α-Si) and silicon diodes. There are many phases in vanadium oxides. 6836 68360D-7 . of SPIE Vol. so its resistance is very low at room temperature. al. but its resistance at room temperature is very high. a controlled mixed phase of VOx with good electrical properties by can be formed by post-annealing in oxygen. TCR up to about 3 %/K at room temperature have been reported [69]. 3. Because high electrical resistance of a device results in a high level of noise. Several groups are active to find processes to obtain high TCR in combination with sufficient low sheet resistance for good noise performance. 58] and is today used in a variety of bolometer products.3 Bolometer temperature sensing materials The selected bolometer temperature sensing material has a large influence on the sensitivity (NETD) of the bolometer. The advantage of 3D bolometer integration is that it allows the use of highperformance mono-crystalline temperature sensing bolometer materials on top of standard ROICs. J. Bolometers made of α-Si can consist of very thin memebranes. At the same time. Thus. M.3. Lv et. Today. a CMOS wafer) in a cost efficient way.2 Amorphous silicon (α-Si) Amorphous silicon (α-Si) is extensively described and used in a large variety of products such as active layer in thin-film transistors for liquid crystal displays.the ROIC wafer using low-temperature adhesive wafer bonding in combination with sacrificial removing of the handle wafer as shown in Figure 7c. The 1/f noise constant is a material parameter that can vary several orders of magnitude for different materials and even small variations of the material composition can dramatically change the 1/f noise constant [35]. small area solar powered photovoltaic devices for consumer products and large area power solar cells. The achievable bolometer dimensions are identical to the ones made with conventional monolithic micromachining techniques. A high temperature coefficient of resistance (TCR) and a small 1/f noise constant are desirable material properties as can be seen from equations 1 to 9. The performance limits of uncooled VOx microbolometer focal plane arrays are reported in [32]. it must be possible to integrate the temperature sensing material together with signal read-out electronics (e. V2O5 and V2O3. in 2000 [68]. including the annealing and oxidation of evaporated vanadium under controlled conditions. [67] report a strong effect of Ti–W co-doping on both the optical and electrical properties of vanadium dioxide thin films grown by a reactive pulsed laser deposition. Y. V2O5 can be formed by ion beam from vanadium metal target in high O2 partial pressure. 38. Vanadium oxide films can be prepared by a variety of methods. Singlecrystal VO2 and V2O5 have TCR above 4 %/K.1 Vanadium oxide (VOx) The vanadium oxide thin films that are used have TCR in the range 2 %/K and 3 %/K at room temperature [57. one way to reduce the NETD of a bolometer FPA may be the use of mono-crystalline temperature sensing bolometer materials with a low 1/f noise constant [32. the most common bolometer temperature sensing materials are vanadium oxide (VOx).g. [62] have shown that through the formation of a sandwich structure of V2O5. V and V2O5 by a conventional sputter method. Han et. al. mono-crystalline materials can have a significantly lower 1/f noise constant as compared to amorphous or poly-crystalline materials. A 3D bolometer integration process is currently being implemented in a project with consortium partners Autoliv. 36. For most materials 1/f noise constant is not very well documented in literature. [63] have obtained TCR value about 4. The hydrogenated amorphous silicon (α-Si) microbolometer arrays that have been developed take advantage of the high TCR. the use of the V2O3 phase showing low resistance is important to the fabrication of low-noise microbolometers.12 %/K. Soltani et. They undergo transition from an insulator or semiconductor to a metal phase at a specific temperature. al. relatively high optical absorption coefficient and the advantage that they can be manufactured using silicon fabrication compatible processes. Y-H. but are difficult to make. The NETD of state-of-the-art uncooled infrared bolometer FPAs is typically limited by the 1/f noise from the bolometer temperature sensing material [32. Proc. however. for bolometers with a low thermal conductance while maintaining a fixed bolometer time constant. Umicore.4 %/K and low sheet resistance <20 kΩ/square by using facing target sputtering and heat treatment. V2O3 has low formation energy and undergoes a transition from semiconductor to metal phase at low temperature. reactive RF sputtering [49.3. which allows for a low thermal mass and consequently. It is reported [64-66] that bolometer characteristics can be improved by using vanadium-tungsten oxide made by low temperature oxidation of vanadium-tungsten metal films. 39]. Commonly used thin film depositions techniques such as evaporation and sputtering will give amorphous or polycrystalline films. Syllaios et. VITO. such as VO2. 60] and pulsed laser deposition [61]. Infineon Technologies SensoNor. They achieved films with TCR 5. 36]. al. A description of amorphous silicon microbolometers technology was published by A. 3.

An advantage of using diodes may be the possibility this gives for smaller pixel size because diodes can normally be made on smaller areas than resistors. but plastic as well. For MEMS based Proc. high-volume applications. The relatively lower electronic performance of low-temperature α-Si devices could be compensated by the cheaper production. Very thin membranes (100nm) can be produced using standard CMOS processes. ultra-low-cost. integration of IR window material with good infrared transmission. for future. which allows for deposition on not only silicon and glass. Other reported alternatives for poly-SiGe film are molecular-beam deposition [79].5 to 6.3.3 %/K.5 %/K at room temperature. of SPIE Vol. 76]. 75.g. an order of magnitude lower than the best resistive devices. conventional bolometers operate with vacuum levels < 0.4 %/K been obtained [86]. At room temperature has a TCR of 4. allows standard metallization annealing temperatures [54. 81].g.4 Bolometer vacuum package Bolometers need to be packaged in vacuum for best performance. The TCR can thereby be high (3. such as plasma-enhanced chemical vapor deposition (PECVD) or sputtering. reduced-pressure chemical vapor deposition (CVD). The Ge content is in the order of 85%.g. or by plasma enhanced chemical vapor deposition [77].3 Silicon diodes The temperature coefficient of the forward voltage of a pn-junction or Schottky barrier junction can be used as sensing principle for the temperature caused by absorption of IR radiation. Important requirements for the packaging of bolometer arrays are: good and reliable hermetic seal. e. Titanium is preferred due to its low thermal conductance. 85]. e. al.2 %/K. The material can also be designed with pyroelectric properties depending on the oxygen stoichiometry. With a forward voltage of about 0. TCR can be controlled with the deposition parameters and e.7 V for silicon pndiodes and about 0.6 V for silicon Schottky-diodes and a voltage change with temperature in the range 1-2 mV/K give a temperature coefficient of about 0. VOx due to the higher 1/f noise of the poly crystalline material. Titanium up to 0. TCR values up to 5. Dong et. A recently reported material with performance claimed to be compatible with VOx is thin-film carbon. Very high TCR values can be obtained around the phase transition temperature when changing from semiconductor to metal. Due to the high deposition and annealing temperatures has the membrane however to be manufactured before metallization [78]. Therefore. which results in low performance detectors [73]. The noise level increase however with higher TCR and a trade-off has to be made to achieve maximum performance. Another alternative is poly crystalline silicon germanium (SiGe) with approximately 30% Ge. making it a candidate for a roll-to-roll processing technique.7 and 2. It is made by depositing a parylene layer and pyrolyze it to carbon at 800°C [87]. Both reliability and cost of MEMS is heavily related to the encapsulation techniques chosen. An alternative to α-Si are different types of amorphous germanium-silicon-oxygen compounds (GexSi1-xOy) grown by reactive sputtering in an Ar or Ar:O2 environment [74. The detector performance that can be obtained is up to now lower than e. The TCR and sheet resistance depends on material properties such as doping concentration. The properties are therefore a direct consequence of the deposition process. By growing very thin (nm) Si/SiGe multi-layers epitaxially is it possible to create a single crystalline material with integrated valence barriers.01 mbar [47]. or ultra-high vacuum vapor deposition [80]. Deposition can be made at low temperature with magnetron sputtering.g.g.3 %/K) and simultaneously. deposition temperature and annealing. The TFT used had a high temperature coefficient of its drain current in the range 1. The advantage is lower thermal conductance. as low as 75 °C. [70.Hydrogenated amorphous silicon (α-Si) is a thin–film amorphous material that can only be produced by a nonequilibrium process. The bolometer diodes can potentially be manufactured in standard CMOS lines. α-Si can be deposited at very low temperatures. Wide-band gap materials have also been investigated. They are easy to integrate with CMOS ROIC process and the 1/f noise is low. which e. but the relatively high 1/f noise lower the potential detector performance. the high quality crystalline material provides very low 1/f-noise characteristics. 3.3.1 %/K have been reported. 3. Another type of thermistor materials are semiconducting YBaCuO.4 Other Bolometer Materials A simple alternative for temperature sensing material are thin film metals. The use of amorphous silicon thin film transistors (α-Si TFT) as the active elements has been reported by L.g. and high yield low cost packaging. TCR is unfortunately also very low (e. and TCR values between 2. 6836 68360D-8 . Perovskite metal-oxide manganites with colossal magneto resistance effect (CMR) have also been proposed for thermal imaging. Advantageous is also the stability of the quantum well thermistor material.35 %/K [72]).8 and 4 %/K have been reported [84. PECVD deposited amorphous silicon carbide [82] and photochemical vapor deposition of nanocrystalline silicon carbide [83]. be varied between 0. 71]. 3.

Standard antireflection coating is deposited on both sides of the window areas to obtain windows without any significant degradation of the final effective NETD. A seal ring is added on the frame areas of the ROIC wafer. The cap-wafer is then aligned and attached to the ROIC wafer with the pixels and sealing take place in vacuum. An important advantage for MEMS based bolometer arrays is the possibility this technology offers for low cost wafer level vacuum packaging with methods that has become available the last few years. The need for packaging of microbolometers in vacuum is related to the fact that heat loss due to thermal conduction from the bolometer structure through the gas gap to the substrate underneath will increase the NETD as indicated in equations 1-9. The window areas are made by deep reactive-ion etching. X. For a specific new volume automotive application.1 mbar for a device with pixel area 50 µm x 50 µm and an air-gap of 20 µm. 6836 68360D-9 .bolometers. The cap-wafer material (window) is made in single-crystal silicon wafers with low oxygen concentration to obtain good IR transmission over the whole wavelength range 8-14 µm. eutectic bonding and soldering. A is the area of the bolometer plate and d is the distance between the plate and the substrate. The most common way of packaging is to build the bolometer chip into a hermetic metal or ceramic package with an IR-transmitting lid build into the package cap. Several alternative methods for wafer level packaging are already developed by the MEMS industry. and metal based bonding that includes thermo-compression bonding. Good descriptions of these methods will be available in a new handbook for MEMS [89]. of SPIE Vol. At higher pressure levels the thermal conductance Gair through the air-gap can be estimated to be: Gair = kair A/d (Eq. the areas on the cap-wafer around the pixel areas are removed by standard wafer dicing saw in order to release the wire bonding pads.01 mbar can be obtained and open up for applications requiring better NETD. The process sequence for the wafer scale packaging allows building of additional thin-film getter structures and Pirani gauge vacuum sensors into each device. Hereby vacuum pressure < 0. The inherently good detector performance developed allows us to use a lower vacuum without compromising overall performance too much. Development of specific methods to measure the quality of the vacuum both on wafer level and in the final devices are investigated. Hermetic methods include low temperature fusion bonding. Wafer-level vacuum packaging may be an important factor in bringing commercialization of micro-bolometer arrays into low-cost. hereby reducing the system cost substantially. this allows the bolometer chip to be integrated directly to the camera housing. This technology allows the pixels to be packaged in a hermetic cavity with the electrical contacts outside the cavity at low cost. glass frit bonding. industry. automotive and environmental monitoring. Proc. Establishing manufacturing schemes by using accelerated testing for checking the actual gas pressure and bonding consistency will hereby be made possible at low cost. Silicon lid wafer with IR coating Silicon lid wafer with IR coating Wire bonding pads Pixels Pixels Substrate wafer (ROIC) Substrate wafer (ROIC) (a) (b) Figure 8: Principle cross-sectional of wafer level packaging of silicon lid wafer with window areas onto ROIC wafer with bolometer pixels (a) after wafer level sealing and (b) after release of wire bonding areas by sawing. high volume applications such as home. [88] have measured the typical thermal conductance and heat capacitance of a microbolometer as a function of pressure at room temperature on test structures. 10) where kair is the thermal conduction for the air. This enables good long term stability and makes the use of getter materials and custom processes obsolete. After sealing. Figure 8 shows the principles of the wafer scale packaging technology that is being developed. we develop a wafer scale package to form lids with silicon IRwindow over the bolometer pixel [55]. He et al. anodic bonding. As a result. Their results shows that thermal conduction through the gas starts to have an effect from a pressure in the range 0. packaging may be done at chip level or wafer level.

To measure the resistance change of a bolometer. 640x512 160x120. Thus. 45]. 640x480 320x240 320x240 384x288 Pixel pitch (µm) 25 37. 5.5 25 Detector NETD (F=1. Bolometer Read-Out Integrated Circuits (ROICs) Bolometer read-out integrated circuits allow measuring the very small resistance changes of individual bolometers that are part of large bolometer arrays. France Mitsubishi. Figure 9a shows scanning electron microscope (SEM) images of a commercial VOx bolometer [32] and Figure 9b shows a VOx umbrella type bolometer [33] that is in the R&D stage. As a result of the bolometer self-heating. Other suppliers include Raytheon in USA. USA L-3.5 30 28 R&D: 17 25 R&D: 17 25 R&D: 17 25-50 25 23. Table 1 shows an overview of the main suppliers. the ROIC is required to handle a higher dynamic range. Commercial and State-of-the-Art Infrared Bolometer Arrays Today. modern FPAs have implemented design features that tolerate and compensate for a floating chip temperature in the specified temperature range [92].640x480 160x120 . USA Bolometer type VOx bolometer VOx bolometer α-Si bolometer VOx bolometer VOx bolometer (standard design) VOx bolometer (umbrella design) VOx bolometer (umbrella design) VOx bolometer VOx bolometer (umbrella design) α-Si bolometer Si diode bolometer VOx bolometer VOx bolometer Array format (pixels) 160x120 . consist of two-layer (umbrella-type) bolometer designs [33.wassenaar. the noise voltage from the ROIC should be sufficiently low to not be the domination noise contribution to the NETD.640x480 320x240 320x240 320x240 . Commercially available bolometer arrays are either made from vanadium oxide (VOx). Early FPAs used thermo-electric-coolers (TEC) to stabilize the FPA chip at a fixed temperature and to provide a low noise voltage. today many infrared cameras do not require a TEC for temperature stabilization for the FPA chip. However. Mitsubishi in Japan. BAE Systems in USA and DRS in USA. 6836 68360D-10 . Most commercially available bolometer FPA make use of standard single-level bolometer designs. the ROIC applies a bias voltage which causes a temperature increase (offset) in the bolometer membrane.640x480 320x240 160x120 . Japan SCD. 20-60 Hz) 35 mK 50 mK 50 mK 30-50 mK 50 mK 35 mK 50 mK 30-40 mK 50 mK 35-100 mK 50 mK 75 mK 50 mK DRS. Japan NEC. of SPIE Vol. L-3 in USA. Company FLIR. the bolometer technology and specifications for existing products and for bolometer arrays that are in the R&D stage. 90-92]. USA ULIS.4. the dominating suppliers of high-performance uncooled infrared bolometer arrays are Flir Systems in USA. NEC in Japan and SCD in Israel. Table 1: Commercial and state-of-the-art R&D uncooled infrared bolometer arrays. USA BAE. amorphous silicon (α-Si) or silicon diodes. As can be seen from equations 1 to 9. Ulis in France.320x240 320x240 . USA Raytheon. with VOx being the dominating technology. Column parallel read-out architectures with integrated AD conversion are commonly used in commercial FPAs [32. Israel Proc. some of the FPAs with very small bolometer pixels that are in the R&D stage. Highperformance uncooled infrared bolometer FPAs are subject to export control under the Wassenaar agreement (www.

machine vision and consumables. 2. 43. 6. Camera module integration. Outlook and Future Developments Development of micro-bolometer technology has the last 5 years continued to dramatically improve performance and resolution. First level packaging needs to be integrated in the chip fabrication process using wafer-level technologies and integrated getter materials. Vacuum packaging. Mass-market application will become the technology driver. This changes the development focus from performance and resolution [33. 3. The next generation detectors will have a pixel pitch of 17 µm. power consumption and cost of uncooled technology. • Medium resolution. high performance imaging for surveillance and military – mega pixel arrays with lowest possible NETD. This will open up for many producers and in long term also lower manufacturing cost for smaller series production. the cost for uncooled infrared imaging chips are decreasing accordingly. e. and the production yield and testing requirements will be of larger importance for the final production cost. High-end products for niche markets will in the long run benefit from the achievements and manufacturing infrastructure resulting from the development of mass-market products. 6836 68360D-11 . similar as digital CMOS cameras are produced today. The demonstrated performance is getting closer to the theoretical limit and the gap to photonic detectors is getting smaller. 80x30 pixel arrays with a NETD of about 150 mK @F1. microbolometer detectors are now produced in larger volumes than all other IR array technologies together. For the future we can divide the technology requirement and thus development efforts towards the following main market segments: • High resolution. Uncooled infrared bolometer arrays are reaching performance levels which previously only were possible with cooled infrared photon detectors.g. With a continuously increasing market volume (>100 000 units per year to date). s__a '1 (b) Figure 9: (a) Commercial VOx bolometer design with 28 µm x 28 µm pitch from BAE [32] and (b) R&D VOx bolometer with umbrella design and 17 µm x 17 µm pitch from DRS [33]. low performance image based sensing for automotive safety. Conclusions Bolometer arrays have become the technology of choice for low-cost infrared imaging systems used in both civil and military applications. 44] to low cost production [54. Foundry manufacturing. State-off-the-art performance is for many applications good enough. With the advantages regarding weight. Only standard materials and fabrication processes that can be used in CMOS and MEMS foundry fabrication. The main development issues are: 1. 93. medium performance imaging for automotive and security – 320x240 pixel arrays with a NETD of about 35 mK @F1. 94].(a) TJ — —j s_-a -. This puts additional requirement on high temperature stable detector material. of SPIE Vol. To increase detector sensitivity. Low resolution arrays will be completely integrated with driving electronics and optics. • Low resolution. new temperature sensor materials with low 1/f noise and high Proc. Initial development cost and complexity of the manufacturing will be of less importance for large volume production.

74-76. Lubke. R. Doering.from ceramic hybrid arrays to direct thin film integration”.64.E. Weinstein. Selders. Ohnstein.10. W. Samuel Pierpont Langley”. Watton. Journal of Microelectromechanical Systems. Cole.6. Baltes. “The Infrared Pioneers – III. Carney. pp.E. Muralt.E.TCR are required. R.195-206. 6836 68360D-12 . Foss. T.J. Ferroelectrics. Schneeberger. IRIS DSG.A. C. E. R.29. Vol. C. Feb. 1984.43-49.S. Infrared Physics. E. Nielsen. Brand. C. Foss. Vol. Vol. K.42. pp.S. “Integrated uncooled infrared detector imaging arrays”.J.C. J. pp. “Monolithic silicon uncooled focal planes for the HIDAD program”. ”Construction and characteristics of evaporated nickel bolometers”. Blevin. 1986. 1956.57-64. No. Han. Proc. R.4. No.E. Billings.102-116. making post-processing deposition impossible. Hartmann.5. W. pp. “Ferroelectric IR bolometers . 1965. M. “Monolithic silicon uncooled focal plane development for high density array development program (HIDAD)”. pp.3.429-435.H.E. GOMAC. R. H. Mass-market applications will focus the development on low-cost manufacturing and integration of the whole IR camera module. Review of Scientific Instruments.A. 1991. Infrared Physics. Essen. IRIS DSG. Wood. West Germany. The Detection and Measurement of Infrared Radiation. Solid-State Sensor and Actuator Workshop 1992. 1957. Hyde. “Thin-film resistance bolometer IR detectors – II”. 377-385. Proc. Philips. 1988. O. Holmen. R. P.W.A. Chasmar. Cole. “Advances in uncooled silicon monolithic IR 2D arrays”.E. pp. Heterogeneous 3D integration processes will be developed to solve this issue. Barr. This requires trade-off optimization of the performance related to cost for each part in the system. B. “Micromachined infrared detectors based on pyroelectric thin films”. R. Paul. F.A. "An investigation of the properties of bolometers made by vacuum evaporation". E. pp. P. Vol. As example can the requirement on a very low vacuum pressure be exchanged with a better performing detector material. pp.A. Vol. Higashi. Reports on Progress in Physics.J. N. Qiuting. R. J.26. Aiken. D.17. Higashi. Wood. References [1] [2] Websters Dictionary. Oxford. K. R. R. Kruse. Vol. “HIDAD – a monolithic silicon uncooled infrared imaging focal plane”. R. Proc. Special thanks to John Franks and Jan Van Nylen from Umicore for providing the illustration in Figure 2b and to Jan-Erik Källhammer from Autoliv for valuable feed-back for improvements of the paper.C. A.J. R. “Progress in silicon monolithic uncooled IR focal plane arrays”. Clarendon Press. “High-sensitivity thin-film bolometers”. IRIS DSG. No.E.2. pp.18. Brown. Barr. B.E. pp. N. Cole. Smith. 1991. W.132-135. 1946. Liddiard. C. Higashi.1. 1990. of SPIE Vol. 2001. Electronics. M. Sensor 1982 – Sensor Technology and Temperature Measurement. R.E. The manufacturing needs to be aligned with standard CMOS and MEMS foundry processing. “Thin-film enhances bolometer's sensitivity". 1980. Han. R. R. Menolfi. A. C. Proc.A. B. Wood. Proc. Liddiard. Gosch. R. Review of Scientific Instruments.19. Journal of Scientific Instruments. N. pp.184. 1996. pp.141-150. Higashi. K. Loepfe. Carlson.503-510. Vol. Schaufelbuhl. Wood. Proc. Han. 2001. U. Waelti. C. Infrared Physics. J. O. B. Proc. “The production of film type bolometers with rapid response”. Jones.A.L.A. Munch. IRIS DSG.24. H. No. “Large-area bolometers of evaporated gold".A.J. Hilton Head Island. “Uncooled low-cost thermal imager based on micromachined CMOS integrated sensor array”. Yosihara. Science of Light. Carter. Han. Vol. This is required for large volume fabrication and the technology development progress will even faster benefit from the achievements continuously gained in the whole semiconductor field. “Thin-film resistance bolometer IR detectors”. No. 1947. 1989. USA. [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] Proc. Wood.F. pp.B. R. Wood.10. R. M. Higashi. F. The most promising materials are fabricated at relatively high temperature. Vol.1339-1388.A.PIMS” for the fruitful cooperation.H. 28. W.E. Acknowledgments The authors thank all co-workers in the Eurimus-project “Pedestrian Injury Mitigation System . 1963. Vol.

USA. [22] R. C. San Diego. Infrared Physics and Technology. Baur. Orlando. J. Tohyama. A. Malyarov. [44] D. Gordon. San Diego. Elliott. Amsterdam. J.101-106.43. Proc. “Characterization of transferbonded silicon bolometer arrays”. USA. SPIE 2004.A. Proc. pp. USA.5406. Ray. S. G. M. Vol. “Performance limits of uncooled VOx microbolometer focal plane arrays”. SPIE 2003. USA. E. [27] A. Vol.S. in P. http://en. M. Hellström. Jansson. Orlando. J. Leary. Rogalski. Kluwer Academic Publishers. Decharat. in H. SPIE 2003.A. Gordon and Breach Science Publishers. Proc. P. Filachev. USA. pp. Hewitt. T. Rumbaugh. [26] T. M. Kruse. Ray. Marasco. pp. [35] V. C. CH. Wood. Arrays. Laser Focus World. L. Backer. 5406. C. AMAA 2003. Germany. pp. Niklaus. Capper. Letellier. SPIE 1993. “Monolithic silicon microbolometer arrays”. Almada. S. R. C. Sasaki. N.5074. Emmett. Dainese. R.doi. 2002. Zerov. of SPIE Vol. 2005. Sessler. R.G. “Infrared microbolometer sensors and their application in automotive safety”. Hewitt. USA. Semiconductors and Semimetals. D.6542. V. C. V. Proc. 531-540. pp. Weinheim.W. Kruse.W. pp.548-556. “Uncooled infrared sensor performance”. Advanced Micro and Nanosystems. Tissot. A. Infrared Physics and Technology. M. CMOS . Ouvrier-Buffet. Vol. [28] A Rogalski.L. pp. Orlando. and Applications. 1999.6542. Yon. T. Fedder. J. [38] F. C. Balceraka. KGaA. [34] P. Proc. Vol. 447-453. Kurashina. Stemme. pp. Decharat. Yon. [23] P.62940F1-62940F11. N. Tissot. G. USA. J. pp. Vol. [39] F. Baur. Orlando. Han. J. T. 1998. Gordon. E. pp. Murphy. WILEY-VCH Verlag GmbH & Co. [42] D. Mottin. Källhammer. SPIE 2004. pp. J.253-258. Proc. 5406.6542. 2000. Kennedy. D. USA. SPIE 2003. SPIE 2004.J. E. [41] B. “Performance improvements for VOx microbolometer FPAs”.1M.L. Clarke “Recent development of ultra small pixel uncooled focal plane array at DRS. Orlando. “Uncooled IR focal plane arrays”. Vol. Russia. Vol. D. 1993. Rapach. Sessler.L. Kruse. P. B. N. June. “Can the 300 K radiating background noise limit be attained by uncooled thermal imagers ?”. Wyles. Ponomarenko. C. Joswick. Williams.L.437-446. pp. Infrared Detectors. M. Van Lue. J. [48] F. Pettersson. USA. pp. SPIE 2004. L.MEMS. Oda. Taubkin. pp. Miyoshi. Buttler. pp. A.187-210. C. Hierold.363-378. Asbrock.45-121. Van Lue. Kennedy. pp. “Performance improvements for VOx microbolometer FPAs”. Biancardini.531-540.5406. Moscow. R. D. E. Proc. S.65421U. Proc.1016/j. Haggblad. Trujillo. [45] J. Stemme. A.1-1Y.4820.3. Orlando. [36] P. Blackwell.280-498. C. USA. ”Calculational modeling of the main characteristics of an uncooled linear microbolometer array”. 2000. A. Kohin. 6836 68360D-13 . J.12. Bellingham. SPIE 2004. 5406. Micromachined transducers sourcebook. O. Vol. Wyles. McGraw-Hill. C. Ushakova. “Advances in uncooled technology at BAE systems”.6542. Niklaus. Vol. R. The Netherlands. [25] P. SPIE 2004.149-174. J. Geneczko.A. G. SPIE 2007. Proc. J. K. Bach. G. D. Proc. Khrebtov.47. Uncooled Thermal Imaging.200-207. Hewitt. Gordon. I. http://dx. A.7. Blackwell. Opto-Electronics Review. Academic Press. USA. T. Infrared Detectors and Emitters: Materials and Devices.W. “CMOS-based Thermal Sensors”. “Infrared detectors: an overview”. pp. O'Donnell.12. T. No. Systems. Proc.J.2007. D.1Y. J. Wyles. Proc.infrared. [24] R.5406. Richwine. [43] R. “Uncooled Infrared Bolometer Arrays Operating in a Low to Medium Vacuum Atmosphere: Performance Model and Tradeoffs”. Vol. Wan. 5126. USA. “Micromachined bolometer arrays achieve low-cost imaging”.6294. [37] E. E. Skatrud. SPIE Press.-E. H. SPIE 2007.T.L. Proc. Proc. E. R. Wyles. Kruse. Wood. Navy. Proc. Proc. Foss.2. G. Orlando. [30] Source: U. A. Vol. Boston. J.2020. Vol. Jardine. [29] A. Freyvogel. Wyles. Pejnefors. “A comprehensive model for bolometer element and uncooled array design and imaging sensor performance prediction”. Stemme. in press.J. J. [47] D. I.A. Tabata. Ajisawa. 1997. Black.wikipedia. Kohin. Bisotto. Orlando. A. C. 2001. J. Wood.1-1M. M. U. pp. Howard. S. C. Baur. Robas. “Uncooled amorphous silicon technology enhancement for 25 µm pixel pitch achievement”. USA. S.153-157. SPIE 2007. [40] R. [32] M.A. Vol.65421Z. [46] S. “New thermally isolated pixel structure for highresolution uncooled infrared FPAs”. Orlando. Ito. M.T. Vol. Li. D.W.7. L. in P.F. Niklaus. of SPIE 2006.D.001. “17 µm pixel 640 x 480 microbolometer FPA development at BAE Systems” Ray. Vol. Martin.Y. Journal of Optical Technology. Vol. Jansson. J. Sessler. Orlando USA.[21] R.521-530. M. Orlando.08. “Uncooled microbolometer infrared sensor arrays”. S. K. Berlin. Soodb.A Kovacs. No. Korvink. “Infrared focal plane arrays: state of the art and development trends”. [31] G. O. 2007. Brand. USA. Seattle. Peysha. “Performance Model for Uncooled Infrared Bolometer Arrays and Performance Predictions of Bolometers Operating at Atmospheric Pressure”. SPIE 2007. Murphy. G. Murphy. Vol. Akin. USA. Proc. Bain. 1-15.J. Asbrock. USA.-E. pp. K. [33] C. 428-436. N. 52-85. pp. M. Vol. Vol. M. Boston. Dereniak.L. Wallgren. pp. pp. Van Lue. pp. Wyles. 2004. I. Baltes. Wood. “640 × 512 17 µm microbolometer FPA and sensor development”.

Eriksson. Vol. Pettersson.50. J. pp.1167-1177. Kurashina. 2005.4369. Stemme.397-404. pp. Heredia. Proc.Y. I. [54] C. Wissmar. Shannon. pp. Vieider. Chi-Anh. A. R. K-T.1-6. Shin. [66] N. C. Vol. 6836 68360D-14 .315. L. Struct. Vol. Kumar. Niklaus. Vol. “Characterization of uncooled bolometer with vanadium tungsten oxide infrared active layer”. “A high performance single-chip uncooled a-Si TFT infrared sensor”. Haddad.-Q. Journal of Micromechanics and Microengineering. Moon: “Excess noise in vanadium tungsten oxide bolometric material”. Transducers 2003. Vercammen. VanHulsel.312. B. Andersson. 2004. J. Jakobsen. [64] Y-H.450-452. Journal of Microelectromechanical Systems. pp. pp. Wu and Z.2. “Far infrared low-cost uncooled bolometer for automotive use”. Stemme. Niklaus. Yi. Applied Physics Letters. G. J.109. Mansi. Sensors and Actuators A: Physical. Franks. G. D. 2006. No. “AUTHENTIC: a very lowcost infrared detector and camera system”. [50] S.87-92. W. Proc. M. Vol. [51] F. Ericsson. Kruzelecky. H-J. Vol. Xiang. 99. R. pp. VanNylen. NEC Research & Development. Moon. Vol. J. “Fulfilling the pedestrian protection directive using a long-wavelength infrared camera designed to meet the performance and cost targets”. J. Källhammer. [52] F. J. Ericsson. Junique. SPIE 2001. H-J. Gopal.4969-4972.22. J. Chang-hong. pp. Vol. Akin. S. Vercammen. Dong. pp.. Stemme. B. No. Niklaus. pp. K. Surface and Coating Technology. 2001. Niklaus. J. 2005.. S. E. De la Hidalga. SPIE 2006. No. Bold. Pettersson. SPIE 2003. pp. No.10. [71] L. pp. H-K. Yue. F. Sabuncuoglu Tezcan. Ritchey. F. Porter.c. 2003. Vol. “ Effects of Ti–W codoping on the optical and electrical switching of vanadium dioxide thin films grown by a reactive pulsed laser deposition”. E. R. Eriksson. pp.227-238. [53] J. Proc. Kvisterøy. Zhang et al.5. H. Vol. Liu. Han. Vol. Lu. Ajisawa. K-T.50. No. Vol. Proc. S. reactive sputtering and annealing process”. MRS 2000. 2001. Halldin. [62] Y-H. Proc. pp. Yi. Kawahara. “Study of a pulsed laser deposited vanadium oxide based microbolometer array”. M. S. Jaramillo. A. Pettersson. X.1-2. L. Bi-feng. Thin Solid Films. G. Vol. G. “Performance of 320 x 240 bolometer-type uncooled infrared detector”.53-58.55-61.1958-1960. Andersson. pp. No. Kim et. [68] A. Narayandass. Applied Physics Letters. Kim. “Fabrication and performance of microbolometer arrays based on nanostructured vanadium oxide thin films”. 3. F. H. S. “Micromachined uncooled IR bolometer linear array using VO2 thin films”.1. “Fabrication of vanadium oxide thin film with high-temperature coefficient of resistance using V2O5/V/V2O5 multi-layers for uncooled microbolometers”. R.-Y. [56] F. H. The Electrochemical Society 2003. Franks. Vieider. S. Vol. Y. Gooch.-H. A. [63] Y. SPIE 2007.. Hu. Vol. pp. Abs. Shin and S.-E. [70] L.123. [69] A. pp.1L. A14. Stemme. Han. Ma. H.. pp. Germany. U.12. Orlando. Oda. G.J.. R. 2001. [65] N. D. Edwards. Manoravi. P.254101-3. Infrared Physics & Technology.1-1L.38-41. Forsberg. Kang. Strasbourg. Smart Mater. M. D.1. [57] C. Berlin. [55] T.031101/1-031101/28. Schimert. Moon. “Infrared responsivity of uncooled VO~2-based thin films bolometer”. A. “Low-cost far infrared bolometer camera for automotive use”.86. Brookfield. “Fabrication and characterization of integrated uncooled infrared sensor arrays using a-Si thin-film transistors as active elements”. F. Sylaios. T. Chen. C.188-192. Vol. L. T. 2007. M. “Low-cost uncooled infrared detectors in CMOS process”. “Wafer-level membrane transfer bonding of polycrystalline silicon bolometers for use in infrared focal plane arrays”. Acta physica Sinica.-E. McCardel. pp. Choi. AMAA 2007. ”Low temperature electrical properties of a boron-doped amorphous silicon bolometer”. Halldin. F. “Preparation of vanadium oxide thin films with high temperature coefficient of resistance by facing targets d.6198. Proc. Vol. [72] M.Y. G. pp. Vol. I. W. No. J. H-J.16. pp. H. J. Kim. Niklaus. X.1. Stemme. 2003.4820. H. “Thermal characterization of surface-micromachined silicon nitride membranes for thermal infrared detectors”. X. A. Liu. Stemme. B. al. 2003. ”Enhanced characteristics of an uncooled microbolometer using vanadium-tungsten oxide as a thermometric material”. Xin-jian. Chi-Anh. Proc.44. Chen. [59] N. M.4. USA. Stemme.-E.85. T. Journal of Microelectromechanical Systems.[49] P. Chaker. G. Eminoglu. Karunagaran. France. H. International Journal of Infrared and Millimeter Waves . Soltani. Shin. VanHulsel. K. Mangalaraj. [61] R. 2003. 2005. A. S. 2007. R. J. Eriksson.11.260-264. Kälvesten. P. A. USA. Vol. T.6. [60] S. of SPIE Vol.696-700.14. 201. J.74-84. Liu. Kälvesten. Lambkin. Part 3.6542. D. J.Y. D. Dong.1. Smart Mater. Kvisterøy.425. Sasaki. Källhammer. Journal of Applied Physics: Applied Physics Reviews. Vol. Tregilgas. Orlando. [67] M. Tanaka. Y. Struct. Proc. H.J. Eriksson. Vol. J. Vieider. J. Proc. S. VanNylen. S. Tanrikulu. U.F. 2007. 1997. Yue.881.11. “Adhesive wafer bonding”. Sensors & Actuators A. Joseph and V. Margot. E. Gutmann. A. No. Park.609. Mathewson.T. P. Källhammer. Savage. Lv. Jakobsen. Vol. Niklaus. P.. R. “A new concept for CMOS-compatible fabrication of uncooled infrared focal plane arrays using wafer-scale device transfer bonding”. “Amorphous silicon microbolometer technology”. S.102-113. V.509-513. M. Wissmar. [58] C. T. H. Torres.

Part A: Vacuum.635-638. pp. A. Vol.407-414. 2006. Neuzil. Surfaces and Films. [93] P. Lee. [75] D. Lehto. Grishin. Karunasiri. A. Vol. N. J. Rana.748755.11. 2000.11. “Semiconducting YBaCuO microbolometers for uncooled broad-band IR sensing”. Motooka: Handbook of Silicon Based MEMS Materials and Technologies. Robert. Cherepov. [78] V. Infrared Physics and Technology.2552. Susan. J. A. T. [81] S. M. Y. Gotoh. Ahmed.514. Chistokhin. Liddiard. “High signal to noise ratio quantum well bolometer materials”. W. Vol. Y-C. Orlando. Ambrosio.995-1003. 2002. Proc. K. Butler. Kosarev. N. T. pp. Ringh. N.2. “Sangrador IR uncooled bolometers based on amorphous GeSiO on silicon micromachined structures”.6542. Vol. N. to be published by William Andrew Publishing. L.87. Proc. U. A. Tal. No.19. Matsumoto. Vol.1844-1850. pp. Vol. Jansson. C.Touvignon. De Moor. Zeng.4945.21. Sridhar. N. M. H. Proc. Applied Physics Letters. Tilli. Z. Arnaud.264-73. “Infrared focal plane array incorporating silicon IC process compatible bolometer”.316. [76] A. pp.297302.1. Proc. Y.5126. pp. “Noise behavior of amorphous GeSiO for microbolometer applications”. “Radio Frequency sputtered SiGe and SiGeO thin films for uncooled infrared detectors”. Fieque. J. Vol. 2007. P. Grietens. B. C. Woolaway.Crastes. E. Liu. T. Tanaka. pp. Gillham. Trouilleau. Vol. pp. International Journal of Infrared and Millimeter Waves. Yildiz. Leonov.355-360. Vol. Vol. Lindroos. Aizawa. SPIE 2007. He. Mei.465-472. [80] R. Ichihara. O. Vieider. Sweden.43. Proc. B. “A high performance amorphous Si1-xCx:H thermistor bolometer based on micro-machined structure”. J. A. “FOA/DSTO uncooled IRFPA development”. Verbist. [79] I. Vol. pp. K. 2000.6401. pp. “Fabrication and performance comparison of planar and sandwich structures of micro-bolometers with Ge thermo-sensing layer”. Hoglund. of the SPIE 2001.33502-1-318. [74] E. Vol.3698. “Polycrystalline layers of silicon-germanium alloy for uncooled IR bolometers”. SPIE 1999. Tsukamoto. L. R. “Uncooled amorphous silicon 160 x 120 IRFPA with 25 µm pixel-pitch for large volume applications”. “Monolithic uncooled 8 × 8 bolometer arrays based on poly-SiGe thermistor”. S. Vol. “Improvements in uncooled systems using bias equalization”.7. “Micromachined poly-SiGe bolometer arrays for infrared imaging and spectroscopy”. No.1-1V. [92] W.6.233-235. [90] C.106-109. I. pp. Liddiard. U. Minassian. S. 2005. Wissmar. “Performance of microbolometer focal plane arrays under varying pressure”. Proc. pp.46. Vol.J. Thin Solid Films.K. USA. [84] M. Proc. S. Liger. [86] J. Proc. No. Perova. of SPIE Vol.Ca)xMnO3 membrane on Si for uncooled infrared microbolometer”. "Theoretical analysis of pulse bias heating of bolometers and effectiveness of pulse bias compensation". “A 32*32 parylene-pyrolyzed carbon bolometer imager”. G.8. [91] C. . J. B. Manning. Lim. Tanaka. No. K.2. pp. Ericsson. SPIE 2006. Hijikawa. Yaradanakul. T. IEEE Transactions on Electron Devices. A. R. [83] H. No. USA. pp. A. SPIE 1995. B.5406. Proc. [89] V. Yoon. Kumazawa. P. Vol. Vermeiren. 1996. A. pp. Robinson. Nakazato.54-63. 2008.468-472. A.76070-7610. [85] L.1253-1256. No. Michailovsky. of SPIE 2004.Legras. Parkinson.4. L. Proc. K. MEMS 2006. Torres. pp. Dong. K. Andersson. No. Fomin. “Free-standing epitaxial La1-x(Sr.1-2. Proc. No. A. Clement. Vol.515. L. “Room temperature YBaCuO microbolometers” Journal of Vacuum Science and Technology. “Silicon foundry micro-bolometers . K. C.27. Watabe.322-329. E. Vol. Ringh. [77] M. No. pp. Chiba. Journal of Microelectromechanical Systems. Vol. M. Goessens. S. San Diego.[73] A. Dupont.3698. Electron Device Letters. Tinnes. Y. A. pp. pp.5.644-653. Stockholm. P.264-275. 6836 68360D-15 . Celik-Butler. Orlando.3. C. USA. Honda. Parrish. T. B. P. Yon. Tranducers 1997. No. SPIE 1999. Orlando. Itoh. Yue. M. S. Journal of non-crystalline solids. SPIE 2003. 2003.2. Okuyama. “Electrical properties of photo-CVD boron-doped hydrogenated nanocrystalline siliconcarbide (p-nc-SiC:H) films for uncooled IR bolometer applications”. Butler. Iborra. Phong. Kaushal. Proc.The route to the mass-market thermal imager”. U. No. D. pp. Teranishi. [87] M. Jansson. Endoh. Moreno. Tissot. J.4369. Thin Solid Films. Kim. Vol. Qiu. T. Van Hoof. S. SPIE 2003. USA. 2005. Myong. Tait. Vol. Herrero.18. Du Bois. pp. Almasri. [94] C.1V. [88] X. [82] T.

Sign up to vote on this title
UsefulNot useful